A data textual task execution method

Abstract

The present invention relates to a method for executing a data textualization task, the method comprising: step S31: reading an original simulation data file; step S32: reading the original simulation data file based on a data reading template; step S33: parsing the read data based on a data parsing file; and step S34: filling a textual information file according to the textualization template. The present invention adopts the form of custom rules, combined with professional knowledge of the power system, to analyze data and generate short sentence texts, aggregate similar sentences, and organize the aggregated sentences according to a template to generate a final readable text, forming a complete generation process from data to knowledge text. It also breaks through the limitations of previous text generation systems in which fixed calculations and text generation rely on templates, making the text generation system easier to maintain and the generated final text more rich in form and more professional and comprehensive in content.

Description

Technical field

[0001] The present invention belongs to the technical field of data texting, and in particular relates to a method for executing a data texting task. [Background Technology]

[0002] Power system simulations generate large amounts of real-time operational data and result data. The data structure is complex, making it difficult to directly observe and draw conclusions. Data textualization technology provides a means of compressing and concentrating the large amounts of data generated in power system simulations. It also enables text mining from the textualized information, allowing for multiple uses of the data. Furthermore, this data textualization method simplifies the repetitive work of analysts searching for information within the data, saving valuable time. Therefore, data textualization, converting data information into textual information in a clear and unambiguous form, has become a new approach and a new approach to simulation analysis. This invention uses custom rules, combined with power system expertise, to analyze data and generate short sentence texts. It aggregates similar sentences and organizes these aggregated sentences according to templates to generate a final, readable text, forming a complete process for generating knowledge text from data. This overcomes the limitations of previous text generation systems, which relied on rigid calculations and template-dependent text generation, making the text generation system easier to maintain and generating a richer, more professional, and comprehensive text. [Summary of the invention]

[0003] In order to solve the above problems in the prior art, the present invention proposes a data text task execution method, which includes:

[0004] Step S31: reading the original simulation data file;

[0005] Step S32: reading the original simulation data file based on the data reading template;

[0006] Step S33: parsing the read data based on the data parsing file;

[0007] Step S34: Fill in the textual information file according to the textual template.

[0008] Furthermore, the original simulation data file is a power flow simulation calculation data file.

[0009] Furthermore, the power flow simulation calculation data files are stored in a calculation example folder according to different electrical components, and the content of a single file is the simulation calculation data of a type of electrical components.

[0010] Furthermore, the content of each file is in the form of a two-dimensional table, wherein the rows of the two-dimensional table represent simulation data generated by a certain physical quantity of an electrical component within a unit simulation time, and the columns of the two-dimensional table represent observable physical quantities of such components.

[0011] Furthermore, the power flow simulation calculation data files have the characteristic of being associated between files, and the data values ​​of a certain row and column in the first file can be associated with the data values ​​of a certain row and column in the second file.

[0012] Furthermore, the two associated data values ​​are equal.

[0013] Furthermore, the reading of the original simulation data file based on the data reading template is specifically as follows: in view of the characteristics of the concept object model, the template items are filled in the data reading template based on the correspondence between the template items and the concept object attributes.

[0014] Furthermore, after the data reading of the template is completed, the data texting task reads all the call functions required for template parsing to perform subsequent template parsing.

[0015] Furthermore, the parsing of the read data based on the data parsing file is specifically as follows: a cyclic call is made to the template parsing function, during each call, the template parsing function reads the data parsing file to parse the concept objects and their horizontal relationships; all template items are parsed through cyclic calls until all template items are parsed.

[0016] Furthermore, the textual information file is located on a distributed storage device, so as to be applicable to the big data environment of current power system simulation calculations.

[0017] The beneficial effects of this invention include: using custom rules, combined with power system expertise, to analyze data and generate short sentence texts, aggregating similar sentences, and organizing these aggregated sentences according to templates to generate readable text, thus forming a complete process for generating knowledge text from data. This overcomes the limitations of previous text generation systems, which rely on fixed calculations and templates for text generation, making the text generation system easier to maintain and generating a richer, more professional, and comprehensive text. It also avoids the reliance of current artificial intelligence algorithms on data volume, allowing conclusions to be drawn even when data volume is insufficient.

Brief Description of the Drawings

[0018] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application, but do not constitute an improper limitation of the present invention. In the drawings:

[0019] Figure 1 It is a schematic diagram of the data parsing template of the present invention.

[0020] Figure 2 It is a schematic diagram of a data reading template of the present invention.

[0021] Figure 3 It is a timing diagram of the textual task execution process of the present invention.

[0022] Figure 4 It is a schematic diagram of the calling relationship during the execution of the textual task of the present invention. [Specific implementation method]

[0023] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The exemplary embodiments and descriptions are only used to explain the present invention but are not intended to limit the present invention.

[0024] like Figure 1 As shown, a data text task execution method applied by the present invention is described in detail;

[0025] The simulation data textualization is performed based on the logical relationships between conceptual objects, relations, and entities. The data textualization task is regarded as an entity, called E, and the entity is assumed to exist. The entity is composed of several facts, which are derived from a priori judgments or logical reasoning. There is no relationship between facts, and facts are just a description or list.

[0026] Facts are composed of conceptual objects and the relationships between them; conceptual objects contain one or more conceptual object attributes; conceptual objects can transcend facts and become the constituent elements of different facts, but cannot be used to directly constitute entities, but only exist based on facts; attributes are used to describe the state of conceptual objects;

[0027] A fact is a collection of one or more conceptual objects. A collection of facts constitutes an entity. Facts are recorded as:

[0028] E={Fact1{01, 02, 03...}, Fact2{01, 02, 03...}...} (1);

[0029] Preferably: the relationship is a horizontal relationship between conceptual objects; the horizontal relationship includes a logical operation relationship, a numerical operation relationship, a self-defined function relationship, etc.; assuming that the conceptual object is x n , n is a positive integer greater than 1, then about x n The horizontal relationship expression can be written as:

[0030]

[0031]

[0032] For example, for the textualization of power flow calculation data, a power flow calculation data textualization task is an entity. Regardless of the content of the entity, the textualization task will exist inherently. The entity only outputs factual statements, that is, statements whose logical values ​​(including probabilistic truth values) in horizontal relationships are true. The statements are the basis of facts and are regular expressions set by the system user, which conform to formal logic. For the textualization of power flow calculation data, due to its large data volume and complex structure, it is necessary to use data reading templates to construct conceptual objects.

[0033] Preferred: Integrate expert experience in horizontal relationship construction, convert numerical data into knowledge text, and use data parsing templates to build horizontal relationships;

[0034] The textualization task first requires the construction of a textual model, which is then maintained and operated using software design. The textual template serves as the carrier of entities in the textual model. The data parsing template serves as the carrier of facts, and horizontal relationships can be set in the data parsing template. For power system simulation data, the key to building a textual model is the construction of the textual template, data parsing template, and data reading template.

[0035] The data textualization task execution method specifically includes the following steps:

[0036] Step S1: constructing a conceptual model for text generation; setting the corresponding relationships among data reading template, data parsing template, and textualization template based on the logical relationships among conceptual objects, relationships, and entities;

[0037] By setting the corresponding relationship between the data reading template, data parsing template, and textualization template, the attributes in the concept object, the relationship between the attributes, and the relationship between the template items are constructed. In the subsequent textualization process for a specific simulation data file, it is necessary to instantiate the above template files according to the characteristics of the simulation data file.

[0038] Step S11: constructing a conceptual object, specifically: setting a conceptual object for the electrical components targeted by the power system simulation data, and constructing a data reading template based on the conceptual object, using the data reading template as a carrier of the conceptual object; it can be seen that the purpose of constructing the conceptual object is achieved by constructing the data reading template;

[0039] Preferably: the data reading template includes several key items, and the key items correspond to several attributes of the concept object;

[0040] In the process of data textualization, the data textualization task is an entity, which is composed of several facts. Facts are derived from a priori judgments or logical reasoning. There is no relationship between facts. Facts are just a description or list.

[0041] Preferred: Concept objects that need to be constructed include: busbars, AC lines, DC lines, generators, transformers, loads and other electrical components;

[0042] Preferably, the attributes of the conceptual object include: rows and columns of electrical component related data, read data set, unique identifier of data set, data cross-row reading, and data association reading;

[0043] Preferably: according to the characteristics of the concept object model, several template items are set in the data reading template to correspond to the concept object attributes; the specific corresponding relationship is shown in the following table;

[0044] Template Item A Concept object name Concept object attributes Concept object name Template Item B Result Set Concept object attributes The data set read Template Item C Primary Key Concept object attributes Unique identifier for a data set Template Item D Reading step Concept object attributes Reading data across rows Template Item E Outer Join Concept object attributes Data association reading

[0045] Step S12: constructing horizontal relationships between conceptual objects; specifically, selecting referenced conceptual objects, performing relationship analysis on the conceptual objects, and filling in the data analysis template based on the relationship analysis results;

[0046] Horizontal relationship construction is the extraction of data knowledge. The knowledge in the data mainly exists in the association relationship between data. Therefore, horizontal relationships include the extraction of numerical relationships, logical relationships, equation relationships, or custom relationships between data.

[0047] Preferred: Integrate experts' experience in analyzing data relationships, conduct data mining on knowledge texts formed based on horizontal relationships, and use the results of data mining to construct horizontal relationships;

[0048] Preferably, the horizontal relations between conceptual objects include logical operation relations (HRL), numerical operation relations (HRN), and user-defined function relations (HRF).

[0049] Preferably: multiple horizontal relationships can be set in one parsing template, and concept objects can be called in the horizontal relationships;

[0050] The correspondence between template items and several items of horizontal relationships is shown in the following table;

[0051]

[0052] Step S13: constructing entities of the data textualization task, specifically: constructing entities according to the attributes of the data textualization task, and constructing a textualization template accordingly;

[0053] The entity construction according to the attributes of the data textualization task is specifically as follows: setting a set of facts contained in the entity corresponding to the data textualization task; that is, a set consisting of a set of concept objects and horizontal relationships between the concept objects; constructing template items of the textualization template based on the entity, the correspondence between the template items and the custom information and the horizontal relationships; and achieving the purpose of outputting a textual report by constructing the textualization template;

[0054] The entity is the entire content of the report document. The construction of the entity model is to determine the content set of the document output. According to the design of the textual model, the content set is the truth value item of all horizontal relationships. The design of the entity model tends to be flexible in the report document and convenient in defining the relationship set. The entity model includes: custom information and horizontal relationship calls.

[0055] Preferably: the text template includes custom information of text entities and horizontal relationship calls;

[0056] The correspondence between textual template items and several items of entity model is shown in the following table;

[0057]

[0058] Step S14: generating a data text information file; specifically, initializing the data text information file, and saving the data text related information based on the text template in the subsequent text process;

[0059] The initialization data text information file specifically includes: creating a new text information file, and setting the size of the text information file according to the size of the original simulation data file;

[0060] The data textualization information file is a data file with a data structure that supports text mining, provided by the data textualization prototype system. To support text mining, the data structure of the information file will record all the relevant information when generating each factual statement in the next textualization task. At the same time, it provides a storage format based on factual statements to facilitate searches based on factual statements.

[0061] Step S2: preparing the simulation calculation data for data textualization; specifically, setting the data reading template, data parsing template, and textualization template;

[0062] The data text preparation for the simulation calculation data includes the following steps:

[0063] Step S21: setting the name of the horizontal relationship group and the name of the conceptual object referenced in the horizontal relationship in the data parsing template;

[0064] Preferably: the horizontal relationship group name and the referenced concept object name are both in English characters;

[0065] Step S22: setting the horizontal relationship name in the data parsing template;

[0066] Preferably: horizontal relationship names are in English characters;

[0067] Step S23: setting the horizontal relationship description in the data parsing template;

[0068] Preferably: the horizontal relationship description can be Chinese characters, English characters, numbers, etc.;

[0069] Step S24: setting a horizontal relationship expression in the data parsing template;

[0070] Preferred: Horizontal relational expressions support logical expressions, function names, etc. Among them, logical expressions support numerical operations and logical relational operations;

[0071] Step S215 sets the concept object name in the data reading template;

[0072] Preferred: Concept object names are in English characters;

[0073] Step S26: setting the concept object description field in the data reading template;

[0074] Preferred: The description of the concept object can be in Chinese characters, English, numbers and other characters;

[0075] Step S27: Set the file name of the conceptual object data source and the name of the returned result set in the data reading template;

[0076] Step S28: setting the primary key of the conceptual object dataset in the data reading template;

[0077] Preferably: the primary key is the column number of the conceptual object data file;

[0078] Step S29: setting the associated read data of the conceptual object in the data read template;

[0079] Preferably: according to the unique structure of the simulation data, the association between the conceptual objects A and B, and between the data content of the specified row and column of object A and the row number of object B can be set;

[0080] Step S210: configuring and calling the set horizontal relationship name in the text template, and filling in custom information as needed;

[0081] Step S3: Instantiate the data textualization task; specifically, read the original simulation data file, read the original simulation data file based on the data reading template, parse the read data based on the data parsing file, and fill the textual information file according to the textualization template;

[0082] Preferably, the original simulation data file is a power flow simulation calculation data file; the power flow simulation calculation data file is stored in a calculation example folder according to different electrical components, and the content of a single file is the simulation calculation data of a class of electrical components; the content of each file is in the form of a two-dimensional table, the rows of the two-dimensional table represent the simulation data generated by a certain physical quantity of the electrical component within a unit simulation time, and the columns of the two-dimensional table represent the observable physical quantities of this class of components; in addition, the data format can be varied in that it is possible to present the observable physical quantities of the electrical components and the simulation data per unit time in multiple rows;

[0083] Preferably: the power flow simulation calculation data file has the feature of file association, which can associate the data value of a row and column in the first file with the data value of a row and column in the second file; the two associated data values ​​are equal; for example: in the data of electrical component A, a row and column L a C a The value is associated with a row and column L in the B electrical component data b C b The data value can be simply written as A(L a C a )=B(L b C b );

[0084] The reading of the original simulation data file based on the data reading template is specifically as follows: in view of the characteristics of the concept object model, the template items are filled in the data reading template based on the corresponding relationship between the template items and the concept object attributes;

[0085] After the data reading template is completed, the data text task reads all the call functions required for template parsing for subsequent template parsing;

[0086] The parsing of the read data based on the data parsing file is specifically as follows: a template parsing function is called cyclically, and in each calling process, the template parsing function reads the data parsing file to parse the concept objects and their horizontal relationships; all template items are parsed by cyclic calling until all template items are parsed;

[0087] The filling of the text information file according to the text template is specifically: filling the text information file according to the text template with the data analysis result;

[0088] Preferably, after initiating the execution of the textual task, initializing the textual information file;

[0089] Preferably, the text information file is located on a distributed storage device, so as to be applicable to the big data environment of current power system simulation calculation;

[0090] The prepared data "36-node case data" is used as the original data file, and the folder size is 2.66M, which contains data description files and data result files such as busbar, AC line, transformer, DC line, generator, and load; in the process of texting the data, you need to first configure the reading template, parsing the last shift, and texting the template. Then, select "Scenario Management" on the texting system interface, create a new scenario "Job 2", select the corresponding template, and import the original data file. Click "Analyze" on the main interface to get the results of this analysis; check the local project folder "Project Path Name" + "resultData" to see the temporary files generated according to the rules. The temporary file size is 50.26KB;

[0091] Optimized: Data texting is considered to be completed using distributed devices; for example, it runs on two node machines, achieving the experimental purpose of multi-machine distributed processing; at the same time, the configurable characteristics of the generation template and generation process are realized in the text generation process, and only about 2M of data is used to generate relatively accurate text cases, meeting the design requirements and goals;

[0092] In the several embodiments provided by the present invention, it should be understood that the disclosed methods and terminals can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the module division is merely a logical function division, and other division methods may be used in actual implementation.

[0093] In addition, the technical solutions in the above embodiments can be combined and replaced with each other if no contradiction occurs.

[0094] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0095] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0096] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the invention is defined by the appended claims rather than the above description, and it is intended that all changes that fall within the meaning and scope of the equivalent elements of the claims be included in the present invention. Any accompanying figure marks in the claims should not be regarded as limiting the claims involved. In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple modules or devices stated in the system claims may also be implemented by one module or device through software or hardware. Words such as first, second, etc. are used to indicate names and do not indicate any particular order.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A data textualization task execution method, characterized in that: The method comprises: Step S31: reading the original simulation data file; Step S32: reading the original simulation data file based on the data reading template; Step S33: parsing the read data based on the data parsing file; Step S34: filling the textual information file according to the textual template; The parsing of the read data based on the data parsing file is specifically as follows: a template parsing function is called cyclically, and in each calling process, the template parsing function reads the data parsing file to parse the concept objects and their horizontal relationships; all template items are parsed by cyclic calling until all template items are parsed; The text information file is located on a distributed storage device; it is suitable for the big data environment of current power system simulation calculation; The original simulation data file is a power flow simulation calculation data file; The power flow simulation calculation data files are stored in a calculation example folder according to different electrical components, and the content of a single file is the simulation calculation data of a type of electrical components; The content of each file is in the form of a two-dimensional table. The rows of the two-dimensional table represent the simulation data generated by a physical quantity of the electrical component in unit simulation time, and the columns of the two-dimensional table represent the observable physical quantities of such components. The data files of power flow simulation calculation have the characteristic of file association. The data value of a row or column in the first file can be associated with the data value of a row or column in the second file. The two data values ​​associated are equal.

2. The data textualization task execution method according to claim 1, characterized in that: The reading of the original simulation data file based on the data reading template specifically includes: filling in the template items in the data reading template based on the corresponding relationship between the template items and the concept object attributes according to the characteristics of the concept object model.

3. The data textualization task execution method according to claim 2, characterized in that: After the data reading template is completed, the data text task reads all the call functions required for template parsing for subsequent template parsing.

Images