CN114357196A - Knowledge graph-based ROS message type searching method and system - Google Patents
Knowledge graph-based ROS message type searching method and system Download PDFInfo
- Publication number
- CN114357196A CN114357196A CN202210031341.XA CN202210031341A CN114357196A CN 114357196 A CN114357196 A CN 114357196A CN 202210031341 A CN202210031341 A CN 202210031341A CN 114357196 A CN114357196 A CN 114357196A
- Authority
- CN
- China
- Prior art keywords
- message
- name
- knowledge
- graph
- ros
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000012545 processing Methods 0.000 claims description 25
- 238000000605 extraction Methods 0.000 claims description 11
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 239000002131 composite material Substances 0.000 description 15
- 230000011218 segmentation Effects 0.000 description 10
- 230000018109 developmental process Effects 0.000 description 8
- 238000011161 development Methods 0.000 description 7
- 150000001875 compounds Chemical class 0.000 description 6
- 238000005070 sampling Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 240000008168 Ficus benjamina Species 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a knowledge graph-based ROS message type searching method and a knowledge graph-based ROS message type searching system. The method comprises the following steps: acquiring a message packet name and a message name of the ROS message; extracting feature words based on the message packet name and the message name; constructing a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words; searching for ROS message types based on the knowledge-graph. The invention provides the knowledge map among the message packet, the message and the characteristics of the existing message type, is beneficial to a user to refer to more related information, and can improve the searching efficiency and speed.
Description
Technical Field
The invention relates to the technical field of robot operating systems, in particular to a knowledge-graph-based ROS message type searching method and system.
Background
The Robot Operating System (ROS) has become the most popular robot development framework over the past few years, a loosely coupled development framework based on message passing between nodes. In ROS, messages are a key part of ensuring proper communication between different nodes of the robot program. Usually a complete robot program will contain a large number of messages.
The composition of the message in ROS is shown in the official document in fig. 1. The general message composition method is mainly two types:
the ROS has built-in basic data types, such as int, float, pool, string, etc., and the array composed of these basic data types, such as int [ ], float [ ], pool [ ], string [ ], etc.
A composite data type, such as the Point message shown in fig. 2, is a composite message in the geometry _ msgs provided by the authority, that is, a self-defined message, and a developer can directly use the message as a message data type to define variables in the message.
By the definition mode, a developer can directly use the message as a message data type without directly using the basic data type to define the message. A message is thus exactly a message data type. In the actual robot program development process, when a compound message type needs to be used, it is not easy to find the compound message type. In the ROS the messages are all stored in message packets. As the point message in fig. 2 is stored in the geometry message package geometry _ msgs, a message package usually contains multiple messages. The number of message bundles officially included by the current ROS is 264. Wherein the number of composite message types is 3244. When these composite data types are used, the packet name of the message packet in which the message is located needs to be provided. Because different message packages may contain the same message name, the ROS needs to know which message package the message type is in when parsing the composite message types. However, in the actual development process, sometimes the developer does not know which message packet the message type required by the developer is in, because he does not know which composite data type can be used, and only knows some characteristic information of the currently required message type, such as a message related to navigation, a speed message, or the like. At this time, the developer only knows the task characteristics of the message to be completed currently, and does not know which composite message types conforming to the task characteristics can be used. The possible message types can only be queried in a natural language description manner when looking up through the ROS community or a general search engine. However, the ROS community does not provide a tool dedicated to query message types, and the results of the query often cannot completely meet the needs of developers when a general-purpose search engine is used for querying, and a relatively rich query statement needs to be input.
Sometimes, even if the developer knows what the type of compound data he needs is, it needs to look up the message by entering the name of the message package. And the longest message package name character length is 28 and the average message package name character length is 15. As shown in fig. 3, package names are often composed of multiple words together and constitute a word block and mixed with abbreviations. The packet names of message packets are often very lengthy and complex, and difficult to remember and spell correctly. The problem cannot be completely solved by the fuzzy query mode based on the regular expression provided by the ROS search engine due to reasons of misspelling of the user or forgetting words by default and the like.
Disclosure of Invention
In view of the above problems, an object of the present invention is to provide a method and a system for searching for ROS message types based on a knowledge-graph, so as to search for ROS message types efficiently and quickly.
In order to achieve the purpose, the invention provides the following scheme:
a knowledge-graph-based ROS message type searching method comprises the following steps:
acquiring a message packet name and a message name of the ROS message;
extracting feature words based on the message packet name and the message name;
constructing a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words;
searching for ROS message types based on the knowledge-graph.
Optionally, the extracting feature words based on the message packet name and the message name specifically includes:
carrying out prefix and suffix removal processing on the message packet name and the message name;
extracting initial feature words based on the processed message packet name and the processed message name;
and carrying out data processing on the initial characteristic words to obtain characteristic words.
Optionally, the method further comprises: prefix removing processing is carried out on the message packet name, a message packet name library is constructed,
optionally, the knowledge graph further includes website information of the message package.
Optionally, the searching for ROS message types based on the knowledge-graph specifically includes:
determining a message packet name to be searched from the message packet name library according to the input initial message packet name;
and searching the ROS message type from the knowledge graph according to the message packet name to be searched.
Optionally, the searching for ROS message types based on the knowledge-graph specifically includes:
and searching the ROS message type from the knowledge graph according to the input characteristic words to be searched.
The invention also provides a knowledge-graph-based ROS message type searching system, which comprises:
the message package name and message name acquisition module is used for acquiring the message package name and the message name of the ROS message;
the characteristic word extraction module is used for extracting characteristic words based on the message packet name and the message name;
the knowledge graph building module is used for building a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words;
and the ROS message type searching module is used for searching the ROS message type based on the knowledge graph.
Optionally, the feature word extraction module specifically includes:
the first processing unit is used for carrying out prefix and suffix removal processing on the message packet name and the message name;
the extraction unit is used for extracting initial characteristic words based on the processed message packet name and the processed message name;
and the second processing unit is used for carrying out data processing on the initial characteristic words to obtain the characteristic words.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention discloses a knowledge graph-based ROS message type searching method, which comprises the following steps: acquiring a message packet name and a message name of the ROS message; extracting feature words based on the message packet name and the message name; constructing a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words; searching for ROS message types based on the knowledge-graph. The invention provides the knowledge map among the message packet, the message and the characteristics of the existing message type, is beneficial to a user to refer to more related information, and can improve the searching efficiency and speed.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a diagram of the manner in which messages are composed in the ROS;
FIG. 2 is a custom data type Point;
FIG. 3 is the first 20 message packet names in alphabetical order among all message packets;
FIG. 4 is a message named ActionDescription;
FIG. 5 is a message length distribution graph for 202 messages;
FIG. 6 is a flow chart of a knowledge-graph based ROS message type search method of an embodiment of the present invention;
FIG. 7 is a schematic diagram of a knowledge-graph based ROS message type search method of an embodiment of the present invention;
FIG. 8 is a message name and its characteristic information;
FIG. 9 is a diagram of a message packet name and its characteristic information;
FIG. 10 is a presentation case of English text word segmentation;
FIG. 11 is a knowledge-graph;
FIG. 12 is a sub-graph presentation of a message type knowledge-graph;
fig. 13 is a message type search flow diagram.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In order to know the actual development situation of the message in the current ROS open source project, the invention collects the 12 items with the highest number of the tags of the star in the gitubs as the ROS projects, and counts the self-defined message files in the 12 items. The statistics show a total of 202 message files, i.e., 202 custom messages. As shown in fig. 4, is the content of a message of length 5 named ActionDescription. The present invention refers to variables defined in a message as message variables, and the number of message variables as the length of the message. Generally, a message contains a plurality of message variables. The more message variables, i.e. the longer the message, the more complex the message. These messages of the current statistics contain a total of 2480 message variables. Fig. 5 shows the message length distribution of the 202 messages.
Statistics show that each item contains an average of 16.83 custom messages, and each message contains an average of 12.277 message variables. It can be seen that the number of custom messages in the project is relatively high, with an average message variance per message also exceeding 12. As can be seen from fig. 5: the proportion of messages with a length of less than 5 is only 30.69%. The proportion of messages of length greater than 10 is 39.6%. The proportion of messages of length greater than 20 is 16.34%. There are a small number of messages whose length exceeds 50. Even with messages of 149 a length. A large percentage of messages in an actual development project are long in message length. In addition, the invention counts the use condition of the compound message type. Statistics show that the number of messages using the composite message type among 202 messages is 34, with a proportion of only 16.83%. And in 2480 message variants of 202 messages, the composite message type is 79, accounting for only 3.18%. It can be seen that the frequency of use of the composite message type is low and that a large number of messages use more of the underlying data type.
One important reason for this is that it is difficult for developers to find the composite message types that they need. It is difficult to find a composite type of message that meets the current needs. First, the number of officially included composite message types is 3244. A large number of compound message types have not been used or known, or even known to exist. Secondly, only the package name can be searched when searching in the ROS search engine, but the message name cannot be searched directly. However, the information of the type of message contained in the packet cannot be known from the packet name alone. Finally, in many cases when composing a message, the developer only knows the features that the composed message has. Such as a three-dimensional coordinate point. Characterized by point, 3d, space, etc., it is difficult and not comprehensive to find the required message type under current search engines based only on these.
In addition, 3244 composite message types are included in the current official. Many message types already have the same or similar messages disclosed that can be reused. Even though the developer does not need to fully use the existing complex message types, the existing message types related to the current task can still provide a reference for the developer or copy a part of the message, thereby reducing the development amount and difficulty.
The invention aims to provide a knowledge-graph-based ROS message type searching method and a knowledge-graph-based ROS message type searching system, which are used for efficiently and quickly searching ROS message types.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
As shown in fig. 6-7, the method for searching for ROS message types based on knowledge-graph provided by the present invention comprises the following steps:
step 101: and acquiring the message packet name and the message name of the ROS message.
And accessing an ROS community pack library website https:// index. ROS. org/packages/page/1/time/, inputting msg to query all message packs, and copying and downloading the pack names of the message packs and the message file names in the message packs through the local computer to obtain message pack name data A and message name data B.
Step 102: and extracting feature words based on the message packet name and the message name. The method specifically comprises the following steps: carrying out prefix and suffix removal processing on the message packet name and the message name; extracting initial feature words based on the processed message packet name and the processed message name; and carrying out data processing on the initial characteristic words to obtain characteristic words.
And removing redundant information of the front part and the back part of the message packet name data A and the message name data B by using a local computer and a programmed python program to obtain processed message packet name data A 'and processed message name data B'.
Since the message is composed with the characteristic information of the message, the characteristic information of the message is extracted to inquire the needed message. The invention has two benefits over using natural language queries: on one hand, the input of a developer is simplified, the query is convenient, and the characteristics of the problem are better met; on the other hand, the method is easier to realize and easier to use. The initial message feature collection scheme is to extract feature information of a message from a description information text of the message using a natural language processing algorithm. But in practice two problems are encountered. The first problem is that the message packets are distributed among different repositories, which has great difficulty in acquiring data. The second problem is that there is a considerable proportion of message packets and messages within them that do not contain descriptive information and the characteristics of the messages cannot be derived from the descriptive text. As shown in fig. 8 and 9, the present invention observes and analyzes the message packet name data a 'and the message name data B' and finds that the message name and the message packet name contain characteristic information of many messages. It is possible to attempt to extract feature information from the message name and the message bundle name. In addition, since each message will have a name, there is no problem with having no source of characteristic information. The present invention requires verification of whether the features contained in the naming can effectively characterize the features of the message.
Random sampling investigation is conducted on the message packet name data A 'and the message name data B'. Because the messages in the same message packet have the same code style, the characteristics of all the messages in the message packet can be represented only by extracting the first message from the message packet and analyzing the result. It is only necessary to randomly sample from the 264 message packets and analyze the first message in the message packet. The invention randomly selects ten message packets, and analyzes and finds that half of the messages have no description information and lack of comparison samples. Therefore, the invention randomly extracts three times respectively, and ten times are extracted each time. Four random samplings take forty samples out of 39 different samples. The sampling proportion was 14.77% of the total. The invention manually extracts the characteristics in the message name and the message packet name and the characteristics in the description information and compares the extracted characteristics in the message name and the message packet name. Table 1 is a comparative illustration of a sample of a sampled portion of a message, the third column being manually extracted features from the named name, the fourth column being description information of the message, wherein the bold portions are manually extracted features from the description information. Table 2 is the final analytical statistics. It can be seen that the percentage of messages without descriptive information and messages named to fully characterize the feature is 82%. There are no messages named that cannot characterize a message. The remaining seven parts of the characterizable message are analyzed to find that five of them can include the main characteristics of the message. The other two are shown in table 3, because the text content of the description information is many and contains a large amount of introductory content, the characteristics in the naming are difficult to represent all the information, but the naming contains the main characteristic information of the message. It can be considered that the features extracted by the message name and the message bundle name can sufficiently characterize the message.
Table 1 results of partial sampling
TABLE 2 statistical results of the sampling survey
| Categories | Number of | Percentage of |
| Sampling sample | 39 | 1000 |
| Message without description | 19 | 487 |
| The characteristics of the message name and the packet name can fully represent the description text information of the |
13 | 333 |
| The characteristics of the message name and the packet name can partially represent the description text information of the |
7 | 179 |
| The characteristics of the message name and the packet name can not represent the description text information of the |
0 | 00 |
TABLE 3 naming of two messages in a sample survey messages that can only partially characterize the characteristics of the messages
The invention collects all message bundles and messages from the bundle repository of the ROS official community for a total of 264 message bundles, which contain 3244 messages. However, manually extracting features from a total of 3508 message bundle names and message names is a very time-consuming and laborious task. Therefore, a library of English text word segmentation modules named word nja in the programming language python is used for automatically extracting the characteristic information in the names. Fig. 10 is a word segmentation demonstration case using the word segmentation module. It can be seen that the word segmentation module can completely separate the english words which are concatenated together. The module is therefore able to meet the requirements of the present invention for extracting features in message names and message bundle names. Next, feature extraction is performed on the message packet name data a 'and the message name data B'. And performing word segmentation and feature extraction on the message packet name data A 'and the message name data B' by using a word segmentation module on a python programming platform of the local computer to acquire 9800 pieces of feature word data D.
After analyzing the data D, the problems of repetition, word form change, noun abbreviation, single letter and the like are found. The following are examples of the kind of problem that arises, respectively.
v+ing:Bounding、Charging、Planning
v+ed:Stamped、Desired、Planned
n+s/es:Controls、Trajectories
Abbreviation conf-configuration, pos-position, nav-navigation
Prepositions of for, in, at, of, on, with
The be verb and the assistant verb am/is/are, do
Repetition of characteristic words
Therefore, secondary data processing is required for the extracted feature words.
The processing can not be modified only by designing simple computer processing program rules. For example, when dealing with a running noun, not all words suffixed with ing are in the form of a running noun, such as string. Some words with a suffix ed are also noted when dealing with past segmentations, e.g., speed is not a form of past segmentations. It is also necessary to note words suffixed with s when dealing with nouns. Generally, character strings with a length less than 3 are meaningless character strings, but still a few meaningful character strings need to be considered separately. Completion of the word abbreviation is the most intractable place. Not only is it necessary to know the word before the abbreviation but also to determine which word the abbreviation represents based on various information of the message. Even though the same abbreviation, may not be the same word in different message names. This is the most time consuming place. And for prepositions, be verbs and auxiliary verbs, only a computer program needs to be designed and all the prepositions, the be verbs and the auxiliary verbs are deleted. On the problem of removing the repeated characteristic words, only data D needs to be put into a list of a local computer python program, and the repeated characteristic words are automatically removed through a sort () method built in the python language. The invention designs eight different python processing programs on the local computer respectively aiming at each problem type. And processing the data D sequentially through eight processing programs to obtain feature word data E.
Step 103: and constructing a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words.
Through the relation analysis of the message, the message packet and the feature words, a complex graph structure is found among the message, the message packet and the feature words. One message packet contains a plurality of messages. Some messages of the same name are contained in different message packages. A message packet contains one or more features. A message contains one to more features. Some features may be included in a plurality of different messages and message packets. Fig. 11 is a simplified diagram showing a partial relationship among the three. In addition, the message packages and the messages all have web pages for describing the content of the messages in detail, and developers sometimes need to check the detailed information of the messages after finding the required compound message type. The invention does not crawl and reorganize all the unstructured description information, but adds the web addresses of the web pages into the description information of the message packets. The information is completely crawled, so that time and labor are wasted, and the good display effect of the original webpage is still not achieved after the information is crawled and recombined. Therefore, only the web address of the web page needs to be provided. The invention only needs to provide the function of searching the composite message type meeting the task requirement. In conclusion, the method using the knowledge graph can better solve the searching problem. Firstly, the graph structure of the knowledge graph is consistent with the graph structures among the message packet, the message and the feature words. Second, the knowledge-graph can add rich information into the graph nodes. Finally, because some potential associations exist among different characteristic words, such as semantic associations, association of affiliations and the like, the associations can be better found by using the knowledge graph, and thus a foundation is laid for the next recommendation function. The invention selects the popular noe4j graph database to construct the knowledge graph.
Entity linkage is needed to construct the knowledge graph, and the structural relationship among different entities is represented in the graph. But how the structural relationship is determined by the problem itself and the solution. The present invention proposes to link all features onto the message package in this problem. Because some of the feature words come from the message bundle and do not belong to the message. These features can only be queried by means of links to message packets. In addition, in actual use, all messages must know the message packet to which the message belongs. Thus linking all features to the message bundle. The method uses a programmed python program on a local computer to put the data A and the data B together in a mode that the messages need to be under the message packets, and uses a wordninja module in the python language to perform word segmentation on the data A and the data B to obtain data F.
And sequencing the data F into a required sequence according to the structure of the map by using a python program programmed on a local computer, adding the website information of the message packet, and then importing the data F into a CSV file to obtain data G.
On the neo4j database on the local computer, the message type knowledge graph is constructed by importing the local computer data G into the neo4j database using the load command of neo4 j. FIG. 12 is a partial sub-graph structure display of knowledge-graphs of the present invention.
Step 104: searching for ROS message types based on the knowledge-graph. The method specifically comprises the following steps: determining a message packet name to be searched from the message packet name library according to the input initial message packet name; and searching the ROS message type from the knowledge graph according to the message packet name to be searched. Or searching the ROS message type from the knowledge graph according to the input characteristic words to be searched.
The neo4j database query program is designed according to different query modes for users to use. The step of querying is from the feature- > message package- > message or directly from the message package- > message. FIG. 13 is a flow chart of a message type query.
The user sometimes knows the message packet in which the message is located, and can directly input the message packet name to inquire the message in the packet. However, it is not easy to correctly enter the message bundle name because the message bundle name is lengthy and complicated and contains many acronyms, and many times a developer is likely to enter an error when entering the message bundle name of a query. Although the search engine of the ROS provides developers with fuzzy search functions for forward matching and backward matching, sometimes the user does not know the long enough front and back parts of the package name in the actual search process. If the query part is too short, a large number of irrelevant results can be returned, and the user needs to continuously page and search. Thus, the present invention builds a library of message package names separately on the local computer by removing the data a prefixes. Because the problems of user input error and the like cannot be solved through the regular expression, the invention queries the message packet name in the message packet library by using a character string similarity ranking mode. The similarity calculation method of the character strings adopts a minimum edit distance algorithm and an inclusion matching algorithm. The minimum editing distance refers to the minimum number of editing operations required to change from one string to another string. Permitted editing operations include replacing one character with another, inserting one character, and deleting one character. Generally, the smaller the edit distance, the greater the similarity of the two strings. The invention calls a module named fuzzy wuzzy in python on a computer to realize the character string similarity ranking.
Compared with the ROS community and the traditional search engine, the method for inquiring the message type based on the characteristics can enable the user to inquire the message type needed by the user more quickly, and therefore development efficiency is improved. And a structural network among the three of 'message packet-message-characteristics' of the existing message type can be provided, so that the user can refer to more related information.
The invention also provides a knowledge-graph-based ROS message type searching system, which comprises:
the message package name and message name acquisition module is used for acquiring the message package name and the message name of the ROS message;
the characteristic word extraction module is used for extracting characteristic words based on the message packet name and the message name;
the knowledge graph building module is used for building a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words;
and the ROS message type searching module is used for searching the ROS message type based on the knowledge graph.
The feature word extraction module specifically comprises:
the first processing unit is used for carrying out prefix and suffix removal processing on the message packet name and the message name;
the extraction unit is used for extracting initial characteristic words based on the processed message packet name and the processed message name;
and the second processing unit is used for carrying out data processing on the initial characteristic words to obtain the characteristic words.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. A knowledge-graph-based ROS message type searching method is characterized by comprising the following steps:
acquiring a message packet name and a message name of the ROS message;
extracting feature words based on the message packet name and the message name;
constructing a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words;
searching for ROS message types based on the knowledge-graph.
2. The knowledge-graph-based ROS message type searching method of claim 1, wherein said extracting feature words based on said message packet name and said message name, specifically comprises:
carrying out prefix and suffix removal processing on the message packet name and the message name;
extracting initial feature words based on the processed message packet name and the processed message name;
and carrying out data processing on the initial characteristic words to obtain characteristic words.
3. The knowledge-graph-based ROS message type searching method of claim 1, further comprising: and carrying out prefix removal processing on the message packet name to construct a message packet name library.
4. The knowledge-graph-based ROS message type searching method of claim 1, wherein said knowledge-graph further comprises the web address information of the message package.
5. The knowledge-graph-based ROS message type searching method of claim 3, wherein said knowledge-graph-based ROS message type searching specifically comprises:
determining a message packet name to be searched from the message packet name library according to the input initial message packet name;
and searching the ROS message type from the knowledge graph according to the message packet name to be searched.
6. The knowledge-graph-based ROS message type searching method of claim 1, wherein said knowledge-graph-based ROS message type searching specifically comprises:
and searching the ROS message type from the knowledge graph according to the input characteristic words to be searched.
7. A knowledge-graph based ROS message type searching system, comprising:
the message package name and message name acquisition module is used for acquiring the message package name and the message name of the ROS message;
the characteristic word extraction module is used for extracting characteristic words based on the message packet name and the message name;
the knowledge graph building module is used for building a knowledge graph based on the message packet name, the message name and the structural relationship among the feature words;
and the ROS message type searching module is used for searching the ROS message type based on the knowledge graph.
8. The knowledge-graph-based ROS message type searching system of claim 7, wherein said feature word extraction module specifically comprises:
the first processing unit is used for carrying out prefix and suffix removal processing on the message packet name and the message name;
the extraction unit is used for extracting initial characteristic words based on the processed message packet name and the processed message name;
and the second processing unit is used for carrying out data processing on the initial characteristic words to obtain the characteristic words.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210031341.XA CN114357196A (en) | 2022-01-12 | 2022-01-12 | Knowledge graph-based ROS message type searching method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210031341.XA CN114357196A (en) | 2022-01-12 | 2022-01-12 | Knowledge graph-based ROS message type searching method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114357196A true CN114357196A (en) | 2022-04-15 |
Family
ID=81109169
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210031341.XA Pending CN114357196A (en) | 2022-01-12 | 2022-01-12 | Knowledge graph-based ROS message type searching method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114357196A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110399470A (en) * | 2018-04-24 | 2019-11-01 | 微软技术许可有限责任公司 | Conversation message processing |
| CN113297349A (en) * | 2021-05-18 | 2021-08-24 | 中国人民解放军国防科技大学 | Knowledge graph-based ROS software package recommendation method and system |
-
2022
- 2022-01-12 CN CN202210031341.XA patent/CN114357196A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110399470A (en) * | 2018-04-24 | 2019-11-01 | 微软技术许可有限责任公司 | Conversation message processing |
| CN113297349A (en) * | 2021-05-18 | 2021-08-24 | 中国人民解放军国防科技大学 | Knowledge graph-based ROS software package recommendation method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110399457B (en) | Intelligent question answering method and system | |
| CN107797991B (en) | Dependency syntax tree-based knowledge graph expansion method and system | |
| CN110390006B (en) | Question-answer corpus generation method, device and computer readable storage medium | |
| CN1924858B (en) | Method and device for fetching new words and input method system | |
| CN109947952B (en) | Retrieval method, device, equipment and storage medium based on English knowledge graph | |
| US20040221229A1 (en) | Data structures related to documents, and querying such data structures | |
| JP2001167087A (en) | Device and method for retrieving structured document, program recording medium for structured document retrieval and index preparing method for structured document retrieval | |
| US9740767B2 (en) | Systems and methods for analyzing failed and successful search queries | |
| CN110909016B (en) | Repeated association detection method, device, equipment and storage medium based on database | |
| KR100835706B1 (en) | System and method for korean morphological analysis for automatic indexing | |
| CN110795526B (en) | Mathematical formula index creating method and system for retrieval system | |
| CN113515600B (en) | Automatic calculation method for spatial analysis based on metadata | |
| CN101288071A (en) | Method and computer unit for determining computer service names | |
| JP4237813B2 (en) | Structured document management system | |
| Cortez et al. | A flexible approach for extracting metadata from bibliographic citations | |
| CN114357196A (en) | Knowledge graph-based ROS message type searching method and system | |
| CN105426490A (en) | Tree structure based indexing method | |
| CN108614821B (en) | Geological data interconnection and mutual-checking system | |
| Bo et al. | Towards an efficient searching approach of ros message by knowledge graph | |
| Doerr et al. | A method for estimating the precision of placename matching | |
| CN110909128B (en) | Method, equipment and storage medium for carrying out data query by using root list | |
| CN112183074A (en) | Data enhancement method, device, equipment and medium | |
| CN111782958A (en) | Recommendation word determining method and device, electronic device and storage medium | |
| CN110930189A (en) | Personalized marketing method based on user behaviors | |
| JP2007026116A (en) | Concept search system and concept search method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |