WO2022156446A1

WO2022156446A1 - Method and apparatus for determining summary of search result, and electronic device

Info

Publication number: WO2022156446A1
Application number: PCT/CN2021/138921
Authority: WO
Inventors: 李悦; 张志凌; 李顶圣
Original assignee: 华为技术有限公司
Priority date: 2021-01-19
Filing date: 2021-12-16
Publication date: 2022-07-28
Also published as: CN114817520A

Abstract

The present application relates to the technical field of the Internet, and provides a method and an apparatus for determining a summary of a search result and an electronic device, wherein the electronic device may be a server, and the method comprises: determining a preset length of characters starting from a starting position of a target paragraph, evaluating the last sentence of text of the preset length of characters to obtain a first evaluation result and, on the basis of the first evaluation result, determining a first capturing position of the target paragraph, the first capturing position being any one of the following: the end position of text obtained after supplementing the last sentence of text, the end position of the last sentence of text, and the end position of text obtained after deleting the last sentence of text; and, on the basis of the first capturing position, determining a summary of the search result from the target paragraph. The present method can flexibly adjust the position of the text captured in a webpage such that the information expressed in the last sentence of the obtained webpage summary is more complete, thereby improving the presentation effect of the webpage summary and enhancing the reading experience of the user.

Description

Method, device and electronic device for determining abstract of search result

This application claims the priority of the Chinese patent application with the application number 202110072051.5 and the application title "Method, Apparatus and Electronic Equipment for Determining the Abstract of Search Results", which was submitted to the State Intellectual Property Office on January 19, 2021, the entire contents of which are by reference Incorporated in this application.

technical field

The present application relates to the field of Internet technologies, and in particular, to a method, apparatus and electronic device for determining an abstract of a search result.

Background technique

With the rapid development of Internet technology, people are more and more accustomed to using search engines to find the information they need. The search engine can provide the URL and abstract of the corresponding webpage according to the search query submitted by the user, and the abstract is used to describe the general content of the webpage, and the user can determine whether to browse the webpage corresponding to the abstract by reading the abstract.

In the related art, the text of a certain length of characters is usually intercepted from the webpage text as the webpage summary. The reference value is very small, and it will also lead to poor presentation of the abstract, thereby affecting the user's reading experience.

SUMMARY OF THE INVENTION

The present application provides a method, device and electronic method for determining an abstract of a search result, which can flexibly adjust the position of text intercepted in a webpage, so that the information expressed in the last sentence of the obtained webpage abstract is more complete, thereby improving the webpage abstraction. The presentation effect improves the user's reading experience. The specific plans are as follows:

In a first aspect, an embodiment of the present application provides a method for determining an abstract of a search result, including:

Determine a preset length of characters from the starting position of the target paragraph, wherein the target paragraph is the paragraph where the abstract of the search result is located;

Evaluate the last sentence of text of the preset length characters to obtain a first evaluation result, wherein the target paragraph is the paragraph where the abstract of the search result is located;

According to the first evaluation result, the first interception position of the target paragraph is determined, wherein the first interception position is any one of the following: the end position of the text obtained by supplementing the last sentence of text, the Describe the end position of the last sentence of text, and the end position of the text obtained after deleting the last sentence of text;

According to the first interception position, an abstract of the search result is determined from the target paragraph, wherein a character length corresponding to the abstract of the search result is within a preset length range.

The method for determining the abstract of the search result provided by the embodiment of the present application determines a preset length of characters from the starting position of the target paragraph, evaluates the last sentence of the determined preset length of characters, and obtains the first evaluation according to the evaluation. As a result, the first interception position is determined. It can be seen that the embodiment of the present application flexibly determines the first interception position according to the first evaluation result, so that the last sentence of the text of the abstract of the search result determined according to the first interception position can be a sentence with relatively complete semantics. The text improves the reference value of whether the user browses the web page. When the summary of the search results is displayed on the client, the rendering effect is better, which reduces the amount of invalid reading by the user, improves the user's reading experience, and improves the user's click on the summary. Web page probability. In addition, the first interception position determined in the embodiment of the present application is the end position of the text obtained after supplementing the last sentence of text, the end position of the last sentence of text, or the end position of the text obtained after deleting the last sentence of text. According to the different first evaluation results, the first interception positions determined in the application embodiment are also different, and the determined first interception positions are further flexibly determined, so that the determined abstract has a better presentation effect.

In an optional embodiment, the first evaluation result is: the end of the last sentence of text is incomplete topic information;

The first interception position is: the end position of the text obtained by supplementing the subject information at the end of the last sentence of text.

In this embodiment, the topic information at the end of the last sentence of text is evaluated, and the obtained first evaluation result is that the end of the last sentence of text is incomplete topic information. When a user is reading a text, for subject information such as movie and TV drama titles, book titles, etc., the user usually prefers to read the complete content of these subject information to better understand the text. The first evaluation result obtained in this embodiment is the end of the last sentence. For incomplete subject information, when determining the first interception position, the first interception position may be determined as a position that can complete the above subject information according to the first evaluation result, so that the user can better understand the text.

In an optional embodiment, the first evaluation result is: the semantics of the last sentence of text is incomplete;

The first interception position is: the end position of the text obtained after the last sentence of text is supplemented into a complete sentence of text.

In this implementation, the semantics of the last sentence of text is evaluated, and the first evaluation result obtained is that the semantics of the last sentence of text is incomplete. When the semantics of a sentence of text is incomplete, the user usually cannot know the specific content represented by the text after reading the text. Therefore, the first evaluation result obtained in this embodiment is that the semantics of the last sentence is incomplete. At the first interception position, the text of the last sentence can be completed according to the first evaluation result, which is convenient for the user to read the text.

In an optional embodiment, the first evaluation result is: the character length corresponding to the last sentence of text is less than the first preset length;

The first interception position is: the end position of the text obtained after deleting the last sentence of text.

In this embodiment, the length of the character corresponding to the last sentence of text is evaluated, and the obtained first evaluation result is that the length of the character corresponding to the last sentence of text is less than the first preset length. Through this embodiment, it can be quickly evaluated that the semantic integrity of the last sentence of text is low and the reference value for users to understand the web page is low, so that when determining the first interception position, it can be quickly determined that the above preset length characters The last sentence is removed to more quickly identify snippets that show better search results to users.

In an optional embodiment, the first evaluation result is: the number of words corresponding to the last sentence of text is less than a second preset number;

In this embodiment, the number of words corresponding to the last sentence of text is evaluated, and the obtained first evaluation result is that the number of words corresponding to the last sentence of text is less than the second preset number. Since the number of words in a sentence of text can usually reflect the semantic integrity of a sentence of text, evaluating the number of words in the last sentence of text can more accurately evaluate the semantic integrity of the last sentence of text and the reference value for users to understand web pages. Through the embodiments of the present application, it can be quickly assessed that the semantic integrity of the last sentence of text is low and the reference value for users to understand the web page is low, so that when the first interception position is determined, it can be quickly determined that the above preset length The last sentence of characters is removed to more quickly identify snippets that show better search results to users.

In an optional embodiment, the first evaluation result is: the proportion of characters other than words and words in the last sentence of text is greater than a preset proportion;

In the embodiment of the present application, the proportion of characters other than words and words in the text of the last sentence is evaluated, and the obtained first evaluation result is that the proportion of characters other than words and words in the text of the last sentence is greater than the preset proportion. Users usually understand abstracts more fully through words and phrases that have actual semantics. Characters other than words and words in abstracts are usually less helpful for users to understand abstract content. Therefore, words and words in texts are usually the ones that users understand. The more effective information of the abstract, the embodiment of the present application can estimate that the proportion of effective information in the last sentence of text is small, and when determining the first interception position, it can be quickly determined to delete the above-mentioned last sentence of text with a preset length of characters. .

In an optional embodiment, the first evaluation result is: the end of the last sentence of text is a punctuation mark;

The first interception position is: the end position of the text obtained after deleting the punctuation mark at the end of the last sentence of text.

When the abstract is displayed, if there is a punctuation mark at the end of the abstract, the reference value of this punctuation mark to the user's reading comprehension is very small, and the punctuation mark at the end may also affect the display effect of the abstract on the client side. Therefore, when evaluating the last sentence When there is a punctuation mark at the end of the text, you can delete the punctuation mark at the end to make the summary display better.

In an optional embodiment, the first interception position is: the end position of the text obtained after deleting the last sentence of text;

The determining the abstract of the search result from the target paragraph according to the first interception position includes:

It is determined that the character length corresponding to the text located before the first interception position in the target paragraph is less than the lower limit of the preset length range;

Determine the second interception position, wherein, the second interception position is: the end position of the text obtained after the last sentence of text is supplemented;

A summary of the search result is determined from the target paragraph based on the second clipping position.

In this embodiment, when the first interception position is the end position of the text obtained after deleting the last sentence of text, if the character length corresponding to the text in the target paragraph before the first interception position is less than the lower limit of the preset length range , indicating that the length of the characters corresponding to the text before the first interception position is too small. In this case, the summary of the determined search result may contain very little information, so that the user cannot obtain the effective information of the search result from the summary. In this case, the abstract of the search result is determined from the target paragraph according to the second interception position, since the second interception position is the end position of the text obtained after the last sentence of text is supplemented, the characters corresponding to the abstract of the obtained search result can be increased. length, so that users can learn more about the search results from the snippet.

In an optional embodiment, the determining the abstract of the search result from the target paragraph according to the second clipping position includes:

Determine that the character length corresponding to the text before the second interception position in the target paragraph is greater than the upper limit of the preset length range;

The preset length characters are determined as an abstract of the search result.

If the character length corresponding to the text before the second interception position in the target paragraph is greater than the lower limit of the preset length range, it means that the character length corresponding to the text before the second interception position is too long. The number of entries is too small, or the determined abstract cannot be displayed completely, which affects the layout of the page. In this case, the preset length of characters is determined as the abstract of the search result, so that the determined abstract will not be too long. , and it will not be too short, which is conducive to the layout of the page.

In an optional embodiment, the first interception position is: the end position of the text obtained after the last sentence of text is supplemented;

It is determined that the character length corresponding to the text located before the first interception position in the target paragraph is greater than the upper limit of the preset length range;

Determine the third interception position, wherein, the third interception position is: the end position of the text obtained after deleting the last sentence of text;

A summary of the search result is determined from the target paragraph based on the third clipping position.

In this embodiment, when the first interception position is the end position of the text obtained by supplementing the last sentence of text, if the character length corresponding to the text before the first interception position in the target paragraph is greater than the second preset length, it means that The character length corresponding to the text before the first interception position is too long. In this case, since the display space of each search result on the display page is usually determined, the summary of the determined search result may not be displayed completely. In this case, the abstract of the search result is determined from the target paragraph according to the third interception position, since the third interception position is the end position of the text obtained after deleting the last sentence of text, the character length corresponding to the abstract of the obtained search result can be shortened, Enables the determined search results to be fully displayed in the display page.

In an optional embodiment, the determining the first interception position of the target paragraph according to the first evaluation result includes:

It is determined that the first evaluation result is any of the following: the end of the last sentence of text is complete topic information, the semantics of the last sentence of text is complete, and the character length corresponding to the last sentence of text is not less than the first preset. The length, the proportion of characters other than words and words in the last sentence of text is not greater than the preset proportion;

Evaluate the last sentence of text to obtain a second evaluation result, wherein the second evaluation result is different from the evaluation content corresponding to the first evaluation result;

According to the second evaluation result, a first interception position of the target paragraph is determined.

When the first evaluation result is that the end of the last sentence of text is the complete topic information, the semantics of the last sentence of text is complete, the length of the characters corresponding to the last sentence of text is not less than the first preset length, or the last sentence of text except for words and words When the proportion of characters is not greater than the preset proportion, it means that when evaluating the last sentence of text, the last sentence meets the evaluation requirements, that is, the last sentence meets certain integrity requirements. In this case, in order to make the obtained The completeness of the abstract is better and the presentation effect of the abstract is better. The last sentence of text can be further evaluated with different contents to obtain a second evaluation result, and the first interception position can be determined according to the second evaluation result.

In an optional embodiment, the determining the abstract of the search result from the target paragraph according to the first interception position includes:

Evaluate the last sentence of the text in the target paragraph before the first interception position to obtain a third evaluation result;

According to the third evaluation result, determine the fourth interception position of the target paragraph;

A summary of the search result is determined from the target paragraph based on the fourth clipping position.

In this embodiment, after the first interception position is determined, the integrity of the last sentence of the text in the target paragraph before the first interception position may still be low. The last sentence of the text before an interception position is evaluated, a fourth interception position is determined according to the obtained third evaluation result, and an abstract of the search result is determined from the target paragraph according to the fourth interception position. That is to say, after each interception position is determined in this embodiment, the last sentence of the text in the target paragraph that is located before the most recently determined interception position is re-evaluated until the last sentence of the text before the most recently determined interception position. The text of the last sentence of the text satisfies the requirement of complete display of the abstract, so that the completeness and display effect of the abstract of the determined search result can be improved.

In a second aspect, an embodiment of the present application also provides a device for determining a summary of a search result, including:

The evaluation module is used to determine the preset length characters from the starting position of the target paragraph, and evaluate the last sentence of text of the preset length characters to obtain the first evaluation result, wherein the target paragraph is the search result. The paragraph in which the summary of the results is located;

A determination module, configured to determine a first interception position of the target paragraph according to the first evaluation result, wherein the first interception position is any one of the following: a text obtained by supplementing the last sentence of text The ending position of the text, the ending position of the last sentence of text, the ending position of the text obtained after deleting the last sentence of text, according to the first interception position, the summary of the search result is determined from the target paragraph, wherein , the character length corresponding to the abstract of the search result is within a preset length range.

In an optional embodiment, the first evaluation result is: the character length corresponding to the last sentence of text is less than the first preset length, or the proportion of characters other than words and words in the last sentence of text greater than the preset ratio;

The determining module is specifically used for:

In an optional embodiment, the determining module is specifically configured to:

It is determined that the first evaluation result is any of the following: the end of the last sentence of text is complete topic information, the semantics of the last sentence of text is complete, and the character length corresponding to the last sentence of text is not less than the first preset. Length, the proportion of characters other than words and words in the last sentence of text is not greater than the preset proportion;

In an optional embodiment, the determining module is specifically used for:

In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor, a memory, and an interface;

The processor, the memory, and the interface cooperate with each other, and the processor is configured to perform the method of any one of the first aspects.

In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the processor executes the first aspect The method of any of the above.

In a fifth aspect, the embodiments of the present application further provide a computer program product, which, when the computer program product runs on a computer, causes the computer to execute the method described in any one of the first aspects.

Description of drawings

1a is a schematic diagram of a search page of an example search engine provided by an embodiment of the present application;

Figure 1b is a schematic diagram of a browser page;

2 is a schematic diagram of content corresponding to a search result displayed by an example of a browser provided by an embodiment of the present application;

3 is a schematic structural diagram of an example of a search system provided by an embodiment of the present application;

4 is a schematic flowchart of an example of a method for determining an abstract of a search result provided by an embodiment of the present application;

FIG. 5 is an example diagram of determining an abstract of a search result from a target paragraph provided by an embodiment of the present application;

6 is another example diagram of determining an abstract of a search result from a target paragraph provided by an embodiment of the present application;

7 is a schematic flowchart of another example of a method for determining an abstract of a search result provided by an embodiment of the present application;

8 is a plurality of example diagrams of the last sentence of a preset length of characters provided by an embodiment of the present application;

9 is a schematic diagram of an example of word segmentation of text provided by an embodiment of the present application;

10 is a schematic diagram of another example of word segmentation of text provided by an embodiment of the present application;

11 is an example diagram of an example of determining an abstract of a search result from a target paragraph provided by an embodiment of the present application;

12 is a schematic flowchart of another example of a method for determining a summary of a search result provided by an embodiment of the present application;

13 is a schematic diagram of displaying on a browser the abstracts of each search result determined using the method for determining the abstract of the search result provided by the embodiment of the present application;

14 is a schematic flowchart of another example of a method for determining an abstract of a search result provided by an embodiment of the present application;

15 is a schematic flowchart of an example of determining a target paragraph provided by an embodiment of the present application;

16 is a schematic structural diagram of an apparatus for determining a summary of a search result provided by an embodiment of the present application;

FIG. 17 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings in the embodiments of the present application. Wherein, in the description of the embodiments of the present application, unless otherwise stated, “/” means or means, for example, A/B can mean A or B; “and/or” in this document is only a description of the associated object The association relationship of , indicates that there can be three kinds of relationships, for example, A and/or B, can indicate that A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present application, "plurality" refers to two or more than two.

Hereinafter, the terms "first", "second" and "third" are only used for descriptive purposes, and should not be understood as indicating or implying relative importance or implying the number of indicated technical features. Thus, a feature defined as "first", "second", "third" may expressly or implicitly include one or more of that feature.

In the following, application scenarios of the embodiments of the present application are introduced.

Users usually use search engines to search for the information they need. Specifically, users can enter a search query on the client. The client can be an electronic device such as a mobile phone, tablet computer, smart watch, computer, etc., and a browser can be installed on the client. The application program corresponding to the search engine can be installed. The search query can be one or more keywords, a piece of text or a formula, etc., or other search queries (such as a picture). After entering the search query, you can query each web page matching the search query as the search result, and obtain the title, hyperlink and abstract of each web page queried, and send the obtained title, hyperlink and abstract to the client, the customer The terminal displays the received title, hyperlink and summary on the display screen for the user to view. The abstract of the search result can describe the general content of the web page, and can reflect why the web page is relevant to the search query input by the user.

FIG. 1 a is a schematic diagram of a search engine page of a browser, FIG. 1 b is a schematic diagram of a browser page, and FIG. 2 is a schematic diagram of content corresponding to a search result displayed by the browser. For example, the user may enter a search query (eg, wood floor) in the search box 102 of the search engine page of the browser shown in FIG. 1a; or, as shown in FIG. 1b, the user may also enter the search query in the address bar 101 of the browser Search query (for example, water heater type), after receiving the search query, the browser can jump to the search engine page and display the search result. FIG. 2 shows the titles and abstracts of each search result displayed by the browser after the user enters the search query of “wooden floor” in the search box 102 in FIG. 1a. In FIG. 2, the browser also displays the Uniform Resource Location of the search results. (uniform resource locator, URL), when there is an image in the web page corresponding to the search result, the browser also displays the image corresponding to the search result. The last sentences of the second, third and fourth abstracts in Figure 2 correspond to "Yiqiang", "otherwise" and "here" respectively. It can be seen that the last sentences of the second, third and fourth abstracts are not helpful for users to browse the webpage. Provide more valuable reference information, so that when the last sentence of text is displayed on the browser, the rendering effect of the abstract is not good, which will increase the invalid reading volume of users and affect the user experience.

In order to solve the above problem, the embodiments of the present application provide a method and apparatus for determining an abstract of a search result, which can improve the presentation effect of a webpage abstract and improve the user's reading experience.

The executive body of the method for determining the summary of the search result provided by the embodiment of the present application may be a server, the server may be a server of a search engine, or a server of other search systems, and the server can crawl web pages and associate information with web pages Stored in the information database, the server can perform information search in the information database according to the search query received from the client.

Below, the method for determining the abstract of the search result provided by the present application will be introduced through the accompanying drawings and embodiments. In the following embodiments, the execution subject is a server as an example for description.

FIG. 3 is a schematic structural diagram of a search system 300 according to an embodiment of the present application. As shown in FIG. 3 , the method for determining the abstract of a search result provided in this embodiment of the present application may be applied to a search system 300 , and the search system 300 may include a client 310 and a server 320 , and the client 310 and the server 320 are connected through a network 330 . The server 320 is configured to receive the search query sent by the client 310 through the network 330, determine the abstract of the search result according to the search query, and send the determined abstract to the client 310 through the network 330. The server 320 can also be used to Network 330 crawls individual web pages or other information. The client 310 is configured to acquire the search query input by the user through the network 330, send the search query to the server 320, receive the abstract of the search result sent by the server 320, and display the received abstract on the display. Client 310 may be an electronic device. The search system 300 may include one server 320, may include multiple servers 320, may include one client 310, or may include multiple clients 310, which is not limited in this application. In FIG. 3 , the search system 300 includes three servers 320 and two clients 310 .

In one embodiment, a browser 311 may be installed on the client 310, and the browser 311 may provide an interface of a search engine, so that a user can send a search query to the search engine through the browser 311. The server 320 may be a server 320 corresponding to a search engine. The search engine may search the index database for search results matching the search query input by the user, and may obtain web pages corresponding to the search results selected by the user from the information database.

Hereinafter, the method for determining the abstract of the search result provided by the present application will be introduced through specific embodiments.

Example 1

FIG. 4 is a schematic flowchart of a method for determining an abstract of a search result provided by an embodiment of the present application. As shown in FIG. 4 , the abstract of a search result may be determined according to the following steps S401 to S416, and the execution subject of steps S401 to S416 is: server.

Step S401: Acquire the text information of the search result, and determine the target paragraph from the text information.

The above target paragraph is the paragraph where the abstract of the search result is located.

Figures 5 and 6 show two examples of determining the abstract of the search result from the target paragraph. For example, as shown in FIG. 5, the determined paragraph of the abstract may be the following first paragraph 5a: "Mary is one of the most popular singers and dancers. Mary is born in August 4th, 1969. She is very good at singing and acting. Her songs are always pop music, but there are also countryside music, hip-hop, and rock. She composed many of her own songs.”

For another example, as shown in Fig. 6, the determined paragraph of the abstract may also be the following second paragraph 6a: "The process of injection molding is: at a certain temperature, by stirring a completely molten plastic material, injecting it with high pressure The mold cavity is cooled and solidified to obtain a molded product. The advantages of the injection molding method are that the production speed is fast, the efficiency is high, the product integrity is good, the production process can be automated, the output products have various shapes, accurate product dimensions, and products are easy to replace. The specific process of injection molding is based on the content of Chapter 3 of "Production Methods of Plastic Products". The parameters in the injection molding process mainly include injection pressure, injection time, injection temperature, pressure holding time and temperature, back pressure pressure."

The above-mentioned first paragraph 5a and second paragraph 6a are examples when the paragraph where the abstract is located is in English and Chinese respectively. The specific content of the search results and the corresponding language are different, and the content and language of the paragraph where the abstract is located are also different. The second paragraph 6a is the paragraph where the two abstracts corresponding to the search results determined for the two different search queries are located.

Step S402: Determine a preset length of characters from the start position of the target paragraph.

For example, when the preset length is 140 characters, for the first paragraph 5a above, as shown in Figure 5, the preset length characters 5b determined by the server are: "Mary is one of the most popular singers and dancers. Mary is born in August 4th, 1969. She is very good at singing and acting. Her songs are", the determined preset length characters are non-Chinese characters; for the second paragraph 6a above, as shown in Figure 6, the server determines The preset length of the characters 6b is: "The process of injection molding is: at a certain temperature, by stirring the completely molten plastic material, injecting it into the mold cavity with high pressure, and cooling and solidifying to obtain a molded product. The advantages of the injection molding method The production speed is fast, the efficiency is high, the product integrity is good, the production process can be automated, the output products have various shapes, accurate product dimensions, the products are easy to replace, and can be formed into complex shapes. The specific process of injection molding is based on "plastics". Products", the determined preset length characters are Chinese characters.

Wherein, the determined preset length characters include spaces. Since characters such as spaces and punctuation marks will occupy display space when the summary of the search result is displayed on the client, the determined preset length of characters is the actual character length including spaces.

Step S403: Evaluate whether the number of words corresponding to the last sentence of text with the preset length of characters is not less than the second preset number.

The evaluation result obtained in step S403 is the first evaluation result.

The number of words corresponding to the last sentence of text may be the sum of the number of Chinese single-character words and the number of non-Chinese words contained in the last sentence of text. Among them, non-Chinese words can include English words, German words, French words, etc., the number of words corresponding to punctuation marks is zero, at least one consecutive number is a non-Chinese word, and consecutive numbers and English characters are 1 Non-Chinese words, consecutive means adjacent and not separated by spaces. In this embodiment, a single-character word can be understood as a single character.

The above-mentioned second preset number may be any number from 2 to 5, or may be another smaller number.

For example, the second preset number is 4. As shown in FIG. 5 , the last sentence text of the preset length characters 5b determined by the server from the starting position of the first paragraph 5a is "Her songs are", which contains The number of words 3 is less than the second preset number 4, and the server evaluates that the number of words corresponding to the last sentence of text of the preset length characters 5b is less than the second preset number, that is, the evaluation result of step S403 is no.

For another example, as shown in FIG. 6 , the last text of the preset length characters 6b determined by the server from the starting position of the second paragraph 6a is “the specific process of injection molding is “plastic products”, which contains The number of words 15 is greater than the second preset number 4, and the server evaluates that the number of words corresponding to the last sentence of text of the above-mentioned preset length characters 6b is greater than the second preset number, that is, the evaluation result of step S403 is Yes .

The following is a few examples to illustrate how to determine the number of words corresponding to the last sentence of text.

For example, the text of the last sentence is "Rescreened in the Mainland in 3D version in August 2016", the number of single Chinese characters included in this sentence is 9 (9 Chinese single-character words), and the number of non-Chinese words is 3 (respectively 2016, 8, and 3D), so the number of words corresponding to this text is 9+3=12 words.

For another example, the last sentence of text is "Mary Ly is the name of the protagonist", the number of single Chinese characters included in this text is 7 (7 Chinese single-character words), and the number of English words is 2 (2 non-Chinese words), so the number of words corresponding to this text is 7+2=9 words.

For another example, the text of the last sentence is "The Production Method of Plastic Products" was published in 1995. The number of single Chinese characters included in this sentence is 12 (12 Chinese single-character words), and the number of non-Chinese words is 1 (19995), so the number of words corresponding to this text is 12+1=13 words.

Step S403 can also be replaced by step S404.

Step S404: Evaluate whether the character length corresponding to the last sentence of text of the preset length characters is not less than the first preset length.

The evaluation result obtained in step S404 is the first evaluation result. Step S404 is not shown in the figure.

The last sentence of text may contain Chinese characters, English characters, or other non-Chinese characters. In this embodiment of the present application, the character length corresponding to the last sentence of text can be understood as the sum of the characters corresponding to the last sentence of text. Length, that is, the total length of all kinds of characters contained in the last sentence of text. Specifically, the character length corresponding to the last sentence of text may be the sum of the number of Chinese characters and the number of non-Chinese characters contained in the last sentence of text, where one Chinese character is one Chinese character, one English letter is one non-Chinese character, and one A full-width punctuation mark is one Chinese character, a half-width punctuation mark is one non-Chinese character, a number is one non-Chinese character, and spaces are zero characters.

The above-mentioned first preset length can be any length from 3 to 8 characters, or can be other specific lengths. When the language of the target paragraph is different, the value of the first preset length can be different. When the language of the paragraph is English as shown in 5a in FIG. 5, the first preset length can be any value from 10 to 15. When the language of the target paragraph is Chinese as shown in 6a in FIG. 6, the first preset length The length can be any value from 3 to 8, because the semantic integrity of Chinese and English corresponding to the same character length is different (Chinese semantics are more complete), so the first preset length corresponding to Chinese is smaller than English (or other languages that use letters to form words) corresponding to the first preset length. When the first preset length is set to be longer, in step S404, the shorter and incomplete last sentence of text can be deleted with greater probability, so that the end of the abstract of the finally obtained search result is more semantically complete. Since some short sentences may also express complete semantics, when the first preset length is set to be shorter, the probability of mistakenly deleting the last sentence with complete semantics or larger reference value can be reduced in step S404. Those skilled in the art can set the specific value of the first preset length according to the actual scene.

For example, the first preset length is 13. As shown in FIG. 5 , the last sentence text of the preset length characters 5b determined by the server from the starting position of the first paragraph 5a is "Her songs are", and the corresponding If the character length 11 is less than the first preset length 13, the server evaluates that the character length corresponding to the last sentence of text of the preset length characters 5b is less than the first preset length, that is, the evaluation result of step S414 is no.

For another example, the first preset length is 5. As shown in FIG. 6 , the last sentence of the preset length characters 6b determined by the server from the starting position of the second paragraph 6a is “The specific process of injection molding is "Plastic products", the corresponding character length 15 is greater than the first preset length 5, and the server evaluates that the character length corresponding to the last sentence of text of the above preset length characters 6b is greater than the first preset length, that is, step S414 The evaluation result is yes.

For the convenience of understanding, the following examples illustrate how to determine the character length corresponding to the last sentence of text.

For example, the last sentence of text is "Re-released in 3D version in Mainland China in August 2016", the number of Chinese characters included in this sentence is 9 (9 Chinese characters), and the number of numbers is 6 (6 non-Chinese characters) character), the number of English letters is 1 (1 non-Chinese character), so the character length corresponding to this text is 9+6+1=16 characters.

For another example, the last sentence of text is "Mary Ly is the protagonist's name", the number of Chinese characters included in this text is 7 (7 Chinese characters), and the number of English letters is 6 (6 non-Chinese characters), Therefore, the character length corresponding to this text is 7+6=13 characters.

For another example, the text of the last sentence is "The Production Method of Plastic Products" was published in 1995. The number of Chinese characters included in this sentence is 12 (12 Chinese characters), and the number of numbers is 4 (4 non-Chinese characters). character), the number of full-width punctuation marks is 2 (2 Chinese characters), so the character length corresponding to this text is 12+4+2=18 characters.

Steps S403 and S404 can also be understood as evaluating the format of the last sentence of text. For the specific process of determining the last sentence of the preset length of characters in steps S403 and S404, reference may be made to step S703 in the second embodiment.

When the evaluation result of step S403 is negative, or the evaluation result of step S404 is negative, step S405 is performed, and when the evaluation result of step S403 is yes, or the evaluation result of step S404 is positive, step S407 is performed.

Step S405: Delete the last sentence of text.

For example, as shown in FIG. 5, for the above-mentioned first paragraph 5a, the last sentence of the text "Her songs are" of the preset length characters 5b determined from the starting position of the above-mentioned first paragraph 5a can be deleted, and "Her songs are" can be deleted. The following text information 5c is: "Mary is one of the most popular singers and dancers. Mary is born in August 4th, 1969. She is very good at singing and acting." The end position of the text after deleting "Her songs are" For the first interception position, the length of the text information 5c after deleting "Hersongs are" is 127 characters.

Step S406: Determine the end position of the text obtained after deleting the last sentence of text as the first interception position, and determine the first interception position as the latest interception position.

After step S406 is performed, step S408 may be performed.

Step S407: Determine the end position of the last sentence of text as the first interception position, and determine the first interception position as the latest interception position.

After the execution of step S407 is completed, step S410 may be executed.

Step S408: Determine whether the character length corresponding to the text located before the first clipping position in the target paragraph is not less than the lower limit of the preset length range.

For example, when the preset length range is 120 to 200 characters, as shown in FIG. 5 , in the example in step S405, the length of the text information 5c after deleting "Hersongs are" is 127 characters and not less than 120 characters. The lower limit of ~200 is 120.

For another example, when the preset length range is 130 to 200 characters, as shown in FIG. 5 , in the example in step S404, the length of 127 characters in the text information 5c after “Hersongs are” is deleted is less than 130 characters. The lower limit of ~200 is 130.

If the judgment result of step S408 is NO, execute step S409, and if the judgment result of step S408 is YES, execute step S410.

Step S409 : supplementing the last sentence of text as a complete text, and updating the latest interception position to the end position of the text obtained by adding a complete text.

The most recent clipping position described above is used to determine the summary of the search results.

After step S409 is executed, step S415 may be executed, or step S413 may be executed directly.

For example, if the paragraph where the abstract is located is the above-mentioned first paragraph 5a, as shown in FIG. 5 , the text information after the server deletes the last sentence of text with the determined preset length of characters 5b is 5c, which includes 127 characters. The length ranges from 130 to 200 characters. Since 127 is less than 130, the text 5d obtained after adding the last sentence of the preset length of characters 5b to a complete sentence is "Mary is one of the most popular singers and dancers" .Mary is born in August 4th, 1969. She is very good at singing and acting. Her songs are always pop music,", the obtained text includes 158 characters, which is greater than the lower limit of 130.

Step S410: Evaluate whether the end of the text before the latest clipping position is incomplete topic information.

In this embodiment, the latest interception position is determined in steps S406 and S407, and the latest interception position is updated in step S409. The server may update and adjust the latest interception position one or more times according to the text before the determined first interception position. For example, step S409 is updated once, so the server updates the text before the latest interception position every time. to evaluate.

For example, as shown in Figure 5, when the text before the latest interception position is 5d, the last sentence of text is "Her songs are always pop music,", and the server evaluates that the end of "Her songs are always pop music," is not Incomplete subject information.

For another example, as shown in FIG. 6, when the text before the latest clipping position is a preset length of characters 6b determined from the starting position of the target paragraph 6a (because the latest clipping position has not been updated and adjusted yet) , so the text before the latest interception position is the unadjusted preset length characters 6b), and the last sentence of the text 6b obtained after the latest text adjustment is "The specific process of injection molding is based on "plastic products", The server evaluated that "the specific process of injection molding ends with "plastic products" as incomplete subject information "plastic products".

If the evaluation result of step S410 is yes, then step S411 is executed, and if the evaluation result of step S410 is no, then step S412 is executed.

Step S411: Complete the incomplete topic information according to the target paragraph, and update the latest interception position to the end position of the text obtained after the topic information is completed.

For example, as shown in Figure 6, if the text before the latest interception position is 6b, and the last sentence of the text is "the specific process of injection molding is "plastic products", then "the specific process of injection molding is supplemented with "plastic products" For "the specific process of injection molding is "Production Method of Plastic Products", the supplemented text is shown as 6c in Figure 6.

After step S411 is performed, step S412 may be performed.

Step S412: Evaluate whether the semantics of the last sentence of the text before the latest clipping position are complete.

For example, as shown in Figure 6, if the text before the new interception position is the text shown in 6c in Figure 6, the last sentence of the text is "The specific process of injection molding is "Production Method of Plastic Products", this sentence The subject is "the specific process of injection molding", and there is no predicate after the subject. In this application, when a verb or adjective appears after the subject, it can be determined that there is a predicate after the subject. Therefore, it is considered that the text of this sentence lacks a predicate, so it is determined that the semantics of the text "the specific process of injection molding is based on "Production Method of Plastic Products" is incomplete.

For example, as shown in Figure 5, if the text before the latest interception position is the text shown in 5d in Figure 5, and the last sentence of the text is "Her songs are always pop music,", the server evaluates that "Her songs are always pop music," includes the subject "Her songs", the predicate "are" and the object "pop music", so the semantics of the text of this sentence is considered complete.

If the evaluation result of step S412 is yes, step S413 is executed, and if the evaluation result of step S412 is negative, step S414 is executed.

Step S413: Determine the text before the latest clipping position as the abstract of the search result.

In the process of determining the summary of the search result, after evaluating the number of words in the last sentence of text in step S403, after evaluating the character length corresponding to the last sentence of text in step S404, and evaluating the text before the latest interception position in step S410 After the topic information at the end, and after evaluating the semantic integrity of the last sentence of the text before the latest interception position in step S412, it is possible to update and adjust the latest interception position. Therefore, the number of updates of the latest interception position is It could be one time, it could be multiple times. When the server does not adjust the latest interception position after each evaluation, the latest interception position is the first interception position.

Step S414: Complete the semantics of the last sentence of the text before the latest clipping position according to the target paragraph, and update the latest clipping position to the end position of the text obtained after the semantics is completed.

For example, as shown in Figure 6, the text before the latest interception position is the text shown in 6c in Figure 6, and the last sentence of the text is "The specific process of injection molding is "Production Method of Plastic Products", and the server can convert this sentence It is added that "the specific process of injection molding is based on the content of Chapter 3 of "Production Methods of Plastic Products", and the obtained text is shown in 6d in Figure 6. It can be seen that when the semantics are completed, the last sentence becomes "the specific process of injection molding is based on the content of Chapter 3 of "Production Methods of Plastic Products", and the semantics of this sentence are complete.

Step S415: Determine whether the character length of the text before the latest clipping position is not greater than the upper limit of the preset length range.

If the judgment result of step S415 is yes, then step S413 is executed, and if the judgment result of step S415 is no, then step S416 is executed.

Step S416: Update the latest interception position to the first interception position.

For example, as shown in Figure 6, the text before the latest interception position is the text shown at 6d in Figure 6, which is: "The process of injection molding is: at a certain temperature, by stirring a completely molten plastic material, with High pressure is injected into the mold cavity, and the molded product is obtained after cooling and solidification. The advantages of the injection molding method are that the production speed is fast, the efficiency is high, the product integrity is good, the production process can be automated, the output products have various shapes, accurate product dimensions, and It is easy to be replaced and can be made into parts with complex shapes. The specific process of injection molding is based on the content of Chapter 3 of "Production Methods of Plastic Products". The length of the text shown in 6d is 153 characters. If the preset length range is 120 to 150 characters, then the character length 153 of the obtained text is greater than the upper limit of the preset length range. Therefore, update the latest interception position to the end position of 6c (ie, the first interception position), and the obtained text is shown in the figure 6e in 6 is shown.

After step S416 is performed, step S413 is performed.

If the character length of the text obtained after the deletion operation is performed in step S416 is less than the lower limit of the preset length range, the latest interception position is not updated. For example, as shown in Figure 6, the length of the text 6e obtained after the last step of the deletion operation in Figure 6 corresponds to 125 characters. If the preset length range is 130 to 160 characters, and the lower limit is 130 characters, at this time, Even if 125 is less than 130, the latest interception position is no longer updated. When the character length of the abstract exceeds the preset length range, the size of the abstract displayed on the client will take up a larger proportion, which affects the overall display effect of the webpage. Therefore, the abstract is not easy to be too long.

In this embodiment, steps S408 to S416 are to determine the abstract of the search result from the target paragraph according to the first clipping position.

FIG. 13 is a schematic diagram of displaying on a browser the abstracts of each search result determined by using the method for determining the abstract of the search results provided in the first embodiment. After the user enters the search query of "wooden floor", the abstract of each search result is determined using the method for determining the abstract of the search result provided in this embodiment, and the browser displays the content corresponding to each search result as shown in FIG. 13 . After the user enters the search query of "wooden floor", the preset length characters are directly intercepted from the starting position of the target paragraph of the webpage text corresponding to each search result, and the intercepted characters are determined as the webpage summary, and the browser displays each entry. The content corresponding to the search results is shown in Figure 2.

Compared with Fig. 2, Fig. 13 shows that the last sentence of the abstract of the second search result in Fig. 2 is "to be strong", and to evaluate "to be strong", the obtained evaluation result is that "by strong" (2 characters) corresponds to The length of the characters is less than the first preset length (for example, the first preset length is 5 characters), indicating that the length of the characters corresponding to the last sentence of text is small. It can be seen that the last sentence of text is difficult to express effective information, so the last sentence of text "" Delete with "strong", the summary of the search result obtained after deletion is shown in the second search result in Figure 13; the last sentence of the abstract of the third and fourth search results in Figure 2 is "otherwise", "here", In the same way, the character lengths corresponding to the last sentence of the two abstracts are both small, and the last sentence of the text is difficult to express effective information, so delete the last sentence of the two abstracts, and the abstract of the search results obtained after deletion is as follows: As shown in the third and fourth search results in Figure 13; the last sentence of the abstracts of the first and fifth search results in Figure 2 are "What does this shop have to do with whether it is good or not?" There is added formaldehyde", evaluate the last sentence of the abstract of the first and fifth search results, and the obtained evaluation result is that the text character length of the last sentence is not less than the first preset length and the semantics are complete, so these two The text of the last sentence of the abstract of the sixth search result in Figure 2 is "but it also has wood elimination", and the text of the last sentence of the abstract corresponding to the sixth search result is evaluated. The length is not less than the first preset length, and the evaluation result obtained by further evaluation is that the semantics of the last sentence of the abstract of this search result is incomplete, so the last sentence of the abstract of this search result is completed.

As can be seen from FIG. 13 , the last sentence of the abstract of each search result determined by using the solution of the embodiment of the present application is more complete, the presentation effect of the abstract is better, and the display is more friendly.

In this embodiment, first evaluate whether the text of the last sentence of the preset length characters of the paragraph where the abstract is located meets the format requirements, determine the first interception position according to the format evaluation result, determine the first interception position as the latest interception position, and then further evaluate The last sentence of the text before the first interception position, and the latest interception position is further updated according to the evaluation result. In this embodiment, the text before the latest interception position is evaluated multiple times. The latest interception position is updated and adjusted, and the latest interception position is updated multiple times, so that the semantics of the last sentence of the text in the abstract of the determined search result is more complete and more user-friendly.

Embodiment 2

FIG. 7 is another schematic flowchart of a method for determining a summary of a search result provided by an embodiment of the present application. The execution body of this embodiment may be a server. As shown in FIG. 7 , the following steps S701 to S708 may be used to determine the search result Summary.

Step S701: Acquire the text information of the search result, and determine the target paragraph from the text information.

Step S702: From the starting position of the target paragraph, determine the set length of characters.

The implementation process of step S701 and step S702 is similar to the implementation process of step S401 and step S402, and will not be repeated here.

Step S703: Determine the last sentence of text with the preset length of characters.

In this embodiment, the last sentence of the preset length characters may be determined in the following manner: the text after the last first punctuation mark in the preset length characters is determined as the last sentence of text. Specifically, the first punctuation mark may be a period mark, wherein the period mark is one of the punctuation marks, and the period mark is used to indicate pauses of different lengths in spoken language. In a specific embodiment, the above-mentioned first punctuation mark may be any one of the following dots: "," (comma), "?" (question mark), "!" (exclamation mark), ";" (semicolon) ,":"(colon). When any one of the above dots appears in the text, the text before the dot is usually a text with relatively complete semantics, so the first punctuation mark is any of the above dots, which can make the sentence division more reasonable.

FIG. 8 shows a plurality of examples of the last sentence of the above-mentioned preset length characters.

As shown in Figure 8, if the preset length characters determined from the starting position of the above-mentioned target paragraph are as shown in a in Figure 8, the last first punctuation mark is " after "We can watch the news", ”, therefore, the “understanding of domestic and foreign affairs” after this “,” is determined as the last sentence of text with a preset length of characters, and the last sentence of text in a in Figure 8 is framed by a box.

For another example, the preset length characters determined from the starting position of the above target paragraph are shown in b in Figure 8, and the last first punctuation mark is "," after "Therefore", therefore, this " ," after "briging the" is determined as the last sentence of text with a preset length of characters, and the last sentence of text in b in Figure 8 is framed by a box.

For another example, the preset length characters determined from the starting position of the above-mentioned target paragraph are as shown in c in Figure 8, and the last first punctuation mark is "can read gratifying, watch movies on the Internet". ,", therefore, the "medical, shopping" after this "," is determined as the last sentence of text with a preset length of characters. In c in Figure 8, the last punctuation mark is ",", which does not belong to the first sentence. Punctuation marks, therefore, do not use "," to determine the last sentence of text. In c in Figure 8, the last sentence of text is framed by a box.

Optionally, the server may also determine the last sentence of the above-mentioned preset length characters in the following manner: determine that the above-mentioned preset length characters end with a first punctuation mark, and use the second-to-last first punctuation mark in the preset length characters. The text after the symbol is determined to be the last sentence of text.

When the preset length characters end with the first punctuation mark, there is no character after the last first punctuation mark. At this time, if the text after the last first punctuation mark is determined as the last sentence of text, then determine The last sentence of text output is empty, that is, there is no last sentence of text. Therefore, the text after the penultimate first punctuation mark in the preset length characters can be determined as the last sentence of text, so as to obtain the last sentence of text.

As shown in FIG. 8 , if the preset length characters determined from the starting position of the above-mentioned target paragraph are as shown in d in FIG. 8 , it ends with ",", so the last first punctuation mark is the one at the end. ",", the second-to-last first punctuation mark is "." before "Therefore", in this case, the "Therefoer," after the second-to-last first punctuation mark "." can be determined as the default The last sentence of text with a length of characters, in d in Figure 8, the last sentence of text is framed by a box.

Step S704: Evaluate whether the end of the last sentence of text is incomplete topic information.

If the evaluation result of step S704 is yes, then step S705 is executed, and if the evaluation result of step S704 is no, then step S706 is executed.

The above topic information may include: topic name, URL, and may also include other types of topic information. Wherein, the subject name may include the title of film and television drama, character name, book title, musical work title, opera title, folk art performance title, etc., and may also include other subject titles.

Specifically, the following steps S1 to S4 may be used to determine whether the end of the last sentence of text is incomplete topic information.

Step S1: Perform word segmentation on the last sentence of text to obtain at least one word segmentation.

In this embodiment, a machine algorithm can be used to segment the last sentence of text, for example, a forward maximum matching word segmentation algorithm, a bidirectional maximum matching word segmentation algorithm, etc., or other algorithms can be used to segment the last sentence of text. This application does not limit the specific participle method.

In step S1, when the last sentence of text contains only one word or one phrase, a word segment is obtained, and when the last sentence of text contains multiple words or multiple phrases, multiple word segments are obtained.

FIG. 9 and FIG. 10 are schematic diagrams of word segmentation of text in this embodiment, respectively. For example, the text of the last sentence is "the specific process of injection molding is "plastic products", and the result of word segmentation of this text can be shown in Figure 9. For another example, the last sentence of text is "Technology makes our life faster and more convenient", and the result of word segmentation of this text can be shown in Figure 10.

Step S2: It is determined that there is topic information including the last participle of each participle in the subject information database.

Wherein, the above-mentioned subject information database includes a plurality of subject information. Specifically, the topic information base may include: each topic name, each uniform resource locator, and may also include other topic information. The subject information in the subject information base can be updated regularly or irregularly, so that the subject information in the subject information base is more comprehensive.

Specifically, the server may check whether the last word segment is included in the topic information database, and when the last word segment is found, it is determined that topic information including the last word segment in each word segment exists in the topic information database.

Step S3: It is determined that the above-mentioned last participle is incomplete compared with the topic information including the above-mentioned last participle.

For example, the determined last participle is "plastic products", and the subject information containing "plastic products" in the information base is "production method of plastic products", since "plastic products" is incomplete compared to "production methods of plastic products" , it can be determined that the last participle "plastic products" is incomplete compared to the subject information "production method of plastic products" that contains "plastic products".

Step S4: It is determined that the end of the text of the last sentence is incomplete topic information.

The above steps S1 to S4 can determine that the end of the last sentence of text is incomplete topic information.

In this application, other means may also be used to determine that the end of the last sentence of text is incomplete subject information. For example, when it is determined that the last sentence of text includes a punctuation mark that indicates the title of a work, if the last sentence includes the left half of the punctuation mark, but does not include the right half of the punctuation mark, then determine that the end of the last sentence of text is Incomplete subject information. For example, the text of the last sentence is "the specific process of injection molding is based on "plastic products". Since this sentence contains the left half of the book title "", but does not include the right half of """, it is determined that the end of this sentence is incomplete subject information.

Other methods may also be used to determine that the end of the last sentence of text is incomplete subject information, which is not specifically limited in this application.

FIG. 11 shows an example of determining an abstract of a search result. For example, as shown in Figure 11, the determined target paragraph (that is, the paragraph where the abstract is located) 11a is:

"Movies have enriched people's daily lives. This year is a big year for movies. June's "Growing Up Story" is a classic, and September's "From Now on" has also been well received. Many movies have left a deep impression on everyone and received continuous praise. "Running to the Sun" is a classic movie that didn't win many awards that year, but it left a very deep impression on people, and time will tell that the movie is really worth watching."

If the preset length is 75 characters, the preset length of characters 11b determined from the starting position of the target paragraph 11a is: "Movie enriches people's daily life, this year is a big year for movies, and the "Growing Story" in June It is a classic, and "From Now on" in September has also been well received. Many movies have left a deep impression on everyone. The well-received "Run to", the final text of the determined preset length of characters 11b is: "Run to" with constant praise, evaluate "Run to" with constant praise. After the evaluation, the evaluation result obtained is that the last sentence of the text "The end of "Run to" with constant praise is incomplete topic information.

Step S705: Complete the incomplete topic information at the end of the last sentence of text, and determine the end position of the text obtained after the topic information is completed as the first interception position.

Step S706: Evaluate whether the semantics of the last sentence of text are complete.

If the judgment result of step S706 is yes, then step S708 is executed, and if the judgment result of step S706 is no, then step S709 is executed.

Step S708: Determine the end position of the preset length characters as the first cutting position.

Step S709: Complete the last sentence of text, and determine the end position of the text obtained after the last sentence is completed as the first interception position.

After step S709 is executed, step S707 is executed.

For the specific implementation process of step S706 and step S709, reference may be made to step S1203 and step S1204 in the third embodiment, and step S412 and step S414 in the first embodiment, which will not be described in detail here.

Step S707: Determine the text in the target paragraph before the first cut position as the abstract of the search result.

For example, the last sentence of the text 11c obtained by supplementing the topic information is "Running to the Sun", which has been well received. The summary of the search result obtained by supplementing the topic information is shown in 11c in Figure 11. Alternatively, the last sentence of text could be supplemented with "Towards the Sun," which has been well-received.

This embodiment evaluates whether the subject information at the end of the last sentence of text with a preset length of characters is complete, so as to complete the subject information at the end. When the subject information is used, the user usually prefers to continue reading the subject information to better understand the text. In this embodiment, the subject information is supplemented completely, so that the user can read more complete information.

Embodiment 3

FIG. 12 is another schematic flowchart of a method for determining an abstract of a search result provided by an embodiment of the present application. As shown in FIG. 12 , the abstract of the search result may be determined according to the following steps S1201 to S1206 . The execution body of this embodiment may be a server.

Step S1201: Acquire the text information of the search result, and determine the target paragraph from the text information.

Step S1202: From the starting position of the target paragraph, determine the set length of characters.

The implementation process of step S1201 and step S1202 is similar to the implementation process of step S401 and step S402, and will not be repeated here.

Step S1203: Evaluate whether the semantics of the last sentence of text with the preset length of characters is complete.

If the evaluation result of step S1203 is negative, step S1204 is performed, and if the evaluation result of step S1203 is yes, step S1205 is performed.

In a specific embodiment, when it is determined that the last sentence of text lacks any one of a subject, a predicate and an object, it can be determined that the semantics of the last sentence of text is incomplete. For a specific example of evaluating whether the semantics is complete, reference may be made to step S412 of the first embodiment.

In this embodiment, the semantics of the last sentence of text can also be determined to be incomplete in other ways. For example, when the number of words in the last sentence of text is less than the third preset number, it is determined that the semantics of the last sentence of text is incomplete. For example, the third preset number may be any number from 2 to 6, or may be any other smaller number, wherein the words may include single words and phrases in Chinese, words in English, and may also include other languages words and phrases.

For example, as shown in FIG. 11 , evaluating the last sentence of the text of the determined preset length of characters 11b "running towards", which is continuously praised, the obtained evaluation result is that the last sentence of text "running towards" which is constantly praised is lacking. Predicate and object, semantically incomplete.

Step S1204: Complement the last sentence of text as a complete sentence of text, and determine the end position of the text obtained by adding a complete sentence of text as the first interception position.

Step S1205: Determine the end position of the preset length characters as the first interception position.

Step S1206: Determine whether the character length of the text before the first clipping position is not greater than the upper limit of the preset length range.

If the judgment result of step S1206 is yes, then step S1207 is executed, and if the judgment result of step S1206 is no, then step S1208 is executed.

Step S1207: Determine the text in the target paragraph before the first interception position as the abstract of the search result.

For example, as shown in Figure 11, the last sentence of text is supplemented as a complete text, and the last sentence of text becomes "Towards the Sun, which has been well received, is a classic movie", by adding the last sentence of text to a complete text to get A summary of the search results for is shown in 11d in Figure 11.

Step S1208: Delete the last sentence of text, and determine the end position of the text obtained after deleting the last sentence of text as the third interception position.

Step S1209: Determine the text in the target paragraph before the third clipping position as the abstract of the search result.

For the specific processes of step S1208 and step S1209, reference may be made to step S415, step S416, and step S413 in the first embodiment, and details are not described here.

When a user reads a text with incomplete semantics, the user usually prefers to obtain the complete content of the incomplete text to better understand the text. In this embodiment, the sentences are supplemented so that the user can read more complete information.

Below, the method for determining the abstract of the search result provided by the embodiment of the present application will be described in detail.

FIG. 14 is a schematic flowchart of a method for determining an abstract of a search result provided by an embodiment of the present application. As shown in FIG. 14 , the method for determining an abstract of a search result provided by this embodiment of the present application includes the following steps S1410 to S1480. The server in the embodiment shown in FIG. 14 may be the foregoing server 320 , and the client may be the foregoing client 310 .

Step S1410: The server receives the search query sent by the client.

The user can input a search query on the client 310, and after receiving the search query input by the user, the client 310 can send the search query to the server 320 through the network 330 (eg, wired and/or wireless network), and the server 320 can receive the search query. . For example, as shown in FIG. 2, the user can input the keyword "wooden floor" in the search box 201 of the search page of the browser, the keyword is the search query, and the search box 201 of the search page is the search engine's An interface for users to input information.

Step S1420: The server searches for at least one search result matching the above search query.

In this embodiment of the present application, the search results may be any machine-readable and storable documents, for example, the search results may be emails, news, blogs, business directories, electronic versions of printed texts, web pages, and the like. When users search for information on the Internet, the search results are usually web pages. The search results usually include text information, and may also include embedded information such as images, hyperlinks, audio, video, etc., and embedded instructions such as scripting languages.

In this embodiment of the present application, the server may search for at least one search result matching the above search query from the index database of the search engine. Wherein, the index database stores the index information of each webpage, and according to the index information, webpages that meet specific conditions can be quickly found. For example, the index information can include the URL of the webpage corresponding to each word segment and the keyword of the webpage, and can also include Other information about the webpage corresponding to each word segment. In the process of web page search, the server can store the URL of each web page, the hyper text markup language (HTML) code of the web page, the title of the web page and other web page information in the information database. The index information of each web page is established in the index database. The complete information of the webpage is stored in the information database. When the user views a webpage, the server can obtain the complete information of the webpage from the information database, and send the complete information of the webpage to the client, so that the client can display the complete information of the webpage. The complete content of the web page.

The server may look up search results after receiving the above search query. The search results may be web pages that match the search query entered by the user. The search result matches the search query. It can be that the title of the search result contains the keyword entered by the user, or the title of the search result contains a synonym of the keyword entered by the user, or the search result contains all or part of the text entered by the user, or the search result belongs to The field is the same or similar to the field to which the search query entered by the user belongs, and the search result may also be other cases that match the search query. For example, as shown in FIG. 2 , if the keyword input by the user during the search is "wooden floor", the server may determine the web page whose title contains "wooden floor" as the search result.

Step S1430: The server scores each of the found search results, and sorts each of the found search results in descending order of scores.

Among them, a higher score indicates a better match between the search result and the search query.

Specifically, when the search result is a web page, the server may score the searched web page according to at least one of the following conditions:

The number of occurrences of the search query in the webpage (e.g., the more occurrences, the higher the score);

where the search query appears in the web page (e.g. in the title with a higher rating);

The number of times the web page was viewed (for example, the more times the page was viewed, the higher the score), etc.

Step S1440: The server acquires the text information of each of the first preset number of search results in order from front to back, and determines the target paragraph from the acquired text information.

Wherein, for each search result, the target paragraph is the paragraph where the abstract of the search result is located.

The server determines the target paragraph from the acquired text information, that is, the server determines the target paragraph corresponding to the text information from each acquired text information.

The above-mentioned first preset number may be any number, for example, may be 10, 50, 100, 200, 500, etc., or may be other more or less number, the larger the first preset number is set, the server determines The more search results are output, the more search results are presented to the user, and the smaller the first preset number is, the faster the server determines the search results.

In the embodiment of the present application, taking the search result as a webpage as an example, when the server stores the complete information of each crawled webpage in the above-mentioned information database, the server can obtain the text information of the search result from the information database.

Optionally, for each of the first preset number of search results, the server may determine the target paragraph in any one of the following determination manners.

Determination method 1: The server segments the text information according to the carriage return in the text information of the search result, scores each paragraph, and selects at least one paragraph in descending order of scores to determine the target paragraph. It is understandable that the character length of some paragraphs after the paragraph is very short. In this case, a paragraph may not meet the requirements of abstract interception. Therefore, the target paragraph may include one paragraph or two or more. paragraph.

The target paragraph determined in this way has a high degree of matching with the user's search query.

Determination method 2: Segment the text information according to the carriage return in the text information of the search result, and score each paragraph. At least one paragraph adjacent to the paragraph is determined as the target paragraph.

The target paragraph determined in this way can not only have a high degree of matching with the user's search query, and when the target paragraph includes two or more paragraphs, the coherence of each paragraph included in the target paragraph can be better, so that the Semantic coherence and readability of summaries of identified search results are better.

The third determination method: the server segments the text information according to the carriage return in the text information of the search result, and determines the starting paragraph of the text information of the search result as the target paragraph. This method can quickly and easily determine the paragraph in which the abstract is located.

The server may also determine the paragraph in which the abstract is located in other ways, which is not specifically limited in this application.

In the above determination method 1 and determination method 2, the higher the score of the segment, the higher the matching degree of the segment and the search query. Therefore, the server can score each segment based on the following factors: the search query in the segment The number of occurrences (for example, the more occurrences, the higher the score), the position of the segment in the search results (for example, if the segment is the title or the first paragraph, the score is higher), the scoring method of the segment is not here. Repeat.

Step S1450: The server determines a preset length of characters from the starting position of the above-mentioned target paragraph.

The server determines the preset length of characters, which may be to determine the preset length of Chinese characters, or to determine the preset length of English characters, or to determine the preset length of Korean characters, or to determine the preset length of text in other languages, or to determine Preset lengths of characters in a mix of languages.

In an implementation manner, the text information of the search result may have the same preset lengths corresponding to different language types, such as the example in step S402 of the first embodiment, so that the applicability of the solution provided in this application can be improved. stronger. For different languages, the number of lines and spaces occupied by characters of the same length and the same font size in the web page are often different. Therefore, in another implementation manner, for the text information of the search results in different languages, pre- The length can also be different. Those skilled in the art can set the preset length according to the actual situation, which is not specifically limited here. The preset length may be, for example, any length from 100 to 200 characters, or may be a length of other numbers of characters. When the preset length is set longer, the number of characters in the determined summary is also larger, so that users can learn more about the search results; when the preset length is set shorter, the number of characters in the determined summary is also larger. Each search result occupies less space on the webpage when displayed on the display of the client, so that the same webpage can display more search results. Those skilled in the art can set the specific value of the preset length according to actual needs.

Step S1460: The server evaluates the determined last sentence of text with a preset length of characters to obtain a first evaluation result.

In this application, the last sentence of text with a preset length of characters may be determined according to step S703 of the second embodiment, and then the last sentence of text is evaluated.

In this embodiment of the present application, evaluating the last sentence of text with a preset length of characters determined from the starting position of the above-mentioned target paragraph may be, as in step S403 in the first embodiment, evaluating the number of words corresponding to the last sentence of text, or step S404 evaluates the character length corresponding to the last sentence of text. It can also be to evaluate whether the end of the last sentence of text is incomplete topic information as in step S704 in the second embodiment, or it can be to evaluate the semantics of the last sentence of text as in step S1203 in the third embodiment. completeness.

The obtained first evaluation result may be that the number of words corresponding to the last sentence of text in step S403 of the first embodiment is less than the second preset number (that is, the number of words is less), or it may be the last sentence in step S404 of the first embodiment. The character length corresponding to the text is less than the first preset length (that is, the character length is shorter), it can also be that the end of the last sentence of text in step S704 of the second embodiment is incomplete theme information, or it can be the step of the third embodiment The semantics of the last sentence text in S1203 is incomplete. The obtained first evaluation result may also be other results that can reflect the integrity of the text of the last sentence, which is not specifically limited in this application.

Step S1470: The server determines the first clipping position of the target paragraph according to the above-mentioned first evaluation result.

Wherein, the first interception position may be the end position of the text obtained by supplementing the last sentence of text of the target paragraph. The end position of the text obtained after the topic information is completely supplemented, and the first interception position may also be the end position of the text obtained after the last sentence of text is supplemented into a complete sentence in step S1204 of the third embodiment. The first interception position may also be the end position of the text obtained after deleting the last sentence of text of the target paragraph. For example, the first interception position may be the end position of the text obtained after deleting the last sentence of text in step S405 of the first embodiment. The first clipping position may also be the end position of the last sentence of the target paragraph. For example, the first clipping position may be the summary of the search result determined in step S708 of the second embodiment or step S1205 of the third embodiment.

In this embodiment of the present application, to supplement the last sentence of the target paragraph, as in step S1204 in the third embodiment, the last sentence of text may be supplemented into a complete sentence, or as in step S705 in the second embodiment, the last sentence may be added to the text. The subject information at the end of the text is complete. Specifically, as shown in 11d in FIG. 11 of the example in the third embodiment, to supplement the last sentence of text as a complete sentence of text, it may be to add the target paragraph before the next first punctuation mark of the above-mentioned last sentence of text is added to the end of this last sentence of text.

In one embodiment, after the first interception position of the target paragraph is determined, the text located before the first interception position can be adjusted, and a new new interception position can be re-determined, so that the text before the new interception position corresponds to the character length Within the preset length range, for example, a new cutting position can be re-determined according to steps S408 and S409 of the first embodiment. The preset length range may be 80 to 350 characters, and may also be other character length ranges. This application does not specifically limit the preset length range, and those skilled in the art can set it according to actual needs. In this implementation manner, if the character length corresponding to the text located before the first interception position is within the preset length range, the character length of the abstract of the determined search result can be kept within a reasonable length range, and the page length will not be too long. The number of displayed search results is too small, nor is it too short to cause the summary to contain too little valid information.

After the first clipping position of the target paragraph is determined, if the character length corresponding to the text in the target paragraph before the first clipping position is short, it is difficult for the user to know the general content of the search result through the abstract. When the summary of each search result is displayed on the client or other monitors, the layout space occupied by the summary of each search result is usually determined, and the display settings such as font size, line spacing, and word spacing of the summary of each search result are also OK, if the length of the text corresponding to the text in the paragraph where the abstract is located before the cut-off position is long, the determined abstract will not be fully displayed on the client or other monitors, resulting in a less complete abstract of the displayed search results. Poor, the display effect is not good. Therefore, in this application, the character length corresponding to the text located before the interception position in the target paragraph is set within the preset length range, so that the user can better understand the general content of the search result through the abstract, and the determined The summary is completely displayed on the client or other displays, so that the summary of the displayed search results is more complete and the display effect is better.

In this embodiment of the present application, when the above-mentioned first evaluation result is that the number of characters or words corresponding to the last sentence of text is small, such as when the evaluation results of steps S403 and S404 in Embodiment 1 are no, or the semantics of the last sentence of text is negative When it is incomplete, such as when the evaluation result of step S412 in the first embodiment is no, when the evaluation result of step S1203 in the third embodiment is no, or when the end of the last sentence of text is incomplete topic information, as in the first embodiment When the evaluation result of step S410 is yes, and when the evaluation result of step S704 of the second embodiment is yes, the above-mentioned last sentence of text can be deleted, as in step S405 of the first embodiment, and the above-mentioned last sentence of text can also be supplemented, as in the embodiment In step S705 of the second embodiment and step S1204 of the third embodiment, the end position of the text obtained after deleting the last sentence of text or the text obtained after supplementing is determined as the first interception position of the target paragraph. When the above-mentioned first evaluation result is that the number of characters or words corresponding to the last sentence of text is quite large, or the semantics of the last sentence of text is complete, or the end of the last sentence of text is not incomplete topic information, the above-mentioned last sentence of text may be The end position is determined as the first interception position of the target paragraph, as shown in step S708 of the second embodiment and step S1205 of the third embodiment. In this embodiment of the present application, the first interception position of the paragraph where the abstract is located can be flexibly determined according to the first evaluation result, and those skilled in the art can determine the location of the paragraph where the abstract is located based on the principles of better completeness and better display effect of the abstract in the search result. For the first interception position, the present application does not limit a specific manner of determining the first interception position.

Step S1480: The server determines the abstract of the search result from the target paragraph according to the above-mentioned first clipping position.

The character length corresponding to the abstract of the search result determined in step S1480 is within a preset length range.

Specifically, the server may determine the text in the target paragraph before the first clipping position as the abstract of the search result. For example, step S707 in the second embodiment and step S1207 in the third embodiment are examples of determining the text in the target paragraph before the first interception position as the abstract of the search result.

Alternatively, the server may also determine another clipping position according to the text before the first clipping position, and determine the text in the target paragraph before the other clipping position as the abstract of the search result. For example, steps S405 to S416 in the first embodiment are to determine another interception position according to the text before the first interception position, and determine the text in the target paragraph before another interception position as an example of the abstract of the search result, the first embodiment The latest interception position in is another determined interception position. Sometimes the text in the target paragraph before the first interception position still cannot present the abstract well, for example, the text is too long or too short, the text semantics is still incomplete, etc. In this case, the text before the first interception position can be Another clipping position is further determined, wherein the text before the other clipping position can better present the abstract, so that the determined abstract can be further presented in a better effect.

The solution provided in the embodiment of the present application evaluates the last sentence of text with a preset length of characters starting from the starting position of the target paragraph, and determines the first interception position according to the first evaluation result obtained from the evaluation. It can be seen that the implementation of the present application For example, according to the first evaluation result, the first interception position can be flexibly determined, so that the last sentence of the abstract of the determined search result can be a sentence of text with relatively complete semantics, which improves the reference value for whether the user browses the web page. When displayed on the client, the rendering effect is better, the amount of invalid reading by the user is reduced, the reading experience of the user is improved, and the probability of the user clicking on the webpage corresponding to the abstract is increased. In addition, the first interception position determined in the embodiment of the present application is the end position of the text obtained after supplementing the last sentence of text, the end position of the last sentence of text, or the end position of the text obtained after deleting the last sentence of text. According to the different first evaluation results, the first interception positions determined in the application embodiment are also different, and the determined first interception positions are further flexibly determined, so that the determined abstract has a better presentation effect.

In an embodiment, in step S1460, the last sentence of text can be evaluated in any one of the following evaluation manners to obtain the first evaluation result.

Evaluation method 1: Evaluate whether the character length corresponding to the last sentence of text is less than the first preset length.

For the specific evaluation process of the first evaluation method, reference may be made to step S404 in the first embodiment.

The evaluation result based on the first evaluation method is yes, and the first interception position may be: the end position of the text obtained after deleting the last sentence of text, as shown in steps S405 to S406 of the first embodiment.

Through the first evaluation method, it can be quickly evaluated that the semantic integrity of the last sentence of the text is low and the reference value for the user to understand the web page is small, so that when determining the first interception position, it can quickly determine the preset length of characters. The last sentence is removed to more quickly identify snippets that show better search results to users.

Evaluation method 2: Evaluate whether the number of words corresponding to the last sentence of text is less than the second preset number.

For the specific evaluation process of the second evaluation method, reference may be made to step S403 of the first embodiment.

The evaluation result based on evaluation method 2 is yes, and the first interception position may be: the end position of the text obtained after deleting the last sentence of text, such as steps S405 to S406 in the first embodiment.

Evaluation method 2 can also more quickly determine the summary of search results that are better displayed to the user, and because the number of words contained in a sentence of text usually better reflects the semantic integrity of a sentence of text, therefore, evaluate the words of the last sentence of text. Quantity, can more accurately evaluate the semantic integrity of the last sentence of the text and the reference value for users to understand the web page. Through the evaluation method 2, it can be quickly evaluated that the semantic integrity of the last sentence of the text is low and the reference value for the user to understand the web page is small, so that when determining the first interception position, it can quickly determine the preset length of characters. The last sentence is removed to more quickly identify snippets that show better search results to users.

Evaluation method 3: Evaluate whether the proportion of characters other than words and words in the last sentence of text is greater than the preset proportion.

The above-mentioned words may include single-character words and phrases in Chinese, English words, and may include single-character words and phrases in other languages, and the above-mentioned words may include Chinese characters. For text information, characters other than words and words may include: at least one of numbers, punctuation marks, mathematical symbols, currency symbols, collation symbols, lexicon symbols, and phonetic symbols, and characters other than words and words may also include other Symbols are not specifically limited in this application. The above preset ratio may be any ratio from 40% to 60%, or may be other larger ratios.

The evaluation result based on evaluation method 3 is yes, and the first interception position may be: the end position of the text obtained after deleting the last sentence of text.

Users usually understand abstracts more fully through words and phrases that have actual semantics. Characters other than words and words in abstracts are usually less helpful for users to understand abstract content. Therefore, words and words in texts are usually the ones that users understand. The more effective information of the abstract, the third evaluation method can evaluate that the proportion of effective information in the last sentence of text is small, and when the first interception position is determined, it can be quickly determined to delete the last sentence of text with a preset length of characters.

Evaluation method 4: Evaluate whether the end of the text in the last sentence above is incomplete topic information.

For the specific evaluation process of the fourth evaluation mode, reference may be made to step S704 of the second embodiment.

The evaluation result based on evaluation method 4 is that the first interception position may be the end position of the text obtained by supplementing the subject information at the end of the last sentence of text.

In the evaluation method 4, the last sentence of text is evaluated, that is, the subject information at the end of the last sentence of text is evaluated, and the first evaluation result obtained is that the end of the last sentence of text is incomplete subject information. When users are reading the text, they usually prefer to read the full content of the subject information such as movie and TV series titles, book titles, etc. to better understand the text. Evaluation method 4 can evaluate that the end of the last sentence is incomplete. For the subject information, when determining the first interception position, the first interception position can be determined according to the first evaluation result of the fourth evaluation method as a position that can complete the above subject information, so that the user can better understand the text.

Evaluation method five: evaluate whether the semantics of the last sentence above is incomplete.

For the specific evaluation process of the fifth evaluation mode, reference may be made to step S1203 of the third embodiment.

The evaluation result based on evaluation method 5 is yes, and the first interception position may be: the end position of the text obtained after the last sentence of text is supplemented into a complete sentence of text.

In the fifth evaluation method, the last sentence of text is evaluated, that is, the semantics of the last sentence of text is evaluated, and the first evaluation result obtained is that the semantics of the last sentence of text is incomplete. When the semantics of a sentence of text is incomplete, the user usually cannot know the specific content of the text after reading the text. Therefore, the evaluation method 5 evaluates that the semantics of the last sentence is incomplete. When determining the first interception position , the text of the last sentence can be completed according to the first evaluation result of evaluation method 5, which is convenient for users to read the text.

Evaluation method 6: Evaluate whether the end of the text of the last sentence above is a punctuation mark.

Punctuation marks in assessment method 6 can be dots.

The evaluation result based on evaluation method 6 is yes, and the first interception position may be: the end position of the text obtained by deleting the punctuation mark at the end of the text of the last sentence.

It can be seen from the above evaluation methods 1, 2, 3 and 6 that these four evaluation methods do not analyze the semantics of the last sentence, but directly determine the number of characters, words or words contained in the last sentence. Therefore, the evaluation method 1, evaluation method 2, evaluation method 3 and evaluation method 6 can be summarized as format evaluation.

In the embodiment of the present application, other evaluation methods may also be used to evaluate the last sentence of text with a preset length of characters, wherein the evaluation result can reflect the semantic integrity of the last sentence of text, and the specific evaluation method is not limited . The first interception position is determined according to the first evaluation result. If the first evaluation result is different, the determined first interception position may be the same or different.

In the above embodiment, when the length of characters corresponding to the last sentence of text is less than the first preset length, or when the number of words corresponding to the last sentence of text is less than the second preset number, or the proportion of characters other than words in the last sentence of text is greater than When the preset ratio is used, it means that the last sentence of text is likely to be a sentence with incomplete semantics, or a sentence that does not express any actual semantics. Therefore, the last sentence of text has little reference value for users to understand the search results. In this case, the last sentence can be deleted, and the sentence with less reference value can be deleted, so that the completeness of the determined abstract is better, and the user's reading experience is better when the abstract is displayed on the client. In addition, since the character length or the number of words of the last sentence is small, when the last sentence is directly deleted, the difference between the total character length of the obtained abstract and the above preset length is also small. Other monitors show less impact as well.

When it is determined that the length of characters corresponding to the last sentence of text is less than the first preset length, or when the number of words corresponding to the last sentence of text is less than the second preset number, the last sentence of text may also be supplemented as a complete sentence of text. In this way, the completeness of the determined digest can also be improved.

The above sentence of complete text can be understood as a sentence of text with complete semantics, for example, it can be a sentence of text including subject, location and object.

In one embodiment, when the first evaluation result is to determine that the length of characters corresponding to the last sentence of text is not less than the first preset length, or when the number of words corresponding to the last sentence of text is not less than the second preset number, or it is determined that the last sentence When the proportion of characters other than words in a sentence of text is not greater than the preset ratio, or when it is determined that the end of the last sentence of text is complete topic information, or when it is determined that the semantics of the last sentence of text are complete, the first interception position can be the above The end position of the last sentence of text is step S708 in the second embodiment, and step S1205 in the third embodiment. When the last sentence of text with a preset length of characters is evaluated, the result of the evaluation is that the semantics of the last sentence is complete. In this case, there is no need to delete or supplement the content of the last sentence. In this embodiment, the reason for directly determining the end position of the last sentence of text as the clipping position of the paragraph where the abstract is located is that the last sentence of text with a preset length of characters is already text with complete semantics and better display effect.

In one embodiment, the method for determining the summary of the search result may further include the following step: sending the determined summary of the search result to the client, so that the client displays the received summary on the display.

In an implementation manner, the first clipping position may be: the end position of the text obtained after deleting the last sentence of text. Step S1480 can be implemented by the following steps: determine that the length of the character corresponding to the text located before the first interception position in the target paragraph is less than the second preset length; determine the second interception position, wherein the second interception position is: for the last sentence The end position of the text obtained after the text is supplemented; according to the second interception position, the abstract of the search result is determined from the target paragraph. For a specific example of this implementation manner, reference may be made to step S408 , step S409 and step S413 of the first embodiment. The latest interception position in step 409 of the first embodiment is the above-mentioned second interception position.

The above-mentioned second preset length may be the lower limit of the above-mentioned preset length range. In this embodiment, when the first interception position is the end position of the text obtained after deleting the last sentence of text, if the character length corresponding to the text in the target paragraph before the first interception position is less than the second preset length, it means that the first The length of the characters corresponding to the text before the interception position is too small. In this case, the summary of the determined search result may contain very little information, so that the user cannot obtain the effective information of the search result from the summary. In this case, The summary of the search result is determined from the target paragraph according to the second interception position. Since the second interception position is the end position of the text obtained after supplementing the last sentence of text, the length of characters corresponding to the obtained summary of the search result can be increased, so that the user can You can learn more about the search results from the abstract.

The above-mentioned, according to the second interception position, determines the summary of the search result from the target paragraph, which can be the text before the second interception position in the target paragraph is determined as the summary of the search result, or it can be further based on the text before the second interception position. Other interception positions are determined, and an abstract of the search result is determined according to the other interception positions, until the presentation effect of the determined abstract of the search result meets the requirements.

In a specific embodiment, the above-mentioned determination of the abstract of the search result from the target paragraph according to the second interception position can be achieved by the following steps: determining the character corresponding to the text in the target paragraph before the second interception position If the length is greater than the third preset length, the determined preset length characters are determined as the abstract of the search result. The third preset length is greater than the second preset length, and the third preset length may be the upper limit of the preset length range. For example, the abstract of the search result may be determined according to steps S409, S415, S416, and S413 of the first embodiment.

In this implementation manner, if the character length corresponding to the text before the second interception position in the target paragraph is greater than the third preset length, it means that the character length corresponding to the text before the second interception position is too long. The number of search results is too small, or the determined summary cannot be displayed completely, which affects the typesetting of the page. In this case, the determined preset length characters can be determined as the summary of the search result, which is beneficial to the typesetting of the page. The character length corresponding to the determined abstract is within a preset length range. That is to say, in this embodiment, when the determined length of the digest is short, if further increasing the character length of the digest will cause the digest to be too long, the determined preset length characters are used as the final determined digest.

In an implementation manner, the first clipping position is the end position of the text obtained by supplementing the last sentence of text. Step S1480 can be implemented according to the following steps: determine that the length of the character corresponding to the text located before the first interception position in the target paragraph is greater than the third preset length; determine the third interception position, wherein the third interception position is: after deleting the last sentence of text The end position of the obtained text; the text in the target paragraph before the third interception position is determined as the abstract of the search result. For example, the abstract of the search result may be determined according to steps S1206, S1208, and S1209 of the third embodiment.

The third preset length may be the upper limit of the above-mentioned preset length range. In this embodiment, when the first interception position is the end position of the text obtained by supplementing the last sentence of text, if the character length corresponding to the text in the target paragraph before the first interception position is greater than the second preset length, it means that The character length corresponding to the text before the first interception position is too long. In this case, since the display space of each search result on the display page is usually determined, the summary of the determined search result may not be displayed completely. In this case, the abstract of the search result is determined from the target paragraph according to the third interception position, since the third interception position is the end position of the text obtained after deleting the last sentence of text, the character length corresponding to the abstract of the obtained search result can be shortened, Enables the determined search results to be fully displayed in the display page.

In one embodiment, step S1470 may be implemented in the following steps: determine that the first evaluation result is any of the following: the end of the last sentence of text is complete topic information, the semantics of the last sentence of text is complete, and the characters corresponding to the last sentence of text The length is not less than the first preset length, and the proportion of characters other than words and phrases in the last sentence of text is not greater than the preset proportion; the last sentence of text is evaluated to obtain the second evaluation result; according to the second evaluation result, the target paragraph is determined the first intercept position. The second evaluation result is different from the evaluation content corresponding to the first evaluation result. Specifically, the summary of the search result can be determined according to steps S706 to S709 of the second embodiment, wherein step S706 is the process of evaluating the last sentence of text to obtain the second evaluation result, and steps S708 and S709 are based on the second evaluation result. The process of determining the first intercept position.

When the first evaluation result is that the end of the last sentence of text is the complete topic information, the semantics of the last sentence of text is complete, the length of the characters corresponding to the last sentence of text is not less than the first preset length, or the last sentence of text except for words and words When the proportion of characters is not greater than the preset proportion, it means that when evaluating the last sentence of text, the last sentence meets the evaluation requirements, that is, the last sentence meets certain integrity requirements. In this case, in order to make the obtained The completeness of the abstract is better and the presentation effect of the abstract is better, and the text result of the last sentence can be further evaluated to obtain the second evaluation result. The evaluation content corresponding to the second evaluation result and the first evaluation result may be different. For example, when the evaluation content corresponding to the first evaluation result is whether the end of the last sentence of text contains incomplete topic information, the evaluation content corresponding to the second evaluation result The content can be whether the semantics of the last sentence is complete; when the evaluation content corresponding to the first evaluation result is the character length corresponding to the last sentence, the evaluation content corresponding to the second evaluation result can be whether the end of the last sentence contains incomplete topics. information.

In one embodiment, step S1480 may be implemented by the following steps: evaluating the last sentence of the text in the target paragraph before the first interception position to obtain a third evaluation result; Fourth interception position; according to the fourth interception position, determine the abstract of the search result from the target paragraph. For the specific implementation process of this embodiment, reference may be made to steps S408 to S416 in the first embodiment, and steps S1206 to S1209 in the third embodiment. In the first embodiment, the evaluation result obtained in step S408 is the third evaluation result, steps S409 to S411 determine the fourth interception position, and steps S412 to S416 determine the abstract of the search result according to the fourth interception position. In the third embodiment, the evaluation result obtained in step S1206 is the third evaluation result, the third interception position determined in step S1208 is the fourth interception position, and step S1209 is the process of determining the abstract of the search result according to the fourth interception position.

The above-mentioned fourth interception position can be any of the following: the end position of the text obtained after supplementing the last sentence of the text in the target paragraph before the first interception position, the position in the target paragraph before the first interception position. The end position of the text, the end position of the text obtained by deleting the last sentence of the text in the target paragraph before the first interception position.

In this implementation manner, after the first interception position is determined, the integrity of the last sentence of the text in the target paragraph before the first interception position may still be low. The last sentence of the text before an interception position is evaluated, a fourth interception position is determined according to the obtained third evaluation result, and an abstract of the search result is determined from the target paragraph according to the fourth interception position. That is to say, after each time the first interception position is determined, this embodiment will re-evaluate the last sentence of the text in the target paragraph before the first interception position, so that the text before the determined interception position is The text of the last sentence satisfies the requirement of complete display of the abstract, so that the completeness and display effect of the abstract of the determined search result can be improved.

In one embodiment, the server may determine the target paragraph of the search result as follows. FIG. 15 is a schematic flowchart of determining a target paragraph of a search result according to an embodiment of the present application. As shown in FIG. 15 , the target paragraph of the search result can be determined according to the following steps S1501 to S1506 .

S1501: The server receives the search query sent by the client.

S1502: The server searches for multiple web pages matching the search query, and scores each web page in descending order of matching degree.

Among them, the higher the matching degree, the higher the score.

S1503: The server determines X webpages from the above-mentioned multiple webpages according to the order of the scores from high to low.

X is a positive integer not less than 1. For example, X may be any value from 50 to 300, or may be another specific value. The larger the X setting is, the more web pages the server sends to the client, and the more web pages the client can display.

S1504: The server acquires the text information corresponding to the determined webpage, and segments the text information to obtain at least one paragraph corresponding to each of the X webpages.

S1505: The server scores each paragraph.

Among them, the higher the score, the better the match between the paragraph and the search query.

S1506: The server selects Y paragraphs in descending order of ratings as target paragraphs of the search result.

Usually, a webpage can select a paragraph as the paragraph where the abstract is located. In this case, the value of Y is 1. In other cases, if the text information of a segment of the web page is too small to provide enough information as a summary, in this case, the value of Y can be a positive integer greater than 1.

In one embodiment, step S1606 may be implemented in the following steps a to d:

Step a: Determine i=1.

Step b: The server determines whether the length of the text corresponding to the paragraphs ranked 1 to i is not less than a preset length.

If the judgment result of step b is yes, execute step c, and if the judgment result of step b is no, execute step d.

For the setting method of the preset length in step b, reference may be made to the setting method of the preset length in the foregoing embodiments, which will not be repeated here.

Step c: Determine the paragraphs ranked 1 to i as Y paragraphs.

Step d: determine i=i+1, and execute steps b to c.

Examples of abstract determination methods for search results provided by the present application are described in detail above. It can be understood that, in order to implement the above-mentioned functions, the corresponding apparatuses include corresponding hardware structures and/or software modules for performing the respective functions. Those skilled in the art should easily realize that the present application can be implemented in hardware or a combination of hardware and computer software with reference to the modules and algorithm steps of each example described in the embodiments disclosed herein. Whether a function is performed by hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of this application.

The present application can divide the functional modules of the apparatus for determining the abstract of the search result according to the above method examples. For example, each function can be divided into each functional module, or two or more functions can be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware, and can also be implemented in the form of software function modules. It should be noted that the division of modules in this application is schematic, and is only a logical function division, and other division methods may be used in actual implementation.

FIG. 16 shows a schematic structural diagram of an apparatus for determining an abstract of a search result provided by the present application. The apparatus includes an evaluation module 1610 and a determination module 1620.

In one embodiment, the first evaluation result is: the end of the last sentence of text is incomplete topic information;

In one embodiment, the first evaluation result is: the semantics of the last sentence of text is incomplete;

In an embodiment, the first evaluation result is: the length of the characters corresponding to the last sentence of text is less than the first preset length, or the proportion of characters other than words and words in the last sentence of text is greater than the predetermined length. set proportion;

In one embodiment, the first interception position is: the end position of the text obtained after deleting the last sentence of text;

The determining module is specifically used for:

It is determined that the character length corresponding to the text before the first interception position in the target paragraph is less than the lower limit of the preset length range;

In one embodiment, the first interception position is: the end position of the text obtained after the last sentence of text is supplemented;

The determining module is specifically used for:

In one embodiment, the determining module is specifically used for:

The embodiments of the present application also provide an electronic device. The electronic device provided in this embodiment of the present application may be the server 320 shown in FIG. 3 , and is configured to execute the foregoing method for determining the abstract of the search result. FIG. 17 is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As shown in FIG. 17 , the electronic device provided by the present application includes a processor 1701, an interface 1702, a memory 1703, and a communication bus 1704, wherein the processor 1701, The interface 1702 and the memory 1703 communicate with each other through the communication bus 1704;

The memory 1703 is used to store computer programs;

The processor 1701 is configured to implement the method for determining the digest of the search result described in any one of the foregoing embodiments when executing the program stored in the memory 1703 .

The communication bus mentioned in the above electronic device may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus or the like. The communication bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of presentation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

The interface is used for communication between the above electronic device and other devices.

The memory may include random access memory (RAM), and may also include non-volatile memory (NVM), such as at least one disk memory.

Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (digital signal processing, DSP), a dedicated integrated Circuit (application specific integrated circuit, ASIC), field-programmable gate array (field-programmable gate array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

Embodiments of the present application further provide a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the processor is made to execute the description in any of the foregoing embodiments. The method for determining the summary of the search results.

Embodiments of the present application also provide a computer program product, which, when the computer program product runs on a computer, causes the computer to execute the above-mentioned relevant steps, so as to implement the method for determining an abstract of a search result described in any of the above-mentioned embodiments.

Wherein, the electronic device, computer-readable storage medium, computer program product or chip provided in this embodiment are all used to execute the corresponding method provided above. Therefore, for the beneficial effects that can be achieved, reference may be made to the above-provided method. The beneficial effects in the corresponding method will not be repeated here.

In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are only illustrative, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components may be combined or May be integrated into another device, or some features may be omitted, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.

Units described as separate components may or may not be physically separated, and components shown as units may be one physical unit or multiple physical units, that is, may be located in one place, or may be distributed in multiple different places. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware, or may be implemented in the form of software functional units.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a readable storage medium. Based on such understanding, the technical solutions of the embodiments of the present application can be embodied in the form of software products in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, which are stored in a storage medium , including several instructions to make a device (which may be a single chip microcomputer, a chip, etc.) or a processor (processor) to execute all or part of the steps of the methods in the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read only memory (ROM), random access memory (random access memory, RAM), magnetic disk or optical disk and other media that can store program codes.

The above content is only a specific embodiment of the present application, but the protection scope of the present application is not limited to this. Covered within the scope of protection of this application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

A method for determining an abstract of a search result, comprising:

Determine a preset length of characters from the starting position of the target paragraph, wherein the target paragraph is the paragraph where the abstract of the search result is located;

Evaluate the last sentence of text of the preset length characters to obtain the first evaluation result;

According to the first evaluation result, the first interception position of the target paragraph is determined, wherein the first interception position is any one of the following: the end position of the text obtained by supplementing the last sentence of text, the Describe the end position of the last sentence of text, and the end position of the text obtained after deleting the last sentence of text;

According to the first interception position, an abstract of the search result is determined from the target paragraph, wherein a character length corresponding to the abstract of the search result is within a preset length range.
The method according to claim 1, wherein the first evaluation result is: the end of the last sentence of text is incomplete topic information;

The first interception position is: the end position of the text obtained by supplementing the subject information at the end of the last sentence of text.
The method according to claim 1, wherein the first evaluation result is: the semantics of the last sentence of text is incomplete;

The first interception position is: the end position of the text obtained after the last sentence of text is supplemented into a complete sentence of text.
The method according to claim 1, wherein the first evaluation result is: a character length corresponding to the last sentence of text is less than a first preset length, or, the characters of the last sentence of text other than words and words The proportion of characters is greater than the preset proportion;

The first interception position is: the end position of the text obtained after deleting the last sentence of text.
The method according to claim 1 or 4, wherein the first interception position is: the end position of the text obtained after deleting the last sentence of text;

The determining the abstract of the search result from the target paragraph according to the first interception position includes:

It is determined that the character length corresponding to the text located before the first interception position in the target paragraph is less than the lower limit of the preset length range;

Determine the second interception position, wherein, the second interception position is: the end position of the text obtained after the last sentence of text is supplemented;

A summary of the search result is determined from the target paragraph based on the second clipping position.
The method according to any one of claims 1 to 3, wherein the first interception position is: the end position of the text obtained after the last sentence of text is supplemented;

The determining the abstract of the search result from the target paragraph according to the first interception position includes:

It is determined that the character length corresponding to the text located before the first interception position in the target paragraph is greater than the upper limit of the preset length range;

Determine the third interception position, wherein, the third interception position is: the end position of the text obtained after deleting the last sentence of text;

A summary of the search result is determined from the target paragraph based on the third clipping position.
The method according to any one of claims 1 to 6, wherein the determining, according to the first evaluation result, the first interception position of the target paragraph comprises:

It is determined that the first evaluation result is any of the following: the end of the last sentence of text is complete topic information, the semantics of the last sentence of text is complete, and the character length corresponding to the last sentence of text is not less than the first preset. The length, the proportion of characters other than words and words in the last sentence of text is not greater than the preset proportion;

Evaluate the last sentence of text to obtain a second evaluation result, wherein the second evaluation result is different from the evaluation content corresponding to the first evaluation result;

According to the second evaluation result, a first interception position of the target paragraph is determined.
The method according to any one of claims 1 to 6, wherein the determining the abstract of the search result from the target paragraph according to the first interception position comprises:

Evaluate the last sentence of the text in the target paragraph before the first interception position to obtain a third evaluation result;

According to the third evaluation result, determine the fourth interception position of the target paragraph;

A summary of the search result is determined from the target paragraph based on the fourth clipping position.
A device for determining an abstract of a search result, comprising:

The evaluation module is used to determine the preset length characters from the starting position of the target paragraph, and evaluate the last sentence of the above preset length characters to obtain a first evaluation result, wherein the target paragraph is the search result. The paragraph in which the summary of the results is located;

A determination module, configured to determine a first interception position of the target paragraph according to the first evaluation result, wherein the first interception position is any one of the following: a text obtained by supplementing the last sentence of text The ending position of the text, the ending position of the last sentence of text, the ending position of the text obtained after deleting the last sentence of text, according to the first interception position, from the target paragraph to determine the summary of the search result, wherein , the character length corresponding to the abstract of the search result is within a preset length range.
The device according to claim 9, wherein the first evaluation result is: the end of the last sentence of text is incomplete topic information;

The first interception position is: the end position of the text obtained by supplementing the subject information at the end of the last sentence of text.
The device according to claim 9, wherein the first evaluation result is: the semantics of the last sentence of text is incomplete;

The first interception position is: the end position of the text obtained after the last sentence of text is supplemented into a complete sentence of text.
The device according to claim 9, wherein the first evaluation result is: a character length corresponding to the last sentence of text is less than a first preset length, or, the characters of the last sentence of text other than words and words The proportion of characters is greater than the preset proportion;

The first interception position is: the end position of the text obtained after deleting the last sentence of text.
The device according to claim 9 or 12, wherein the first interception position is: the end position of the text obtained after deleting the last sentence of text;

The determining module is specifically used for:

Determine that the character length corresponding to the text before the first interception position in the target paragraph is less than the lower limit of the preset length range;

Determine the second interception position, wherein, the second interception position is: the end position of the text obtained after the last sentence of text is supplemented;

A summary of the search result is determined from the target paragraph based on the second clipping position.
The device according to any one of claims 9 to 11, wherein the first interception position is: the end position of the text obtained after the last sentence of text is supplemented;

The determining module is specifically used for:

It is determined that the character length corresponding to the text located before the first interception position in the target paragraph is greater than the upper limit of the preset length range;

Determine the third interception position, wherein, the third interception position is: the end position of the text obtained after deleting the last sentence of text;

A summary of the search result is determined from the target paragraph based on the third clipping position.
The device according to any one of claims 9 to 14, wherein the determining module is specifically configured to:

It is determined that the first evaluation result is any of the following: the end of the last sentence of text is complete topic information, the semantics of the last sentence of text is complete, and the character length corresponding to the last sentence of text is not less than the first preset. The length, the proportion of characters other than words and words in the last sentence of text is not greater than the preset proportion;

Evaluate the last sentence of text to obtain a second evaluation result, wherein the second evaluation result is different from the evaluation content corresponding to the first evaluation result;

According to the second evaluation result, a first interception position of the target paragraph is determined.
The device according to any one of claims 9 to 14, wherein the determining module is specifically configured to:

Evaluate the last sentence of the text in the target paragraph before the first interception position to obtain a third evaluation result;

According to the third evaluation result, determine the fourth interception position of the target paragraph;

A summary of the search result is determined from the target paragraph based on the fourth clipping position.
An electronic device, comprising: a processor, a memory and an interface;

The processor, the memory and the interface cooperate with each other for carrying out the method of any one of claims 1 to 8.
A computer-readable storage medium, characterized in that, a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the processor is made to execute any one of claims 1 to 8. method described.