CN111324609A

CN111324609A - Knowledge graph construction method and device, electronic equipment and storage medium

Info

Publication number: CN111324609A
Application number: CN202010096405.5A
Authority: CN
Inventors: 熊俊宇; 魏琪康; 周煜; 钟黎; 刘黎春
Original assignee: Tencent Cloud Computing Beijing Co Ltd
Current assignee: Tencent Cloud Computing Beijing Co Ltd
Priority date: 2020-02-17
Filing date: 2020-02-17
Publication date: 2020-06-23
Anticipated expiration: 2040-02-17
Also published as: CN111324609B

Abstract

The application discloses a knowledge graph construction method, a knowledge graph construction device, electronic equipment and a storage medium, which are used for reducing the difficulty of creating a knowledge graph, and the method comprises the following steps: obtaining tables corresponding to at least two entity types respectively, wherein each table comprises a name column and at least one relation series, the name column comprises entity names of entities belonging to the corresponding entity type, and each relation column comprises an association relation between the entities in the tables and the entities in any other table; respectively generating first sentences corresponding to the entity names according to the entity names extracted from the name columns of the tables, wherein each first sentence is used for adding an entity node corresponding to the entity name in the graph database; respectively generating second sentences corresponding to the association relations according to the association relations extracted from the relation series of the tables, wherein each second sentence is used for adding edges representing the association relations among the entity nodes; and calling the first statement and the second statement generated by the graph database analysis to obtain the knowledge graph.

Description

Knowledge graph construction method and device, electronic equipment and storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for constructing a knowledge graph, an electronic device, and a storage medium.

Background

The construction of knowledge maps typically involves a relatively cumbersome process: firstly, processing data to be processed into a data form meeting the import requirement of a graphic database by a professional according to the requirement of the graphic database; then, manually writing a complex import statement, importing the data to be processed which meets the import requirement into a graph database by using the import statement, and generating a corresponding knowledge graph by the graph database. However, errors are easily caused in the process of manually processing data, the data maintenance difficulty is high, a user needs to learn a using method of the graphic database, and the use threshold is very high.

Disclosure of Invention

The embodiment of the application provides a knowledge graph construction method and device, electronic equipment and a storage medium, so that difficulty in creating the knowledge graph is reduced, and common users can conveniently and efficiently complete creation of the knowledge graph.

In one aspect, an embodiment of the present application provides a method for constructing a knowledge graph, including:

obtaining tables corresponding to at least two entity types respectively, wherein each table comprises a name column and at least one relation series, the name column comprises entity names of entities belonging to the corresponding entity type, and each relation column comprises an association relation between the entities in the tables and the entities in any other table;

respectively generating first sentences corresponding to the entity names according to the entity names extracted from the name columns of the tables, wherein each first sentence is used for adding an entity node corresponding to the entity name in the graph database;

respectively generating second sentences corresponding to the association relations according to the association relations extracted from the relation series of the tables, wherein each second sentence is used for adding edges representing the association relations among the entity nodes;

and calling the first statement and the second statement generated by the graph database analysis to obtain the knowledge graph.

responding to a map building instruction input through a file uploading page, and sending a file to be uploaded to a background server, wherein the file to be uploaded comprises: the system comprises tables corresponding to at least two entity categories respectively, each table comprises a name column and at least one relation series, the name column comprises entity names of entities belonging to the corresponding entity category, and each relation column comprises an association relation between the entities in the tables and the entities in any other table;

and displaying the knowledge graph sent by the background server, wherein the knowledge graph is a graph which is generated by the background server according to the file to be uploaded and is used for describing the association relationship among the entities.

In one aspect, an embodiment of the present application provides a knowledge graph constructing apparatus, including:

the system comprises an acquisition module, a storage module and a display module, wherein the acquisition module is used for acquiring tables corresponding to at least two entity categories respectively, each table comprises a name column and at least one relation series, the name column comprises entity names of entities belonging to the corresponding entity categories, and each relation column comprises an association relation between an entity in the table and an entity in any other table;

the first generation module is used for respectively generating first sentences corresponding to the entity names according to the entity names extracted from the name columns of the tables, and each first sentence is used for adding an entity node corresponding to the entity name in the graph database;

the second generation module is used for respectively generating second sentences corresponding to the association relations according to the association relations extracted from the relation series of the tables, and each second sentence is used for adding edges representing the association relations among the entity nodes;

and the analysis module is used for calling the first statement and the second statement generated by the graphic database analysis to obtain the knowledge graph.

Optionally, wherein the table name of each table is the class name of the corresponding entity category; the first cell of each name column in each table comprises a name column identifier, and the cells except the first cell in each name column comprise entity names; the first cell of each relation series in each table comprises a relation column identification and a class name of an entity class which has an association relation with the entity class corresponding to the table to which the relation column belongs, and any cell except the first cell in each relation series respectively comprises an entity name which has an association relation with the entity corresponding to the cell and belongs to the table corresponding to the class name in the first cell.

Optionally, the second generating module is further configured to: extracting the association relation from any relation series of any table by the following method:

and aiming at any cell except the first cell in any relation series, extracting a first entity name from the cell intersected with the row where any cell is positioned and the column where the name column is positioned, extracting a second entity name from any cell, and determining that an association relation exists between an entity corresponding to the first entity name and an entity corresponding to the second entity name.

Optionally, the first cell of each relationship series further includes a relationship name describing the association relationship.

Correspondingly, the second generation module is further configured to extract a relationship name from the first cell of any one of the relationship series, and the second sentence further includes the relationship name displayed on the side of the knowledge graph.

Optionally, the apparatus further comprises a schema creation module for:

creating category nodes corresponding to the entity categories according to the category names of the entity categories corresponding to the tables;

for any relation series of any table, extracting a class name from a first cell of the relation series, and determining that an association relationship exists between an entity category corresponding to the extracted class name and an entity category corresponding to the table;

and creating edges among the class nodes with the association relation to obtain a schema chart representing the association relation among the entity classes.

Optionally, the apparatus further includes a checking module, configured to check data in each table before generating the first statement and the second statement, and if it is found that there is error data in the table, generate corresponding error information.

Optionally, each table further includes an attribute column, where the attribute column is used to store attribute description information corresponding to each entity name in the table; the first statement corresponding to each entity name also comprises attribute description information corresponding to each entity name.

Optionally, the apparatus further includes a dictionary generating module, configured to generate a dictionary file corresponding to the entity according to the entity name in each table name column and the entity category corresponding to the entity name.

Optionally, each table further includes a synonym column, and the synonym column is used for storing the alias corresponding to each entity name in the table.

Correspondingly, the device also comprises a synonym generating module which is used for generating a synonym file according to the entity name of the name column of each table and the alias corresponding to the entity name extracted from the synonym column.

Optionally, the table corresponding to each entity category is a worksheet in the same table file.

the uploading module is used for responding to a map building instruction input through a file uploading page and sending a file to be uploaded to the background server, wherein the file to be uploaded comprises: the system comprises tables corresponding to at least two entity categories respectively, each table comprises a name column and at least one relation series, the name column comprises entity names of entities belonging to the corresponding entity category, and each relation column comprises an association relation between the entities in the tables and the entities in any other table;

and the display module is used for displaying the knowledge graph sent by the background server, and the knowledge graph is a graph which is generated by the background server according to the file to be uploaded and is used for describing the association relationship among the entities.

Optionally, the display module is further configured to:

after error information sent by the background server is received, error prompt information is displayed, wherein the error information comprises an error table with error data;

and displaying the error table in response to the operation aiming at the error prompt information, wherein the error data in the error table is displayed in a special state.

In one aspect, an embodiment of the present application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the methods when executing the computer program.

In one aspect, an embodiment of the present application provides a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, implement the steps of any of the above-described methods.

In one aspect, an embodiment of the present application provides a computer program product comprising a computer program stored on a computer-readable storage medium, the computer program comprising program instructions that, when executed by a processor, implement the steps of any of the methods described above.

According to the knowledge graph construction method, the knowledge graph construction device, the electronic equipment and the storage medium, a user only needs to input data into a form according to a preset format and upload the form to a knowledge graph construction tool in a background server or terminal equipment, the knowledge graph construction tool can automatically extract entity names and association relations from name columns and relation series of each form edited by the user, sentences for creating nodes and edges are automatically generated, a graphic database is called to complete the creation of the knowledge graph, the complex data operation, the imported sentence compiling and other work during the creation of the knowledge graph are avoided, and people who are not professionals related to the knowledge graph can also create the knowledge graph conveniently and efficiently. In addition, the form format is simple and visual, the data maintenance and modification are easy, and the user who does not know the knowledge map can finish the data arrangement quickly, so that the use threshold is greatly reduced.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1A is a partial document prepared for creating a knowledge-graph using Neo4 j;

FIG. 1B is a diagram of internal data of a csv type entity file;

FIG. 1C is a diagram of internal data of a csv type relationship file;

FIG. 1D is a diagram of an import statement manually written by a user;

FIG. 2 is a schematic view of an application scenario of a knowledge graph construction method provided in an embodiment of the present application;

FIG. 3 is a schematic flow chart diagram of a method for constructing a knowledge graph according to an embodiment of the present application;

FIG. 4 is a diagram of a table for constructing a knowledge graph according to an embodiment of the present application;

FIG. 5 is a partial schematic view of a knowledge-graph obtained by a knowledge-graph construction method provided by an embodiment of the present application;

FIG. 6 is a schematic flow chart diagram of a knowledge graph construction method according to an embodiment of the present application;

FIG. 7 is a schematic diagram of a knowledge graph constructed by a knowledge graph construction method according to an embodiment of the present application;

FIG. 8 is a schematic flow chart diagram of a knowledge graph construction method provided in an embodiment of the present application;

fig. 9A is a schematic interface diagram of a document uploading tool according to an embodiment of the present application;

FIG. 9B is a schematic interface diagram of a document uploading tool according to an embodiment of the present application;

FIG. 10 is a schematic structural diagram of a knowledge graph constructing apparatus according to an embodiment of the present application;

FIG. 11 is a schematic structural diagram of a knowledge graph constructing apparatus according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application.

For convenience of understanding, terms referred to in the embodiments of the present application are explained below:

a graph database: is a type of NoSQL database that applies graph theory to store relationship information between entities. The graph database is a non-relational database that stores relational information between entities using graph theory. Common graphic databases are Neo4j, FlockDB, allegrograph, graph db, InfiniteGraph, etc.

Knowledge Graph (Knowledge Graph): the knowledge domain visualization or knowledge domain mapping map is a series of different graphs for displaying the relationship between the knowledge development process and the structure, and the visualization technology is used for describing knowledge resources and carriers thereof, mining, analyzing, constructing, drawing and displaying knowledge and the mutual relation between the knowledge resources and the carriers. Each node in the knowledge graph represents an entity, and the association relationship between the entities is represented by edges between the nodes.

Any number of elements in the drawings are by way of example and not by way of limitation, and any nomenclature is used solely for differentiation and not by way of limitation.

In a specific practical process, the construction of the knowledge graph generally comprises a relatively complicated process: firstly, processing data to be processed into a data form meeting the import requirement of a graphic database by a professional according to the requirement of the graphic database; then, manually writing a complex import statement, importing the data to be processed which meets the import requirement into a graph database by using the import statement, and generating a corresponding knowledge graph by the graph database. Referring to fig. 1A, a part of files required to be prepared for creating a knowledge graph of a commercial property by using Neo4j, where each file suffixed with ". csv" represents an entity file and a relationship file, where the entity file is a file recording related information of all entities included in a certain entity category, the relationship file is a file recording relationships between entities, for example, related information of all merchants is recorded in the entity file "merchant. Referring to fig. 1B, a row of data in the entity file "merchant. The number, the name of each merchant, the category of the merchant, the exhibition hall, the floor of the merchant, the brief description of the merchant and the like. Referring to fig. 1C, each row in the relationship file "merchant _ to _ status. csv" includes a number of one merchant in the "merchant. csv" file and a number of one exhibition hall in the "status. csv" file, for example, the first row data "1, 1" indicates: the merchant HLKD which is numbered as 1 in the merchant.csv file and the exhibition hall which is numbered as 2 in the state.csv file are in an association relationship. Based on fig. 1B and fig. 1C, when the knowledge graph to be drawn includes a large number of entity types, entities, and association relationships among the entities, the number of files that a user needs to edit is large, the construction efficiency is very low, and the existing data organization method is too complex, and some illegal characters and some data loss, data errors, or format errors that are not easily perceived occur when the data is collated, which may cause a graph to fail to be constructed, but the user may hardly check out errors in the files, which further reduces the construction efficiency, and at the same time, the complex data organization method causes data to be difficult to maintain and change. After the required data is prepared, the user is required to manually write a corresponding import statement, the form of the import statement can refer to fig. 1D, which requires the user to know the use method and the data organization form of the graph database and has certain programming capability, and in addition, when the knowledge graph to be drawn contains a large number of entity types, entities and the association relations among the entities, many complicated import statements need to be written and executed, and once the written import statement is wrong, much time and effort are required to find the mistake and modify the mistake. Therefore, the existing knowledge graph construction method is easy to make mistakes in the data processing process, has high difficulty in data maintenance and modification, needs a user to learn a use method of the graph database, and has low knowledge graph creation efficiency and high use threshold.

To this end, the present application proposes a method for constructing a knowledge graph, which specifies a simple and intuitive data organization form implemented by a table, and specifically, related information of entities included in an entity category may be stored by a table, for example, a row in the table corresponds to an entity, the table may include a name column and at least one relation column, the name column includes entity names of the entities belonging to the corresponding entity category, each relation column includes an association relation between an entity in the table and an entity in some other table, and a user may add information of each entity to a corresponding cell of the corresponding table according to the above format to arrange scrambled data into a form meeting requirements. Then, the knowledge graph construction tool obtains tables corresponding to a plurality of entity categories organized by a user, respectively generates first sentences corresponding to the entity names according to the entity names extracted from the name columns of the tables, respectively generates second sentences corresponding to the association relations according to the association relations extracted from the relation series of the tables, and calls the first sentences and the second sentences generated by analyzing the graph database to obtain the knowledge graph, wherein each first sentence is used for adding entity nodes corresponding to the entity names in the graph database, and each second sentence is used for adding edges representing the association relations among the entity nodes. Therefore, a user only needs to edit the tables according to a preset format, the knowledge graph construction tool can automatically extract entity names and association relations from name columns and relation series of each table edited by the user, sentences for creating nodes and edges are automatically generated, and the knowledge graph is created by calling the graph database, so that people who are not professionals related to the knowledge graph can conveniently and efficiently create the knowledge graph, the complex data operation, the imported sentence compiling and other works during creating the knowledge graph are avoided, the formats of the tables are simple and visual, the data maintenance and modification are easy, the users who do not know the knowledge graph can quickly finish the data sorting, and the use threshold is greatly reduced.

After introducing the design concept of the embodiment of the present application, some simple descriptions are provided below for application scenarios to which the technical solution of the embodiment of the present application can be applied, and it should be noted that the application scenarios described below are only used for describing the embodiment of the present application and are not limited. In specific implementation, the technical scheme provided by the embodiment of the application can be flexibly applied according to actual needs.

Reference is made to fig. 2, which is a schematic view of an application scenario of the knowledge graph construction method provided in the embodiment of the present application. The application scenario includes a terminal device 201 and a background server 202. The terminal device 201 and the backend server 202 may be connected through a wireless or wired network, and the terminal device 201 includes but is not limited to an electronic device such as a desktop computer, a mobile phone, a mobile computer, and a tablet computer. The background server 202 may be a server, a server cluster composed of several servers, or a cloud computing center. The terminal device 201 is installed with an application program capable of editing forms, such as EXCEL, through which a user can create a form for each entity category, fill the relevant information of each entity in the corresponding form, and after completing the form corresponding to each entity category, send the form to the backend server 202 through a file upload tool provided by the backend server 202. The background server 202 is internally provided with a knowledge graph construction tool, the knowledge graph construction tool analyzes data in the table to generate an import statement for adding nodes and edges in the graph database, calls the import statement generated by analyzing the graph database to obtain a knowledge graph, and the background server 202 sends the knowledge graph to the terminal device 201 so that a user can view the generated knowledge graph through the terminal device 201.

Of course, the knowledge-graph construction tool may also be installed in the terminal device 201 to complete the construction of the knowledge-graph by the terminal device 201. After the user finishes editing the form in the terminal device 201, the user may start the knowledge graph construction tool, specify the form to be analyzed through the knowledge graph construction tool, acquire and analyze the form to be analyzed, generate an import sentence with nodes and edges added in the graph database, call the import sentence generated by analyzing the graph database, acquire the knowledge graph, and display the acquired knowledge graph through the terminal device 201.

Of course, the method provided in the embodiment of the present application is not limited to be used in the application scenario shown in fig. 2, and may also be used in other possible application scenarios, and the embodiment of the present application is not limited. The functions that can be implemented by each device in the application scenario shown in fig. 2 will be described in the following method embodiments, and will not be described in detail herein.

To further illustrate the technical solutions provided by the embodiments of the present application, the following detailed description is made with reference to the accompanying drawings and the detailed description. Although the embodiments of the present application provide the method operation steps as shown in the following embodiments or figures, more or less operation steps may be included in the method based on the conventional or non-inventive labor. In steps where no necessary causal relationship exists logically, the order of execution of the steps is not limited to that provided by the embodiments of the present application.

The following describes the technical solution provided in the embodiment of the present application with reference to the application scenario shown in fig. 2.

Referring to fig. 3, an embodiment of the present application provides a knowledge graph construction method, which is applicable to the terminal device or the background server shown in fig. 2, and specifically includes the following steps:

s301, obtaining tables corresponding to at least two entity types respectively, wherein each table comprises a name column and at least one relation series, the name column comprises entity names of entities belonging to the corresponding entity type, and each relation column comprises an association relation between the entities in the tables and the entities in any other table.

In specific implementation, one table corresponds to one entity category, that is, the related information of a plurality of entities belonging to the same entity category is stored in the same table, and a row of cells in the table is used for storing the related information of one entity. The related information of the entity at least comprises: the name of the entity and the association relationship between the entity and the entities in other entity categories, of course, the related information of the entity may be determined according to the application scenario, and may further include, for example, an alias of the entity and attribute description information, and the like, which are not limited to the above-mentioned contents.

In order to quickly and conveniently extract required information from a table, columns corresponding to each type of information are specified in the table in advance, different types of information are stored through exclusive columns, for example, a name column is used for storing entity names of entities, a relation column is used for storing association relations of the entities, and an attribute column is used for storing attribute description information corresponding to each entity name in the table, so that a user can more clearly and intuitively know data forms in the table, the user is assisted to quickly finish data entry, meanwhile, the complexity of data in the table is reduced, and corresponding information can be extracted from a specified column of the table.

It should be noted that only the association relationship between the entities of the two fixed entity categories is stored in a relationship series. For example, for the entity category a, a first column of the entity category a correspondence table may be specified for storing entity names of entities belonging to the entity category a, a second column for storing attribute description information corresponding to each entity, a third column for storing association relationships between each entity and entities in the entity category B, and a fourth column for storing association relationships between each entity and entities in the entity category C.

Taking the distribution of various merchants in a commercial property as an example, the specific format of the table can refer to fig. 4, and the entity categories to be counted in the application scenario include: merchant (Merchant), exhibition hall (state), Location (Location), business category (GFirst), and business sub-category (Genre). The table shown in fig. 4 is a table corresponding to the entity category "merchant", where relevant information of one merchant is recorded in each row of cells, the first column is a name of the merchant, the second column is an alias of the merchant, the third column is an association between the merchant and the operation subcategory, the fourth column is an association between the merchant and the exhibition hall, the fifth column is attribute description information of the merchant, and the sixth column is an association between the merchant and the location, where the first column is a name column, the third column, the fourth column, and the sixth column are relationship series, and the fifth column is an attribute column. Taking the first row of data in the table as an example, the name of the merchant is "HLKD", the alias is "KD", the belonging operation subcategory is "paint wall paint", the merchant is located in the second shop, and the specific position is the second shop floor 1.

In practical applications, the definition of each column in the table may be predetermined, and a user may also freely set each column by inputting a corresponding preset field in a first cell of each column, which is not limited in this embodiment of the present application. Taking fig. 4 as an example, the preset field corresponding to the name column may be "name", the preset field corresponding to the attribute column may be "property", the preset field corresponding to the alias column may be "synonym", and the preset field corresponding to the relationship column may be "relationship to XX", where "XX" represents an entity category having an association relationship with an entity category corresponding to a table to which the relationship column belongs, and a user may fill in the content of "XX" by himself, but XX must be a category name of the entity category, for example, the preset field of the fourth column of the table corresponding to Merchant is "relationship state", which indicates that the fourth column is used for storing an association relationship between a Merchant and a gym.

S302, according to the entity names extracted from the name columns of the tables, first sentences corresponding to the entity names are respectively generated, and each first sentence is used for adding entity nodes corresponding to the entity names in the graph database.

And S303, respectively generating second sentences corresponding to the association relations according to the association relations extracted from the relation series of the tables, wherein each second sentence is used for adding edges for representing the association relations among the entity nodes.

When the method is specifically implemented, one entity name generates a first statement correspondingly, and one incidence relation generates a second statement correspondingly. For example, 100 entity names and 200 associations are extracted in total, and then 100 first statements and 200 second statements are generated.

In specific implementation, the name column, the relationship column, the attribute column, and the like in the table can be identified according to the definition of each column in the table. For example, if the first column in the table is a name column and the second column is a relational series, the entity name is extracted from the first column and the association relationship is extracted from the second column. Of course, the name column, the relationship column, the attribute column, etc. in the table may also be identified according to the preset field.

In specific implementation, the formats of the import statements of different graph databases are different, and therefore, the formats of the first statement and the second statement may be determined according to the requirements of the graph databases, and the embodiment of the present application is not limited.

Taking neo4j as an example, determining the first column as a name column from the table merchat shown in fig. 4, extracting entity names of various merchants from various cells of the name column, where the extracted entity names include "HLKD", "YSWY", "JYLYJ", and the like, and it is necessary to create an entity node for each entity name, that is, a first statement for creating an entity node is generated for each entity name, and the generated first statement may be: creating an entity node named "HLKD" after Merchant { name: 'HLKD' }, neo4j acquires the first statement, determining that a third column is an association series from the table Merchant shown in FIG. 4, determining that the third column stores an association between the Merchant and the operation subclass according to a preset field "relationship to Genre" of the third column, and extracting the association relation from each cell of the third column, taking the first row of data in the table Merchant as an example, the first row of data is associated information corresponding to the Merchant "HLKD", and the information in the third row of the first column is "paint wall paint", wherein the "paint wall paint" is an entity node in the operation subclass (Genre), thus determining that the entity node "HLKD" and the entity node "paint wall paint" have an association relation, generating a second statement for adding an edge between the entity node "HLKD" and the entity node "paint wall paint", for example, the second statement to create an edge may be:

MATCH(a:Merchant),(b:Genre)

WHERE a.name ═ HLKD 'AND b.name ═ paint wall paint'

CREATE(a)>(b)。

After neo4j obtains the second statement, an edge is created between the entity node "HLKD" and the entity node "paint wall paint", thereby indicating that there is an association relationship between the two entity nodes.

And S304, calling the first statement and the second statement generated by the graph database analysis to obtain the knowledge graph.

When the method is specifically implemented, the background server or the terminal equipment inputs the generated first statement and the second statement into the graph database, the graph database creates entity nodes by analyzing the first statement, and creates edges among the entity nodes by analyzing the second statement, so that the creation of the knowledge graph is completed. The first statement and the second statement are both generated according to the requirement of the graph database on the import statement, the graph database is executed by the graph database according to the specific process of creating the knowledge graph according to the first statement and the second statement, the specific process is the prior art, and repeated description is omitted.

According to the knowledge graph construction method, a user only needs to input data into a form according to a preset format and upload the form to a knowledge graph construction tool in a background server or terminal equipment, the knowledge graph construction tool can automatically extract entity names and association relations from name columns and relation series of each form edited by the user, sentences for creating nodes and edges are automatically generated, a graph database is called to complete the creation of the knowledge graph, the complex data operation, the imported sentence compiling and other work during the creation of the knowledge graph are avoided, and people who are not professionals related to the knowledge graph can also create the knowledge graph conveniently and efficiently. In addition, the form format is simple and visual, the data maintenance and modification are easy, and the user who does not know the knowledge map can finish the data arrangement quickly, so that the use threshold is greatly reduced.

To assist the user in filling out and maintaining information in the form more easily, the format of the form is further defined below.

For example, the form names of the various forms used to render the same knowledge-graph are the class names of the corresponding entity classes. Still taking the distribution of various merchants in a commercial property as an example, the entity categories include: referring to fig. 4, the name of the table corresponding to the Merchant is "Merchant", the name of the table corresponding to the exhibition hall is "state", the name of the table corresponding to the Location is "Location", the name of the table corresponding to the business category is "GFirst", and the name of the table corresponding to the business sub-category is "gen". Therefore, the knowledge graph creating tool can determine the entity type corresponding to the form only according to the form name, and is convenient for a user to maintain the form.

For another example, the first cell of each name column in each table used to draw the same knowledge-graph includes a name column identifier, and the cells in each name column other than the first cell include the entity name.

For another example, the first cell of each relationship series in each table used for drawing the same knowledge graph includes the relationship column identifier and the class name of an entity class having an association relationship with the entity class corresponding to the table to which the relationship column belongs, and any cell except the first cell in each relationship series includes the entity name in the table having an association relationship with the entity corresponding to any cell and belonging to the class name in the first cell.

Referring to fig. 4 in particular, the first row of the table is used to fill in the identifier corresponding to each column, and the data type stored in each column can be quickly identified through the identifier. For example, the name column identification may be a preset field "name", the relationship column identification may be a preset field "relationTo", and further, the property column identification may be a preset field "property", and the alias column identification may be a preset field "synonym". The field between the relationship column "relationship to" and the character "_" is a class name of an entity category having an association relationship with the entity category corresponding to the form, the user can fill in the field by himself, for example, the field in the first cell in the third column of the Merchan form is "relationship to Genre _ is what operation subcategory", the third column is used for storing the association relationship between the merchant and the operation subcategory, the content after "_" is the relationship name describing the association relationship, the relationship name can be filled in freely by the user, so that the user can know the specific meaning of the association relationship shown in the third column conveniently, and here, the user can know quickly what operation subcategory the merchant is shown in the third column by the field "what operation subcategory"; for example, the field in the first cell of the fourth column of the Merchant table is "relationship to standard _ located in which shop", which indicates that the fourth column is used to store the association relationship between the Merchant and the exhibition hall, and similarly, the content after "_" can be freely filled by the user, and here, through the field "located in which shop", the user can quickly know that "located in which shop" is indicated by the fourth column. After the corresponding fields are filled in each column in the first row of the form in the manner shown in fig. 4, only the data of each merchant needs to be filled in each cell in the corresponding row according to the definition of each column.

Based on the above, the user can edit the form according to the fields and a certain agreed format, and the specific content in the form can be adjusted according to the data in different fields, so that the data sorting mode is more flexible and universal, the efficiency of sorting the data at the early stage can be improved, and the cost of modifying and correcting the data at the later stage can be improved.

Based on the above definition of the table format, the association relationship can be extracted from any relationship series of any table by the following way: aiming at any cell except the first cell in any relation series, extracting a first entity name from the cell intersected by the row where any cell is located and the column where the name column is located, extracting a second entity name from any cell, and determining that an association relation exists between an entity corresponding to the first entity name and an entity corresponding to the second entity name.

Taking the table Merchant shown in fig. 4 as an example, the first column can be determined as a name column according to the field of the first row and the first column of the table, and therefore, the entity name of each Merchant is extracted from each cell in the first column. From the fields in the first row and the second column of the table, it can be determined that the third column stores the association between the merchant and the business subclass, and the cell C for the nth row in the third column_n,3J will the cell C_n,3The name of the entity in (1) is used as the name of the second entity, and the cell C in the n-th row and the first column is used as the name of the second entity_1,3The entity name in the first sentence is used as a first entity name, the existence of the incidence relation between the entity corresponding to the first entity name and the entity corresponding to the second entity name is determined, and then a second sentence for adding the edge representing the incidence relation between the entity node corresponding to the first entity name and the entity node corresponding to the second entity name is generated according to the incidence relation. Taking the data in the first row of the table shown in fig. 4 as an example, a method for extracting the association relationship from the table is described, extracting "HLKD" from the first row and the first column as a first entity name, extracting "paint wall paint" from the third row and the third column as a second entity name, and determining that the association relationship exists between the entity corresponding to the first entity name "HLKD" and the entity corresponding to the second entity name "paint wall paint", that is, an edge needs to be created between the entity nodes "HLKD" and "paint wall paint". Based on the mode, the association relations stored in the related series are sequentially extracted.

Therefore, the knowledge graph establishing method provided by the application has good universality and reusability, and users in different industry fields can automatically establish the corresponding knowledge graph only by editing data according to the agreed simple format field, so that the establishing efficiency is improved.

Further, based on the relationship name describing the association relationship in the first cell of each relationship series in the table, the knowledge graph construction method of the embodiment of the application further includes the following steps: the relation name is extracted from the first cell of any relation series, and the relation name displayed on the side of the knowledge graph is also included in the second sentence.

Taking neo4j as an example, the information in the first cell in the third column of the table merchat shown in fig. 4 is "relationship to Genre — what business sub-category" and indicates that the third column is an association series, and according to the preset field "relationship to Genre" in the third column, the association relation between the Merchant and the business sub-category stored in the third column can be determined, and the name of the relation is "what business sub-category". Taking the first row of data in the table merchat as an example, the first row of data is related information corresponding to the Merchant "HLKD", and the information in the third column of the first row is "paint wall paint", where "paint wall paint" is one entity node in the business subclass, and therefore, it is determined that there is an association between the entity node "HLKD" and the entity node "paint wall paint", and a second statement for adding an edge between the entity node "HLKD" and the entity node "paint wall paint" is generated, for example, the second statement for creating an edge may be:

MATCH(a:Merchant),(b:Genre)

WHERE a.name ═ HLKD 'AND b.name ═ paint wall paint'

CREATE (a) - [ r: what operator category ] - > (b).

After the neo4j obtains the second statement, an edge is created between the entity node "HLKD" and the entity node "paint wall paint", and the relationship name "what business subcategory" is displayed on the created edge, so as to show the association relationship existing between the two entity nodes to the user, and the specific display manner can refer to fig. 5. Fig. 5 is a schematic partial view of a knowledge graph, each circle in fig. 5 represents an entity node, fields displayed in the circle are entity names of the entity nodes, connecting lines between the entity nodes are edges representing association relationships between the entity nodes, and content displayed on the edges is relationship names corresponding to the edges.

Based on the above definition of the table format, referring to fig. 6, the method for creating a knowledge graph according to the embodiment of the present application further includes the following steps:

s601, creating category nodes corresponding to the entity categories according to the category names of the entity categories corresponding to the tables.

In particular, the table names of the tables used for drawing the same knowledge graph are the class names of the corresponding entity classes, so that the knowledge graph creating tool can obtain the table names of the tables, and thus the class names of the entity classes corresponding to the tables are obtained. Then, a class node is created for each entity class by the knowledge-graph tool, the name of each class node being its corresponding class name. Referring to fig. 7, which shows a schematic diagram of created category nodes, each circle in fig. 7 represents a category node, and the field displayed in the circle is the class name of the category node.

S602, aiming at any relation series of any table, extracting a class name from a first cell of any relation series, and determining that an association relation exists between an entity category corresponding to the extracted class name and an entity category corresponding to any table.

In specific implementation, taking the merchat table shown in fig. 4 as an example, the class name of the entity category corresponding to the table is merchat, and a field belonging to the class name is extracted from a field of the first unit of the relationship column, where a field between "relationTo" and a character "_" is the class name to be extracted, that is, it is determined that the association relationship is stored between merchat and the entity category corresponding to the class name extracted from the relationship column. Taking the third column in fig. 4 as an example, the field "gene" between the relationship list in the first cell of the third column "relationTo" and the character "_" is obtained, so as to determine that there is an association relationship between the entity category merchat and the entity category gene. Taking the fourth column in fig. 4 as an example, a field "standard" between "relationship to" and "character" _ "indicated by the relationship column in the first cell in the fourth column is obtained, so as to determine that an association relationship exists between the entity category Merchant and the entity category standard. For the other relationship columns in the Merchant table shown in fig. 4 and the relationship columns in the tables corresponding to the other entity categories, the association relationship between the entity categories may also be determined by the above method, and details are not repeated.

S603, creating edges among the category nodes with the association relation to obtain a Schema diagram (Schema) for representing the association relation among the entity categories.

In specific implementation, according to the association relationship determined in step S602, an edge is created between the category nodes having the association relationship, so as to connect the category nodes having the association relationship together to obtain a corresponding schema diagram, so that a user can clearly and intuitively know the relationship between the entity categories through the schema diagram.

In particular, the edges between the nodes of the connection category may be directed edges. At this time, the association relationship extracted from the table is a directed association relationship, that is, step S602 specifically includes: for any relation series of any table, extracting a class name from a first cell of any relation series, and determining the association relation between an entity class A corresponding to the extracted class name and an entity class B corresponding to any table as follows: a points to B. Correspondingly, when creating an edge between the category nodes, a directed edge is created between the entity node corresponding to the entity category a and the category node corresponding to the entity category B, and the direction of the directed edge is: a points to B, namely, corresponding directed edges are created among the class nodes according to the directed association relation extracted from the table.

In specific implementation, the relationship name may be extracted from the relationship column, and the relationship name is displayed on a corresponding edge in the schema, where a field after the character "_" in the first cell of the relationship series is the relationship name.

Fig. 7 is a schema diagram created based on the table shown in fig. 4, and based on the table shown in fig. 4, it is possible to determine the following directional association relationships, (1) merchat points to gent, and the corresponding relationship name is "what business subcategory", (2) merchat points to stable, and the corresponding relationship name is "which museum" there is, and (3) merchat points to Location, and the corresponding relationship name is "where" there is, and (4) Location points to stable, and the corresponding relationship name is "which museum" there is, and (5) re points to Location, and the corresponding relationship name is "at", and (6) gent points to GFirst, and the corresponding relationship name is "which business category" belongs to, where the corresponding association relationships of (4), (5), and (6) are not shown in the table shown in fig. 4. And creating directed edges among the corresponding category nodes according to the determined directed association relationship, displaying corresponding relationship names on the created directed edges, and referring to the graph 7 by the obtained schema diagram.

For convenience of description, an edge created based on an association relationship in a table is referred to as a forward edge. On the basis of fig. 7, a corresponding reverse side can be created for each forward side in the schema, wherein the pointing direction of each reverse side is opposite to the pointing direction of its corresponding forward side, and each reverse side displays a reverse identifier in addition to the relationship name of its corresponding forward side for distinguishing from the forward side. For example, the reverse identifier may be the prefix character "r" shown in fig. 7 before the relationship name, the side on which the prefix character "r" is shown is the reverse side, and the side on which the prefix character "r" is not shown is the forward side.

According to the knowledge graph creating method, the Schema graph representing the association relation among the entity categories can be automatically generated according to the form, when graph data are edited, a user does not need to think about the Schema mode of the knowledge graph in advance, and only needs to fill the form according to the relation among the data, so that the workload of the user is reduced.

On the basis of any one of the above embodiments, the method for constructing a knowledge graph provided by the present application further includes the following steps: before the first statement and the second statement are generated, data in each table are checked, and if error data exist in the tables, corresponding error information is generated.

In specific implementation, the steps S302 to S304 may be executed after all the tables uploaded by the user pass verification, so as to prevent the construction failure.

In specific implementation, the knowledge graph construction tool can check data in each table, for example, check information such as file format and file size, and the tables can be analyzed only after the check is passed, and a first statement and a second statement are generated. If the uploaded file format is not the designated format, if not EXCEL, the verification is determined not to be passed, at the moment, error information prompting that the file format is wrong can be generated, and the error information is displayed to a user, so that the user can modify the uploaded file according to the error information. If the file size of the uploaded file is smaller than the minimum threshold or exceeds the maximum threshold, the verification is determined not to pass, error information prompting that the file is too small or too large can be generated at the moment, and the error information is displayed to a user, so that the user can modify the uploaded file according to the error information.

Further, data verification can be performed by the following method: and for any relation series in any table, extracting the class name in the first cell of the relation series, if the extracted class name is different from the class name of any table, determining that error data exists in the table, marking the class name in the first cell of the relation column as error data, and further marking the table as an error table. The error information generated at this time includes an error table and error data. After receiving the error information, the terminal device can display the error table and display the error data in the error table in a special state, wherein the special state can be highlighting, thickening or using a prompt box to select the error data.

Further, data verification can be performed by the following method: and aiming at any relation series in any table, determining a target table corresponding to the class name in the first cell of the relation series, if the entity name in the relation series does not appear in the corresponding target table, determining that error data exists in the table, marking the entity name which does not appear in the corresponding target table in the relation series as error data, and also marking the table as the error table. The error information generated at this time includes an error table and error data. After receiving the error information, the terminal device can display the error table and display the error data in the error table in a special state, wherein the special state can be highlighting, thickening or using a prompt box to select the error data.

In practical applications, the method for performing data verification is not limited to the above-mentioned various methods.

Furthermore, some data in the table can be converted according to an agreed conversion mode, for example, dates, numbers and the like are converted into a unified standardized format. In addition, some illegal characters appearing in the table can be replaced, for example, punctuation marks (such as pause and comma) in English format are replaced by punctuation marks in Chinese format, so that construction failure can be prevented.

On the basis of any of the above embodiments, the attribute description information corresponding to each entity name may be extracted from the attribute column of the table, and according to the entity name extracted from the name column of each table and the attribute description information corresponding to the entity name, first statements corresponding to each entity name are respectively generated, each first statement is used for adding an entity node corresponding to the entity name in the graph database, and the first statement corresponding to each entity name further includes the attribute description information corresponding to each entity name.

Taking neo4j as an example, determining a first column as a name column from the table merchat shown in fig. 4, extracting entity names of various merchants from various cells of the name column, and extracting attribute description information corresponding to the entity names from the attribute column, where the extracted entity names include "HLKD", "YSWY", "JYLYJ", and the like, an entity node needs to be created for each entity name, that is, a first statement for creating an entity node is generated for each entity name, and the generated first statement may be, for example, an entity name "HLKD": (n: Merchant { name: ' Kangao emperor ', description: ' ######}). Therefore, when the user views the knowledge graph, the user can view the attribute description information corresponding to the entity node by clicking the entity node in the knowledge graph.

On the basis of any one of the above embodiments, the method for constructing a knowledge graph provided by the present application further includes the following steps: and generating a dictionary file corresponding to the entity according to the entity name in each table name column and the entity category corresponding to the entity name.

In specific implementation, the dictionary file comprises entity names and entity categories of all entities in each table, and the dictionary file can store the entity names of all the entities in a classified manner according to the entity categories, so that classified display and query can be quickly realized through the dictionary file when subsequent application and query are facilitated.

On the basis of any of the above embodiments, each table further includes a synonym column, and the synonym column is used for storing an alias corresponding to each entity name in the table. Taking fig. 4 as an example, the second column is the alias of each merchant, that is, the second column is a synonym column, and taking the first merchant in the table as an example, the entity name of the merchant is "HLKD", and the alias is "KD".

Based on this, the knowledge graph construction method provided by the application further comprises the following steps: and generating a synonym file according to the entity name of the name column of each table and the alias corresponding to the entity name extracted from the synonym column.

In specific implementation, the entity names and the aliases of the entities extracted from the tables are stored in the synonym file as synonym association. For example, the entity name "HLKD" and its alias "KD" of the first merchant in the table shown in fig. 4 are stored as a synonym association in the synonym file, the entity name "YSWY" and its alias "YS" of the second merchant are stored as a synonym association in the synonym file, and so on. Therefore, in the subsequent query process, if the alias of the merchant is input by the user, the merchant corresponding to the input alias can be determined through the synonym file, and the information of the merchant is fed back to the user.

In particular, for the same merchant, multiple aliases may be stored in the synonym column, and the aliases may be separated by a designated symbol, such as a comma or a semicolon. When a plurality of aliases exist in the same merchant, the entity name of the merchant and the corresponding aliases are used as synonyms and are stored in a synonym file in an associated mode, and when the user inquires the synonym file, the merchant can be determined through the synonym file only by inputting any alias of the merchant, and then the information of the merchant is fed back to the user.

On the basis of any embodiment, in the process of creating a knowledge graph, the table corresponding to each entity category may be a worksheet (sheet) in the same table file, and the worksheet corresponding to each entity category is named according to the class name of the entity category. Taking fig. 4 as an example, 5 worksheets are created in the table file, the first worksheet has the name of gent and is used for storing the related information of the entities in the business sub-category, the second worksheet has the name of GFirst and is used for storing the related information of the entities in the business category, the third worksheet has the name of Location and is used for storing the related information of the entities in the Location category, the fourth worksheet has the name of state and is used for storing the related information of the entities in the exhibition category, and the fifth worksheet has the name of merchat and is used for storing the related information of the entities in the business category. When extracting the form name, extracting the name of each worksheet in the form file, thereby obtaining the class name of the entity category corresponding to each worksheet.

Therefore, a user only needs to create one form file, the information related to the knowledge graph is stored in the form file, and only one form file needs to be uploaded when data are uploaded, so that the situation that data are lost due to excessive forms is prevented, and data maintenance is more convenient and efficient.

Referring to fig. 8, an embodiment of the present application further provides a knowledge graph constructing method, which is applicable to the terminal device shown in fig. 2, and specifically includes the following steps:

s801, responding to a map building instruction input through a file uploading page, and sending a file to be uploaded to a background server, wherein the file to be uploaded comprises the following steps: the system comprises at least two tables corresponding to the entity categories respectively, each table comprises a name column and at least one relation column, the name column comprises entity names of entities belonging to the corresponding entity category, and each relation column comprises association relations between the entities in the tables and the entities in any other table.

In specific implementation, a user may create a form for each entity category through an application program, such as EXCEL, installed in the terminal device, which is capable of editing the form, and fill the relevant information of each entity in the corresponding form. After the forms corresponding to the entity types are completed, the user can be connected with the background server through a file uploading tool provided by the background server and sends the forms to the background server.

In specific implementation, the interface of the file uploading tool may refer to fig. 9A, a user selects a file to be uploaded from files stored locally in the terminal device by clicking a browse button, and clicks a submit button, the file uploading tool performs preliminary verification on the file to be uploaded, such as verification of information such as file format and file size, and generates prompt information that the verification passes after the verification passes, at this time, the user may click a map building button, and the file uploading tool sends the file to be uploaded to the background server after receiving a map building instruction. If the file format of the file to be uploaded is not the specified format, if not EXCEL, determining that the verification is not passed, generating error information for prompting the file format error at the moment, and displaying the error information to a user so that the user modifies the file format of the file to be uploaded according to the error information. If the file size of the uploaded file is smaller than the minimum threshold or exceeds the maximum threshold, the verification is determined not to be passed, at the moment, error information used for prompting that the file to be long-transmitted is too small or too large can be generated, and the error information is displayed to a user, so that the user can modify the file to be uploaded or reselect the file to be uploaded according to the error information.

After receiving a file to be uploaded, the background server respectively generates first sentences corresponding to the entity names according to the entity names extracted from the name columns of the tables, each first sentence is used for adding entity nodes corresponding to the entity names in the graph database, second sentences corresponding to the association relations are respectively generated according to the association relations extracted from the relation series of the tables, each second sentence is used for adding edges representing the association relations among the entity nodes, and calling the first sentences and the second sentences generated by analysis of the graph database to obtain the knowledge graph. For a specific embodiment, refer to steps S301 to S304, which are not described again. And the background server sends the acquired knowledge graph to the terminal equipment.

S802, displaying a knowledge graph sent by the background server, wherein the knowledge graph is a graph generated by the background server according to the file to be uploaded and used for describing the association relationship among the entities.

In specific implementation, the terminal equipment can directly display the knowledge graph after receiving the knowledge graph sent by the server. Referring to fig. 9B, the terminal device may also generate prompt information with successful construction after receiving the knowledge graph sent by the station server, and display the knowledge graph after the user clicks a view button in the prompt information. The file uploading tool can also provide a map query button for a user, and the user can directly access the background server through the map query button and obtain a successfully constructed knowledge map. The user can also select to store the knowledge graph in the local of the terminal equipment, so that the user can conveniently check the knowledge graph in the subsequent process.

Through the file uploading tool, a user can conveniently interact with the background server and check the constructed knowledge graph.

Further, the background server checks the data in the file to be uploaded, and the specific checking method is described above. The background server sends the generated error information to the terminal equipment, wherein the error information comprises an error table with error data; after receiving error information sent by the background server, the terminal equipment displays corresponding error prompt information; after the user clicks the error prompt information, the terminal equipment responds to the operation aiming at the error prompt information and displays the error table, wherein error data in the error table is displayed in a special state to prompt the user which data in the table are abnormal, so that the user is assisted to quickly locate the error data in the table and modify the error data.

Further, after the background server generates the schema, the background server may also send a message that the schema is successfully constructed to the terminal device, after the file uploading tool in the terminal device receives the message that the schema is successfully constructed, a prompt message that the schema is successfully constructed may be displayed, and the user may view the schema by clicking the prompt message. Or, the file uploading tool can also provide a schema map query button for the user, and the user can directly access the background server through the schema map query button and obtain the schema map successfully constructed. The user can also select to save the outline drawing locally on the terminal equipment, so that the subsequent viewing is convenient.

In addition, the file uploading tool can provide buttons for the user to inquire the dictionary file and the synonym file, so that the user can conveniently acquire the dictionary file and the synonym file.

It should be noted that the knowledge graph construction tool is an application program corresponding to the knowledge graph construction method in the embodiment of the present application, and may be installed in any terminal device or background server, and when the application program is run, the terminal device or the background server executes the steps included in any knowledge graph construction method provided in the embodiment of the present application.

The embodiment of the application provides a tool capable of constructing the knowledge graph by one key by editing a simple-format form, so that the data preparation threshold of non-knowledge graph professional workers in the construction of the knowledge graph in each field is greatly simplified, the non-knowledge graph professional workers can understand own data scene only by agreeing according to some simple fields, the knowledge graph required by the workers can be constructed by the knowledge graph constructing tool, and the knowledge graph constructing tool can be applied to various fields such as real estate, hotels, finance, human resources, electric power, government affairs, water conservancy and the like.

As shown in fig. 10, based on the same inventive concept as the above-mentioned method for constructing a knowledge graph, an embodiment of the present application further provides a knowledge graph constructing apparatus 100, which specifically includes: an acquisition module 1001, a first generation module 1002, a second generation module 1003 and a parsing module 1004.

The obtaining module 1001 is configured to obtain tables corresponding to at least two entity categories, where each table includes a name column and at least one relationship column, the name column includes entity names of entities belonging to the corresponding entity category, and each relationship column includes an association relationship between an entity in the table and an entity in any other table.

A first generating module 1002, configured to generate first statements corresponding to the entity names according to the entity names extracted from the name columns of the tables, where each first statement is used to add an entity node corresponding to the entity name in the graph database.

A second generating module 1003, configured to generate second statements corresponding to the association relations according to the association relations extracted from the relation series of each table, where each second statement is used to add an edge representing an association relation between entity nodes.

And the parsing module 1004 is configured to invoke the first statement and the second statement generated by the graph database parsing to obtain the knowledge graph.

Optionally, wherein the table name of each table is the class name of the corresponding entity category; the first cell of each name column in each table comprises a name column identifier, and the cells except the first cell in each name column comprise entity names; the first cell of each relation series in each table comprises a relation column identification and a class name of an entity class which has an association relation with the entity class corresponding to the table to which the relation column belongs, and any cell except the first cell in each relation series respectively comprises the entity name which has an association relation with the entity corresponding to the any cell and belongs to the table corresponding to the class name in the first cell.

Optionally, the second generating module 1003 is further configured to: extracting the association relation from any relation series of any table by the following method:

aiming at any cell except the first cell in any relation series, extracting a first entity name from the cell intersected by the row where any cell is located and the column where the name column is located, extracting a second entity name from any cell, and determining that an association relation exists between an entity corresponding to the first entity name and an entity corresponding to the second entity name.

Correspondingly, the second generating module 1003 is further configured to extract a relationship name from the first cell of any relationship series, and the second sentence further includes the relationship name displayed on the edge of the knowledge graph.

Optionally, the knowledge-graph constructing apparatus 100 according to the embodiment of the present application further includes a schema creation module, configured to:

for any relation series of any table, extracting a class name from a first cell of any relation series, and determining that an association relation exists between an entity category corresponding to the extracted class name and an entity category corresponding to any table;

Optionally, the knowledge graph constructing apparatus 100 according to the embodiment of the present application further includes a checking module, configured to check data in each table before generating the first statement and the second statement, and if it is found that there is error data in the table, generate corresponding error information.

Optionally, the knowledge graph constructing apparatus 100 according to the embodiment of the present application further includes a dictionary generating module, configured to generate a dictionary file corresponding to an entity according to the entity name in each table name column and the entity category corresponding to the entity name.

Correspondingly, the knowledge graph constructing apparatus 100 according to the embodiment of the present application further includes a synonym generating module, configured to generate a synonym file according to the entity name in the name column of each table and the alias corresponding to the entity name extracted from the synonym column.

As shown in fig. 11, based on the same inventive concept as the aforementioned method for constructing a knowledge graph, an embodiment of the present application further provides a knowledge graph constructing apparatus 110, which specifically includes an uploading module 1101 and a presentation module 1102.

The upload module 1101 is configured to respond to a map building instruction input through a file upload page, and send a file to be uploaded to a background server, where the file to be uploaded includes: the system comprises at least two tables corresponding to the entity categories respectively, each table comprises a name column and at least one relation column, the name column comprises entity names of entities belonging to the corresponding entity category, and each relation column comprises association relations between the entities in the tables and the entities in any other table.

The display module 1102 is configured to display a knowledge graph sent by the background server, where the knowledge graph is a graph generated by the background server according to a file to be uploaded and used for describing an association relationship between entities.

Optionally, the presentation module 1102 is further configured to: after error information sent by a background server is received, error prompt information is displayed, wherein the error information comprises an error table with error data; and displaying an error table in response to the operation aiming at the error prompt information, wherein the error data in the error table is displayed in a special state.

The knowledge graph construction device and the knowledge graph construction method provided by the embodiment of the application adopt the same inventive concept, can obtain the same beneficial effects, and are not repeated herein.

Based on the same inventive concept as the aforementioned knowledge graph construction method, the embodiment of the present application further provides an electronic device, which may be specifically a desktop computer, a portable computer, a smart phone, a tablet computer, a Personal Digital Assistant (PDA), a server, and the like. As shown in fig. 12, the electronic device 120 may include a processor 1201 and a memory 1202.

The Processor 1201 may be a general-purpose Processor, such as a Central Processing Unit (CPU), a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component, which may implement or perform the methods, steps, and logic blocks disclosed in the embodiments of the present Application. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware processor, or may be implemented by a combination of hardware and software modules in a processor.

Memory 1202, which is a non-volatile computer-readable storage medium, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The Memory may include at least one type of storage medium, and may include, for example, a flash Memory, a hard disk, a multimedia card, a card-type Memory, a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Programmable Read Only Memory (PROM), a Read Only Memory (ROM), a charged Erasable Programmable Read Only Memory (EEPROM), a magnetic Memory, a magnetic disk, an optical disk, and so on. The memory is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to such. The memory 1202 in the embodiments of the present application may also be circuitry or any other device capable of performing a storage function for storing program instructions and/or data.

Embodiments of the present application provide a computer-readable storage medium for storing computer program instructions for the electronic device, which includes a program for executing the method for constructing a knowledge graph.

The computer storage media may be any available media or data storage device that can be accessed by a computer, including but not limited to magnetic memory (e.g., floppy disks, hard disks, magnetic tape, magneto-optical disks (MOs), etc.), optical memory (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor memory (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND FLASH), Solid State Disks (SSDs)), etc.

The above embodiments are only used to describe the technical solutions of the present application in detail, but the above embodiments are only used to help understanding the method of the embodiments of the present application, and should not be construed as limiting the embodiments of the present application. Modifications and substitutions that may be readily apparent to those skilled in the art are intended to be included within the scope of the embodiments of the present application.

Claims

1. A knowledge graph construction method is characterized by comprising the following steps:

2. The method of claim 1, wherein the table name of each table is a class name of the corresponding entity class;

the first cell of each name column in each table comprises a name column identifier, and the cells except the first cell in each name column comprise entity names;

the first cell of each relation series in each table comprises a relation column identification and a class name of an entity class which has an association relation with the entity class corresponding to the table to which the relation column belongs, and any cell except the first cell in each relation series respectively comprises an entity name which has an association relation with the entity corresponding to the cell and belongs to the table corresponding to the class name in the first cell.

3. The method of claim 2, wherein the association is extracted from any relationship series of any table by:

4. The method of claim 3, wherein the first cell of each relationship series further comprises a relationship name describing the association relationship;

the method further comprises the following steps:

and extracting a relation name from the first cell of any relation series, wherein the second sentence further comprises the relation name displayed on the side of the knowledge graph.

5. The method of claim 2, further comprising:

6. The method of any of claims 1 to 5, further comprising:

before the first statement and the second statement are generated, data in each table are checked, and if error data exist in the tables, corresponding error information is generated.

7. The method according to any one of claims 1 to 5, wherein each table further comprises an attribute column, and the attribute column is used for storing attribute description information corresponding to each entity name in the table;

the first statement corresponding to each entity name also comprises attribute description information corresponding to each entity name.

8. The method of any of claims 1 to 5, further comprising:

and generating a dictionary file corresponding to the entity according to the entity name in each table name column and the entity category corresponding to the entity name.

9. The method according to any one of claims 1 to 5, wherein each table further comprises a synonym column, and the synonym column is used for storing the alias corresponding to each entity name in the table;

the method further comprises the following steps:

and generating a synonym file according to the entity name of the name column of each table and the alias corresponding to the entity name extracted from the synonym column.

10. The method of any one of claims 1 to 5, wherein the form corresponding to each entity category is a worksheet in the same form file.

11. A knowledge graph construction method is characterized by comprising the following steps:

12. The method of claim 11, further comprising:

13. A knowledge-graph building apparatus, comprising:

14. A knowledge-graph building apparatus, comprising:

15. A computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the steps of the method of any one of claims 1 to 12.