WO2021027162A1 - Non-full-cell table content extraction method and apparatus, and terminal device - Google Patents

Abstract

A non-full-cell table content extraction method and apparatus, and a terminal device. The method comprises: obtaining an original table, wherein the original table is a non-full-cell table (S101); reading the original table to obtain a full-cell table, and determining all the cells belonging to the same merged cell (S102); sequentially obtaining the data of each cell from the full-cell table, determining whether the cell with the data being not empty is a title cell or a content cell, and marking the title cell and the content cell (S103); and if the merged cell comprises the title cell, filling each cell of the merged cell with the data of the title cell (S104). By means of the method, the non-full-cell table is regularized into the full-cell table, thereby realizing the accurate reading of the data of each cell, and marking the title cell and the content cell.

Description

Method, device and terminal equipment for extracting content of non-full grid table

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on August 9, 2019, the application number is 201910744823.8, and the title of the invention is "a method, device and terminal equipment for extracting the content of a non-full form form", and its entire content Incorporated in this application by reference.

Technical field

This application belongs to the field of computer technology, and in particular relates to a method, device and terminal device for extracting the content of a non-full grid table.

Background technique

The current table data analysis, for a typical table, that is, a full grid table, the content of each cell in the table can be read out separately. But for tables with merged cells (merged cells can be merged rows, merged columns, and include merged rows and merged columns at the same time), that is, non-full grid tables, although full grid tables can be obtained by reading non-full grid tables, However, the content of each cell in the full grid table cannot be successfully read, because when there is content in the merged cell that has undergone the merge operation, the read results except for the upper left or upper right corner of the merged cell (depending on The content in the cell in the current view direction is retained, and the data in the remaining cells of the merged cell is empty, which is deleted.

In addition, when the form of the table cannot be predicted, it is impossible to determine which cell is the title of the table and which cell is the content of the table, and data cannot be combed accurately, resulting in low accuracy of the established database.

Summary of the invention

technical problem

In view of this, the embodiments of the present application provide a method, device, and terminal device for extracting the content of a non-full grid table, so as to solve the problem that the existing content of each cell of a non-full grid table cannot be read completely and cannot be established with high accuracy. Technical issues of the database.

The solution to the problem

Technical solutions

The first aspect of the embodiments of the present application provides a method for extracting the content of a non-full grid table, including:

Obtaining an original form, the original form being a non-full grid form;

Read the original table to obtain a full table, and determine each cell belonging to the same merged cell;

Obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and mark the header cell and the content cell;

If the merged cell includes a header cell, each cell belonging to the merged cell is filled with the data of the header cell.

The second aspect of the embodiments of the present application provides a device for extracting the content of a non-full grid table, including:

An obtaining module for obtaining an original form, the original form being a non-full grid form;

The reading module is used to read the original table to obtain a full table, and determine each cell belonging to the same merged cell;

The determination module is used to obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to the header cell or the content cell, and compare the header cell and the content cell mark;

The filling module is configured to, if the merged cell includes a header cell, fill each cell belonging to the merged cell with the data of the header cell.

A third aspect of the embodiments of the present application provides a terminal device, including a memory and a processor. The memory stores computer-readable instructions that can run on the processor. The processor executes the computer When reading instructions, the following steps are implemented:

Obtaining an original form, the original form being a non-full grid form;

Read the original table to obtain a full table, and determine each cell belonging to the same merged cell;

Obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and mark the header cell and the content cell;

If the merged cell includes a header cell, each cell belonging to the merged cell is filled with the data of the header cell.

The fourth aspect of the embodiments of the present application provides a computer-readable storage medium that stores computer-readable instructions, and the computer-readable instructions implement the following steps when executed by a processor:

Obtaining an original form, the original form being a non-full grid form;

Read the original table to obtain a full table, and determine each cell belonging to the same merged cell;

Obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and mark the header cell and the content cell;

If the merged cell includes a header cell, each cell belonging to the merged cell is filled with the data of the header cell.

The beneficial effects of the invention

Beneficial effect

Compared with the prior art, the embodiment of the application has the beneficial effect that: through the embodiment of the application, the non-full grid table is regularized into a full grid table, so that the data of each cell can be accurately read, and The title cell and content cell are marked, which improves the accuracy of subsequent data sorting and lays a solid foundation for helping to build a more accurate database.

Brief description of the drawings

Description of the drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only of the present application. For some embodiments, for those of ordinary skill in the art, other drawings can be obtained from these drawings without creative labor.

FIG. 1 is a schematic flowchart of a method for extracting the content of a non-full grid table in an embodiment of the application;

Figure 2 is a schematic flowchart of another method for extracting the content of a non-full grid table in an embodiment of the application

3 is a schematic flowchart of another method for extracting the content of a non-full grid table in an embodiment of the application;

4 is a schematic flowchart of another method for extracting the content of a non-full grid table in an embodiment of the application;

5 is a schematic structural diagram of a device for extracting the content of a non-full grid table in an embodiment of the application;

Fig. 6 is a schematic block diagram of a terminal device in an embodiment of the application.

Invention embodiment

Embodiments of the invention

In order to illustrate the technical solutions described in the present application, specific embodiments are used for description below.

It should be noted that the embodiments in this application and the features in the embodiments can be combined with each other if there is no conflict. Hereinafter, the present application will be described in detail with reference to the drawings and in conjunction with embodiments.

In order to enable those skilled in the art to better understand the solution of the application, the technical solutions in the embodiments of the application will be described clearly and completely in conjunction with the drawings in the embodiments of the application. Obviously, the described embodiments are only It is a part of the embodiments of this application, not all the embodiments. Based on the embodiments in this application, for those of ordinary skill in the art, all other embodiments obtained without creative labor should fall within the protection scope of this application.

In the following description, for the purpose of illustration rather than limitation, specific details such as a specific system structure and technology are proposed for a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from obstructing the description of this application.

It should be noted that the descriptions involving “first” or “second” in the terms in the specification, claims, and drawings of this application are only used to distinguish similar objects, and cannot be understood as indicating or implying them. The relative importance or implicitly indicates the number of the indicated technical features, that is, these descriptions need not be used to describe a specific order or sequence. In addition, it should be understood that these descriptions can be interchanged under appropriate circumstances in order to describe the embodiments of the present application.

FIG. 1 shows the implementation process of a method for extracting the content of a non-full grid table provided by an embodiment of the present application. The process of the method includes steps S101 to S104. This method is suitable for situations that need to read data from non-full grid tables. The method is executed by a non-full grid form content extraction device, which is configured in a terminal device and can be implemented by software and/or hardware. The specific implementation principle of each step is as follows.

S101. Obtain an original table, where the original table is a non-full grid table.

Among them, the terminal device obtains the original table, which is a non-full grid table, and then performs data extraction on the non-full grid table to complete the process of regularizing the non-full grid table into a full grid table.

It should be noted that a non-full grid table is a table that includes merged cells. The merged cell can be a merged row, and/or parallel, and/or a merged cell that includes both merged rows and merged columns. Non-full grid table is a concept opposite to full grid table. A full grid table is a table that does not include merged cells.

In the embodiment of the present application, because only the non-full grid table is processed, it is only required that the original table is a non-full grid table, and the style and content of the non-full grid table are not limited.

Illustratively, the original table is as shown in Table 1 below:

Table 1:

S102: Read the original table to obtain a full table, and determine each cell belonging to the same merged cell.

Among them, POI technology is used to read the original table to obtain a full grid table, and each cell belonging to the same merged cell is determined. That is to say, the full grid table is obtained by reading the non-full grid table through the POI technology, and each cell belonging to the same merged cell in the full grid table is determined. It should be noted that if there is data in the merged cell in the non-full grid table before reading, then in the read result, the data in the upper left or upper right corner of the corresponding merged cell is retained, and the data in the remaining cells The data is empty.

In this embodiment, the POI is Apache POI, which is an open source library of the Apache Software Foundation. The POI provides APIs for Java programs to read and write Microsoft Office format files.

When using the POI technology to read the original table, determine whether the cell belongs to the merged cell and determine the cells included in the merged cell. For example, through getMergedRegion() to obtain the merged area, that is, merged cells, through getFirstRow(), getLastRow(), getFirstColumn(), and getLastColumn() to obtain the starting row, ending row, and starting column of the merged cell, respectively. And the end column, which determines the firstRow, lastRow, firstCol, and lastCol of each merged cell.

It is understandable that by reading the original table through the POI technology, the number of merged cells can also be obtained, for example, the number of merged cells can be obtained by getNumMergedRegions(). This application does not exclude these situations.

Exemplarily, continuing the example of step 101, using POI technology to read the original table, as shown in Table 1, and the corresponding full grid table is obtained as shown in Table 2 below:

Table 2:

[Table 1]

Serial number	Shareholder name	Before the merger	null	After the merger	null
null	null	Contribution	Equity ratio	Contribution	Equity ratio
1	XXX1	800	33.34%	817	33.34%
2	XXX2	800	33.33%	816.5	33.33%
3	XXX3	800	33.33%	816.5	33.33%
null	total	2400	100%	2450	100%

Among them, the cell with "null" read result in the full grid table has empty data.

Using POI technology to read the original table to obtain a full grid table, it will be determined that the cell in the first row and the first column and the cell in the second row and the first column belong to the merged cell of the same merged column; the first row and second column The cell in the second row and the second column belong to the merged cell of the same merged column; the cell in the first row, third column and the first row and fourth column belong to the merged cell of the same merge row Cell; the cell in the first row and fifth column and the cell in the first row and sixth column belong to the merged cell of the same merged row.

S103: Obtain the data of each cell from the full grid table in sequence, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and mark the header cell and the content cell.

Among them, after step 102, the non-full grid table is read as a full grid table, and the data of each cell is read, even if it includes cells with empty data. After that, determine whether the cell whose data is not empty belongs to the header cell or the content cell, and mark the header cell and the content cell.

In the embodiments of the present application, by marking the cell as a title cell or a content cell, the title cell and the content cell are distinguished, which provides an accurate data basis for subsequent data combing, and improves table-based data To establish the accuracy of the database.

Optionally, as an embodiment of the present application, the process of determining whether a cell whose data is not empty in step 103 belongs to a title cell or a content cell, as shown in FIG. 2, includes the following steps 201 to 203. S201: Perform word segmentation processing on cell data to obtain several keywords to be matched, and filter out similar keywords that are similar to the keywords to be matched in a preset title keyword library.

Wherein, the title keyword database is a thesaurus composed of keywords extracted in advance from the title cells of multiple tables.

In the embodiment of the present application, the TF-IDF algorithm can be used to extract the title keyword database of a table from multiple, for example, 100 tables to be extracted.

Exemplarily, the TF-IDF algorithm is used to extract the preset title keyword database from multiple tables as a thesaurus composed of keywords such as amount, serial number, establishment, time, registered capital, and quantity.

The cell data is subjected to word segmentation processing to obtain a number of keywords to be matched, and similar keywords that are similar to the keywords to be matched in the preset title keyword library are filtered out. Among them, the embodiments of this application do not limit the specific word segmentation processing methods, and all existing word segmentation processing methods can be used to implement this application.

Exemplarily, when the data in a non-empty cell is "total amount", use the THULAC tokenizer for word segmentation to obtain two keywords "total" and "amount", and filter out the two keywords that exist in the title Similar keywords in the keyword database are "amount", and different keywords are "total".

Optionally, as an embodiment of the present application, in S201, filtering out similar keywords that are similar to the keywords to be matched that exist in a preset title keyword database, including:

Obtain unmatched target keywords from the title keyword database, match the target keywords with the keywords to be matched obtained after word segmentation processing to obtain the matching degree, continue to obtain the next unmatched target keywords and The target keyword is matched with the keyword to be matched to obtain the matching degree, until there is no unmatched target keyword in the title keyword library, and the target keyword with the highest matching degree and meeting the preset condition is regarded as the similar keyword.

Among them, for each keyword to be matched, each keyword in the title keyword library is matched with the keyword to be matched to obtain a matching degree, and the target keyword with the highest matching degree and meeting preset conditions is taken as the similarity key word. The preset condition may be that the matching degree is greater than or equal to a preset threshold. Those skilled in the art know that the preset threshold is an empirical value, which can be set according to requirements. For example, if the similarity is used to characterize the matching degree, the preset threshold can be 0.8 or 0.9, or it can be any value in the range of the two, and it can also be a number greater than 0.9 and less than or equal to 1. The description is only exemplary, and this application does not specifically limit this.

Optionally, as shown in FIG. 3, matching the target keyword with the keyword to be matched obtained after word segmentation processing to obtain the matching degree includes the following steps 301 to 302.

S301: Pre-make a dictionary containing a large number of characters as a preset character set, and each character in the character set is represented by an N-dimensional vector representing its position in the character set, where N is the number of words in the character set.

Among them, each character in this character set is arranged in sequence and has a corresponding arrangement position in the character set. The embodiment of this application collects a large number of articles in advance, and counts the words contained in these articles, calculates the appearance frequency corresponding to each word, and finally stores the words contained in these articles in the descending order of appearance frequency In this dictionary, the word set in the embodiment of this application is generated. Illustratively, assuming that the character set contains 6 characters, it should be noted that there are far more than 6 characters in the character set that are actually used. This is only an exemplary description, and the character set in the first place Words are represented by the vector [1, 0, 0, 0, 0, 0], the second-ranked word is represented by the vector [0, 1, 0, 0, 0, 0], and so on, the words in the set Each word is represented by a six-dimensional vector.

S302: Split the target keyword and the keyword to be matched into several words, and combine to form the target keyword vector corresponding to the target keyword and the key to be matched of the keyword to be matched by searching the N-dimensional vector of each word The word vector calculates the vector similarity between the target keyword vector and the keyword vector to be matched, and uses the vector similarity as the matching degree between the target keyword and the keyword to be matched.

Understandably, after the word set is established, each target keyword is divided into target keywords, and the N-dimensional vector of each target keyword is searched to form the target keyword vector; the same processing method , Split each keyword to be matched into keywords to be matched, and combine to form a keyword vector to be matched by searching the N-dimensional vector of each keyword to be matched. Then, the similarity between the target keyword vector and the keyword vector to be matched is calculated, and the similarity is taken as the degree of matching between the target keyword and the keyword to be matched.

Optionally, the vector similarity calculation formula can be:

similarity=∑ _i=1 ⁿ (A _i ×B _i )/{[∑ _i=1 ⁿ (A _i ) ² ] ^1/2 ×[∑ _i=1 ⁿ (B _i ) ² ] ^1/2 }.

Among them, the target keyword vector is A, and the element composition is _Ai ; the keyword vector to be matched is B, and the element composition is B _i . In addition, in other embodiments of the present application, the vector similarity calculation method may also use the similarity calculation method in the prior art, which is not specifically limited in this application.

Exemplarily, if the target keyword is split into two target keywords, the N-dimensional vectors of the two keywords are [1, 0, 0, 0, 0, 0] and [0, 1, 0, 0]. ,0,0], the combination of the target keyword vector is [1,1,0,0,0,0]; if the keyword to be matched is split into two keywords to be matched, find the two keywords to be matched The N-dimensional vectors are [0, 1, 0, 0, 0, 0] and [0, 0, 1, 0, 0, 0], and the combination to form the keyword vector to be matched is [0, 1, 1, 0, 0,0]. Calculate the vector similarity between [1, 1, 0, 0, 0, 0] and [0, 1, 1, 0, 0, 0] as 0.5.

S202: Calculate the similarity between a first keyword set composed of a number of keywords to be matched with respect to a second keyword set composed of similar keywords.

Wherein, as shown in FIG. 4, step 202 includes the following steps 401 to 403.

In S401, a number of keywords to be matched are formed into a first keyword set; similar keywords are formed into a second keyword set; and a number of keywords to be matched and similar keywords are formed into a target set.

S402: Calculate a first word frequency vector of the first keyword set relative to the target set, and calculate a second word frequency vector of the second keyword set relative to the target set.

S403: Calculate the similarity between the first word frequency vector and the second word frequency vector as the similarity between the first keyword set and the second keyword set.

Illustratively, the first keyword set {total, amount} composed of several keywords to be matched, and the second keyword set composed of similar keywords is {amount}; the keywords to be matched {total, amount} and similar keys The target set of words {amount} is {total, amount};

Calculate the first word frequency vector of the first keyword set relative to the target set as A=[1,1], and calculate the second word frequency vector of the second keyword set relative to the target set as B=[0,1]; calculate two The similarity between word frequency vectors A and B is 0.707, as follows:

similarity=∑ _i=1 ⁿ (A _i ×B _i )/{[∑ _i=1 ⁿ (A _i ) ² ] ^1/2 ×[∑ _i=1 ⁿ (B _i ) ² ] ^1/2 }=1 /(2 ^1/2 )=0.707.

S203: If the similarity is greater than the preset threshold, the cell is a title cell; otherwise, the cell is a content cell.

It can be understood that the preset threshold in step 203 is an empirical value, which can be set according to requirements. For example, the preset threshold value can be 0.8 or 0.9, or it can be any value in the range of the two, it can also be a number smaller than 0.8, or a number larger than 0.9 and less than or equal to 1. This is only an exemplary description, and this application does not specifically limit this.

The embodiment of the application provides a quantitative way to accurately distinguish whether a cell is a title cell or a content cell, which provides an accurate data basis for subsequent database establishment.

S104: If the merged cell includes a header cell, all cells belonging to the merged cell are filled with data of the header cell.

Among them, it is determined in step 102 which cells belong to the same merged cell, and in step 103, it is determined whether each cell belongs to a header cell or a content cell. Therefore, in step 104, when the merged cell includes a header cell In the case of a cell, each cell belonging to the merged cell is filled with the data of the header cell. In other words, for a cell with non-empty data in a merged cell, if the cell is a header cell, fill the header cell with data that belongs to the same merged cell and the data is empty In the cell.

It should be noted that in the embodiments of this application, only the data of the merged cell including the header cell is filled, that is, when the merged cell includes the header cell, the header cell is filled until the merged cell includes In the remaining cells. Generally, the header cell and the content cell do not belong to the same merged cell, and multiple content cells do not belong to the same merged cell. Therefore, in this embodiment of the application, the merged cell only includes the content unit In the case of grids, the data in the remaining cells are not filled because it may be caused by data upload errors, so as to avoid further filling of data, which will cause greater deviations in subsequent database creation.

Illustratively, step 103 determines that "serial number", "shareholder name", "before merger" and "after merger" are header cells, and it is determined that "serial number" and "name of shareholder" belong to the merger unit of the merger bank Grid; "Before absorption and merge" and "After absorption and merge" belong to the merged cells of the merged column. Fill each cell after the split with the corresponding header cell, as follows:

[Table 2]

Serial number	Shareholder name	Before the merger	Before the merger	After the merger	After the merger
Serial number	Shareholder name	Contribution	Equity ratio	Contribution	Equity ratio
1	XXX1	800	33.34%	817	33.34%
2	XXX2	800	33.33%	816.5	33.33%
3	XXX3	800	33.33%	816.5	33.33%
To	total	2400	100%	2450	100%

So far, the process of converting the merged table into a full grid table is completed.

In summary, through the embodiments of this application, the non-full grid table is normalized into a full grid table, so that the data of each cell can be accurately read, and the title cell and content cell are marked, which improves The accuracy rate of subsequent data sorting has been improved, which lays a solid foundation for helping to establish a more accurate database.

It should be understood that the size of the sequence number of each step in the foregoing embodiment does not mean the order of execution. The execution sequence of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiment of the present application.

Corresponding to the method for extracting the content of a non-full grid table described in the above embodiment, FIG. 5 shows a schematic structural diagram of a device for extracting the content of a non-full grid table provided by an embodiment of the present application. As shown in FIG. 5, The device for extracting the content of the non-full grid table includes:

The obtaining module 51 is configured to obtain an original form, and the original form is a non-full grid form;

The reading module 52 is configured to read the original table to obtain a full grid table and determine each cell belonging to the same merged cell;

The judging module 53 is used to sequentially obtain the data of each cell from the full grid table, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and compare the header cell and the content cell Mark

The filling module 54 is configured to, if the merged cell includes a header cell, fill each cell belonging to the merged cell with the data of the header cell.

Optionally, whether the cell for determining whether the data is not empty belongs to a title cell or a content cell, including:

Perform word segmentation processing on the cell data to obtain several keywords to be matched, and filter out similar keywords that are similar to the keywords to be matched in the preset title keyword library;

Calculate the similarity between the first keyword set composed of several keywords to be matched with respect to the second keyword set composed of similar keywords;

If the similarity is greater than the preset threshold, the cell is a title cell; otherwise, the cell is a content cell.

Optionally, the calculating the similarity of the first keyword set composed of several keywords to be matched with respect to the second keyword set composed of similar keywords includes:

Group several keywords to be matched into a first keyword set; Group similar keywords into a second keyword set; Group several keywords to be matched and similar keywords into a target set;

Calculating a first word frequency vector of the first keyword set relative to the target set, and calculating a second word frequency vector of the second keyword set relative to the target set;

The similarity between the first word frequency vector and the second word frequency vector is calculated as the similarity between the first keyword set and the second keyword set.

Optionally, the filtering out similar keywords existing in a preset title keyword library that are similar to the keywords to be matched includes:

Obtain unmatched target keywords from the title keyword database, match the target keywords with the keywords to be matched obtained after word segmentation processing to obtain the matching degree, continue to obtain the next unmatched target keywords and The target keyword is matched with the keyword to be matched to obtain the matching degree, until there is no unmatched target keyword in the title keyword library, and the target keyword with the highest matching degree and meeting the preset condition is regarded as the similar keyword.

Optionally, the matching the target keyword with the keyword to be matched obtained after word segmentation processing to obtain the matching degree includes:

Pre-create a dictionary containing a large number of characters as a preset character set. Each character in the character set is represented by an N-dimensional vector representing its position in the character set, where N is the number of words in the character set;

Split the target keyword and the keyword to be matched into several words, and by looking up the N-dimensional vector of each word, combine to form the target keyword vector corresponding to the target keyword and the keyword vector to be matched to the keyword to be matched Calculate the vector similarity between the target keyword vector and the keyword vector to be matched, and use the vector similarity as the matching degree between the target keyword and the keyword to be matched.

Those skilled in the art can clearly understand that, for convenience and concise description, the specific working process of the terminal device, module and unit described above can refer to the corresponding process in the foregoing method embodiment, and will not be repeated here.

In the foregoing embodiments, the description of each embodiment has its own focus. For parts that are not detailed or recorded in a certain embodiment, reference may be made to related descriptions of other embodiments.

FIG. 6 shows a schematic block diagram of a terminal device provided by an embodiment of the present application. For ease of description, only parts related to the embodiment of the present application are shown.

In this embodiment, the terminal device 6 may be a local terminal device or a cloud terminal device. The terminal device 6 may include a processor 60, a memory 61, and computer-readable instructions 62 that are stored in the memory 61 and can run on the processor 60. When the processor 60 executes the computer-readable instructions 62, the steps of the above embodiments of the method for extracting the content of a non-full grid table are implemented, or the processor 60 implements the above virtual devices when the computer-readable instructions 62 are executed The functions of each module/unit in the embodiment are, for example, the functions of the modules 501 to 504 shown in FIG. 5.

Exemplarily, the computer-readable instructions 62 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 61 and executed by the processor 60, To complete this application. The one or more modules/units may be a series of computer-readable instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer-readable instructions 62 in the terminal device 6.

The processor 60 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.

The memory 61 may be an internal storage unit of the terminal device 6, such as a hard disk or memory of the terminal device 6. The memory 61 may also be an external storage device of the terminal device 6, such as a plug-in hard disk equipped on the terminal device 6, a smart memory card (Smart Media Card, SMC), or a Secure Digital (SD) Card, Flash Card, etc. Further, the memory 61 may also include both an internal storage unit of the terminal device 6 and an external storage device. The memory 61 is used to store the computer-readable instructions and other instructions and data required by the terminal device 6. The memory 61 can also be used to temporarily store data that has been output or will be output.

A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing relevant hardware through computer-readable instructions, which can be stored in a non-volatile computer. In a readable storage medium, when the computer-readable instructions are executed, they may include the processes of the above-mentioned method embodiments. Wherein, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and/or volatile memory. Non-volatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not a limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Those skilled in the art can clearly understand that for the convenience and conciseness of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned functions can be allocated to different functional units and modules as required. Module completion, that is, divide the internal structure of the device into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist alone physically, or two or more units can be integrated into one unit. The above-mentioned integrated units can be hardware-based Formal realization can also be realized in the form of software functional units. In addition, the specific names of the functional units and modules are only used to facilitate distinguishing each other, and are not used to limit the protection scope of the present application. For the specific working process of the units and modules in the foregoing system, reference may be made to the corresponding process in the foregoing method embodiment, which will not be repeated here.

In the foregoing embodiments, the description of each embodiment has its own focus. For parts that are not detailed or recorded in a certain embodiment, reference may be made to related descriptions of other embodiments.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.

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

If the integrated module/unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, this application implements all or part of the procedures in the above-mentioned embodiments and methods, and can also be completed by instructing relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. Here, when the computer-readable instruction is executed by the processor, the steps of the foregoing method embodiments can be implemented.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still implement the foregoing The technical solutions recorded in the examples are modified, or some of the technical features are equivalently replaced; these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the application, and should be included in Within the scope of protection of this application.

Claims (20)

1. A method for extracting the content of a non-full grid table, which is characterized in that it comprises:

   Obtaining an original form, the original form being a non-full grid form;

   Read the original table to obtain a full table, and determine each cell belonging to the same merged cell;

   Obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and mark the header cell and the content cell;

   If the merged cell includes a header cell, each cell belonging to the merged cell is filled with the data of the header cell.
2. The method for extracting the content of a non-full grid table according to claim 1, wherein the determining whether the cell whose data is not empty belongs to a title cell or a content cell includes:

   Perform word segmentation processing on the cell data to obtain several keywords to be matched, and filter out similar keywords that are similar to the keywords to be matched in the preset title keyword library;

   Calculate the similarity between the first keyword set composed of several keywords to be matched with respect to the second keyword set composed of similar keywords;

   If the similarity is greater than the preset threshold, the cell is a title cell; otherwise, the cell is a content cell.
3. The method for extracting the content of a non-full grid table according to claim 2, wherein the calculation of the first keyword set consisting of a number of keywords to be matched is relative to the second keyword set consisting of similar keywords Similarity, including:

   Group several keywords to be matched into a first keyword set; Group similar keywords into a second keyword set; Group several keywords to be matched and similar keywords into a target set;

   Calculating the first word frequency vector of the first keyword set relative to the target set,

   Calculating a second word frequency vector of the second keyword set relative to the target set;

   The similarity between the first word frequency vector and the second word frequency vector is calculated as the similarity between the first keyword set and the second keyword set.
4. The method for extracting the content of a non-full grid table according to claim 2 or 3, wherein the filtering out similar keywords existing in a preset title keyword database that are similar to the keywords to be matched includes:

   Obtain unmatched target keywords from the title keyword database, match the target keywords with the keywords to be matched obtained after word segmentation processing to obtain the matching degree, continue to obtain the next unmatched target keywords and The target keyword is matched with the keyword to be matched to obtain the matching degree, until there is no unmatched target keyword in the title keyword library, and the target keyword with the highest matching degree and meeting the preset condition is regarded as the similar keyword.
5. The method for extracting the content of a non-full grid table according to claim 4, wherein the matching the target keyword with the keyword to be matched obtained after word segmentation processing to obtain the matching degree comprises:

   Pre-create a dictionary containing a large number of characters as a preset character set. Each character in the character set is represented by an N-dimensional vector representing its position in the character set, where N is the number of words in the character set;

   Split the target keyword and the keyword to be matched into several words, and by looking up the N-dimensional vector of each word, combine to form the target keyword vector corresponding to the target keyword and the keyword vector to be matched to the keyword to be matched Calculate the vector similarity between the target keyword vector and the keyword vector to be matched, and use the vector similarity as the matching degree between the target keyword and the keyword to be matched.
6. A device for extracting the content of a non-full grid table, which is characterized in that it comprises:

   An obtaining module for obtaining an original form, the original form being a non-full grid form;

   A reading module, which is used to read the original table to obtain a full table, and determine each cell belonging to the same merged row cell;

   The determination module is used to obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to the header cell or the content cell, and compare the header cell and the content cell mark;

   The filling module is configured to, if the merged cell includes a header cell, fill each cell belonging to the merged cell with the data of the header cell.
7. The device for extracting the content of a non-full grid table according to claim 6, wherein whether the cell whose data is not empty is determined to be a title cell or a content cell includes:

   Perform word segmentation processing on the cell data to obtain several keywords to be matched, and filter out similar keywords that are similar to the keywords to be matched in the preset title keyword library;

   Calculate the similarity between the first keyword set composed of several keywords to be matched with respect to the second keyword set composed of similar keywords;

   If the similarity is greater than the preset threshold, the cell is a title cell; otherwise, the cell is a content cell.
8. The non-full grid table content extraction device according to claim 7, wherein the calculation of the first keyword set consisting of several keywords to be matched is relative to the second keyword set consisting of similar keywords Similarity, including:

   Group several keywords to be matched into a first keyword set; Group similar keywords into a second keyword set; Group several keywords to be matched and similar keywords into a target set;

   Calculating the first word frequency vector of the first keyword set relative to the target set,

   Calculating a second word frequency vector of the second keyword set relative to the target set;

   The similarity between the first word frequency vector and the second word frequency vector is calculated as the similarity between the first keyword set and the second keyword set.
9. The device for extracting the content of a non-full grid table according to claim 7 or 8, wherein the filtering out similar keywords existing in a preset title keyword library that are similar to the keywords to be matched includes:

   Obtain unmatched target keywords from the title keyword database, match the target keywords with the keywords to be matched obtained after word segmentation processing to obtain the matching degree, continue to obtain the next unmatched target keywords and The target keyword is matched with the keyword to be matched to obtain the matching degree, until there is no unmatched target keyword in the title keyword library, and the target keyword with the highest matching degree and meeting the preset condition is regarded as the similar keyword.
10. The non-full form content extraction device according to claim 9, wherein the matching the target keyword with the keyword to be matched obtained after word segmentation processing to obtain the matching degree comprises:

    Pre-create a dictionary containing a large number of characters as a preset character set. Each character in the character set is represented by an N-dimensional vector representing its position in the character set, where N is the number of words in the character set;

    Split the target keyword and the keyword to be matched into several words, and by looking up the N-dimensional vector of each word, combine to form the target keyword vector corresponding to the target keyword and the keyword vector to be matched to the keyword to be matched Calculate the vector similarity between the target keyword vector and the keyword vector to be matched, and use the vector similarity as the matching degree between the target keyword and the keyword to be matched.
11. A terminal device includes a memory and a processor, the memory stores computer readable instructions that can run on the processor, and is characterized in that when the processor executes the computer readable instructions, the following is achieved step:

    Obtaining an original form, the original form being a non-full grid form;

    Read the original table to obtain a full table, and determine each cell belonging to the same merged cell;

    Obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and mark the header cell and the content cell;

    If the merged cell includes a header cell, each cell belonging to the merged cell is filled with the data of the header cell.
12. The terminal device according to claim 11, wherein whether the cell whose data is not empty belongs to a title cell or a content cell, comprising:

    Perform word segmentation processing on the cell data to obtain several keywords to be matched, and filter out similar keywords that are similar to the keywords to be matched in the preset title keyword library;

    Calculate the similarity between the first keyword set composed of several keywords to be matched with respect to the second keyword set composed of similar keywords;

    If the similarity is greater than the preset threshold, the cell is a title cell; otherwise, the cell is a content cell.
13. The terminal device according to claim 12, wherein the calculating the similarity of the first keyword set composed of several keywords to be matched with respect to the second keyword set composed of similar keywords comprises:

    Group several keywords to be matched into a first keyword set; Group similar keywords into a second keyword set; Group several keywords to be matched and similar keywords into a target set;

    Calculating the first word frequency vector of the first keyword set relative to the target set,

    Calculating a second word frequency vector of the second keyword set relative to the target set;

    The similarity between the first word frequency vector and the second word frequency vector is calculated as the similarity between the first keyword set and the second keyword set.
14. The terminal device according to claim 12 or 13, wherein the filtering out similar keywords existing in a preset title keyword library that are similar to the keywords to be matched includes:

    Obtain unmatched target keywords from the title keyword database, match the target keywords with the keywords to be matched obtained after word segmentation processing to obtain the matching degree, continue to obtain the next unmatched target keywords and The target keyword is matched with the keyword to be matched to obtain the matching degree, until there is no unmatched target keyword in the title keyword library, and the target keyword with the highest matching degree and meeting the preset condition is regarded as the similar keyword.
15. The terminal device according to claim 14, wherein the matching the target key word with the to-be-matched keyword obtained after word segmentation processing to obtain the matching degree comprises:

    Pre-create a dictionary containing a large number of characters as a preset character set. Each character in the character set is represented by an N-dimensional vector representing its position in the character set, where N is the number of words in the character set;

    Split the target keyword and the keyword to be matched into several words, and by looking up the N-dimensional vector of each word, combine to form the target keyword vector corresponding to the target keyword and the keyword vector to be matched to the keyword to be matched Calculate the vector similarity between the target keyword vector and the keyword vector to be matched, and use the vector similarity as the matching degree between the target keyword and the keyword to be matched.
16. A computer-readable storage medium, the computer-readable storage medium storing computer-readable instructions, wherein the computer-readable instructions are executed by a processor to implement the following steps:

    Obtaining an original form, the original form being a non-full grid form;

    Read the original table to obtain a full table, and determine each cell belonging to the same merged cell;

    Obtain the data of each cell from the full grid table in turn, determine whether the cell whose data is not empty belongs to a header cell or a content cell, and mark the header cell and the content cell;

    If the merged cell includes a header cell, each cell belonging to the merged cell is filled with the data of the header cell.
17. The computer-readable storage medium according to claim 16, wherein whether the cell for determining whether the data is not empty belongs to a title cell or a content cell, comprising:

    Perform word segmentation processing on the cell data to obtain several keywords to be matched, and filter out similar keywords that are similar to the keywords to be matched in the preset title keyword library;

    Calculate the similarity between the first keyword set consisting of several keywords to be matched with respect to the second keyword set consisting of similar keywords;

    If the similarity is greater than the preset threshold, the cell is a title cell; otherwise, the cell is a content cell.
18. The computer-readable storage medium according to claim 17, wherein the calculation of the similarity between a first keyword set consisting of several keywords to be matched with respect to a second keyword set consisting of similar keywords ,include:

    Group several keywords to be matched into a first keyword set; Group similar keywords into a second keyword set; Group several keywords to be matched and similar keywords into a target set;

    Calculating the first word frequency vector of the first keyword set relative to the target set,

    Calculating a second word frequency vector of the second keyword set relative to the target set;

    The similarity between the first word frequency vector and the second word frequency vector is calculated as the similarity between the first keyword set and the second keyword set.
19. The computer-readable storage medium according to claim 17 or 18, wherein the filtering out similar keywords existing in a preset title keyword library that are similar to the keywords to be matched includes:

    Obtain unmatched target keywords from the title keyword database, match the target keywords with the keywords to be matched obtained after word segmentation processing to obtain the matching degree, continue to obtain the next unmatched target keywords and The target keyword is matched with the keyword to be matched to obtain the matching degree, until there is no unmatched target keyword in the title keyword library, and the target keyword with the highest matching degree and meeting the preset condition is regarded as the similar keyword.
20. The computer-readable storage medium according to claim 19, wherein the matching the target keyword with the keyword to be matched obtained after word segmentation processing to obtain the matching degree comprises:

    Pre-create a dictionary containing a large number of characters as a preset character set. Each character in the character set is represented by an N-dimensional vector representing its position in the character set, where N is the number of words in the character set;

    Split the target keyword and the keyword to be matched into several words, and by looking up the N-dimensional vector of each word, combine to form the target keyword vector corresponding to the target keyword and the keyword vector to be matched to the keyword to be matched Calculate the vector similarity between the target keyword vector and the keyword vector to be matched, and use the vector similarity as the matching degree between the target keyword and the keyword to be matched.

Images