A method, device, vehicle and storage medium for generating chart information

Abstract

This invention provides a method, apparatus, vehicle, and storage medium for generating chart information. The method involves: determining a first data table for a first event and a second data table for a second event; sending a format command to a natural language processing model to control the model to return first response data in a data-readable format according to a data query engine; sending a database statement from the first response data to the data query engine; responding to the received database statement by obtaining multi-source query results from a multi-source database and performing an aggregation operation on the multi-source query results based on aggregation condition identifiers; and generating first chart information using the aggregated multi-source query results and the chart type from the first response data. This improves the efficiency of chart information generation by obtaining query results for different events from multiple data sources, automatically determining the chart type, and automatically generating chart information based on that chart type.

Description

Technical Field

[0001] The present invention relates to the technical field of chart information generation, and in particular to a chart information generation method, a chart information generation device, a vehicle and a computer-readable storage medium. Background Art

[0002] A relational database is a database that uses a relational model to organize data. It stores data in the form of rows and columns to facilitate user understanding. This series of rows and columns in a relational database is called a table, and a group of tables constitutes a database. If users need to view the indicator statistics returned by a relational database in the form of a chart, they need to manually write SQL language and develop corresponding programs to query data information in the form of a chart; if they want to display data information in different chart types, they need to re-release the program, which leads to low data display efficiency. In addition, when generating chart data, related technologies can only be based on the statistical results of a single data source, and cannot automatically query data from multiple data sources for target events, resulting in low query efficiency. Summary of the invention

[0003] The embodiments of the present invention provide a method, an apparatus, a vehicle and a computer-readable storage medium for generating chart information to overcome the above problems or at least partially solve the above problems.

[0004] The embodiment of the present invention discloses a method for generating chart information, which is applied to a cloud server, wherein the cloud server includes a natural language processing model, a data query engine, and a multi-source database, including:

[0005] Determine a first data table for a first event and a second data table for a second event; the first data table and the second data table have the same aggregation condition identifier;

[0006] Sending a format command to the natural language processing model; the format command is used to control the natural language processing model to return first response data for the first data table and the second data table in a data readable format of the data query engine;

[0007] Sending the database statement in the first response data to the data query engine; the data query engine is configured to, in response to receiving the database statement, obtain a multi-source query result for the database statement from the multi-source database, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier;

[0008] First chart information is generated by using the multi-source query result after the aggregation operation is performed and the chart type in the first response data.

[0009] Optionally, the data query engine has a corresponding database type and further includes:

[0010] Searching for a target field in the first data table and the second data table that has an association relationship with the database type;

[0011] Determine a processing strategy for the target field;

[0012] A target field processing command is sent to the natural language processing model to control the natural language processing model to generate a database statement for the target field based on the target field processing strategy when generating the first response data.

[0013] Optionally, it also includes:

[0014] determining a limit number of results for the multi-source query;

[0015] A quantity control command for representing the limited quantity is sent to the natural language processing model to control the natural language processing model to generate the first response data so that the first response data includes a database statement for limiting the number of generated multi-source query results.

[0016] Optionally, the cloud server is configured with a corresponding display terminal device, and the display terminal device is used to generate and send a chart generation instruction to the cloud server, further comprising:

[0017] Determining whether the chart generation instruction includes a chart type instruction;

[0018] If so, the chart type is determined using the chart type instruction.

[0019] Optionally, the step of determining the chart type by using the chart type instruction includes:

[0020] The chart type instruction is forwarded to the natural language processing model to control the natural language processing model to determine a chart type based on the chart type instruction, and to generate the first response data based on the chart type and the database statement.

[0021] Optionally, it also includes:

[0022] If it is determined that the chart generation instruction does not include a chart type instruction, the natural language processing model is controlled to determine the historical chart information corresponding to the first event and the second event, and the historical chart type corresponding to the historical chart information, and the natural language processing model is controlled to generate the second response data based on the historical chart type and the database statement; the natural language processing model is used to send the second response data to the cloud server;

[0023] The second chart information is generated using the query result and the historical chart type.

[0024] Optionally, the vehicle is used to generate a regeneration instruction for the first chart information, further comprising:

[0025] In response to receiving the regeneration instruction sent by the vehicle, controlling the natural language processing model to determine the historical chart information corresponding to the event and the historical chart type corresponding to the historical chart information; the natural language processing model is used to generate third response data based on the historical chart type and the database statement, and send the third response data to the cloud server;

[0026] The query result and the historical chart type are used to generate third chart information.

[0027] The embodiment of the present invention further discloses a chart information generating device, which is applied to a cloud server, and the cloud server includes a natural language processing model, a data query engine and a multi-source database, including:

[0028] A data table determination module, used to determine a first data table for a first event and a second data table for a second event; the first data table and the second data table have the same aggregation condition identifier;

[0029] a format command sending module, used for sending a format command to the natural language processing model; the format command is used for controlling the natural language processing model to return first response data for the first data table and the second data table in a data readable format of the data query engine;

[0030] a database statement sending module, configured to send the database statement in the first response data to the data query engine; the data query engine is configured to, in response to receiving the database statement, obtain a multi-source query result for the database statement from the multi-source database, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier;

[0031] The first chart information generating module is used to generate first chart information by using the multi-source query result after performing the aggregation operation and the chart type in the first response data.

[0032] The embodiment of the present invention further discloses a vehicle, comprising:

[0033] one or more processors;

[0034] and one or more machine-readable media having instructions stored thereon, which, when executed by the one or more processors, cause the vehicle to perform one or more of the methods described above.

[0035] The embodiment of the present invention further discloses a computer-readable storage medium having instructions stored thereon, which, when executed by one or more processors, enables the processors to execute the method described in the embodiment of the present invention.

[0036] The embodiments of the present invention include the following advantages:

[0037] According to an embodiment of the present invention, a first data table for a first event and a second data table for a second event are determined; the first data table and the second data table have the same aggregation condition identifier; a format command is sent to the natural language processing model; the format command is used to control the natural language processing model to return first response data for the first data table and the second data table in a data readable format of the data query engine; a database statement in the first response data is sent to the data query engine; the data query engine is configured to, in response to receiving the database statement, obtain a multi-source query result for the database statement from the multi-source database, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier; and first chart information is generated using the multi-source query result after the aggregation operation and the chart type in the first response data, thereby achieving the acquisition of query results for different events from multiple data sources, automatically determining the chart type, and automatically generating chart information based on the chart type, thereby improving the efficiency of generating chart information. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flowchart of a method for generating chart information provided in an embodiment of the present invention;

[0039] Figure 2 is another timing diagram of a process of generating chart information provided in an embodiment of the present invention;

[0040] Figure 3 It is a structural block diagram of a chart information generating device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0042] In order to enable those skilled in the art to better understand the embodiments of the present invention, some technical terms involved in the embodiments of the present invention are explained below.

[0043] ChatGPT4.0 is a natural language processing model that is improved and upgraded based on the GPT (Generative Pre-trained Transformer) model series. The GPT model is a language model pre-trained using unsupervised learning and can handle tasks in various natural language environments.

[0044] Compared with its predecessor ChatGPT 3.0, ChatGPT 4.0 has improved generation capabilities and conversational interactions, with better context understanding, generation control, and response quality.

[0045] MySQL is a relational database management system. In terms of web applications, MySQL can also be an RDBMS (Relational Database Management System) application software. Relational databases store data in different tables instead of putting all data in one large warehouse, which increases speed and flexibility.

[0046] SQL (Structured Query Language) is a database language with multiple functions such as data manipulation and data definition. This language is interactive and can provide great convenience for users. Database management systems should make full use of SQL language to improve the work quality and efficiency of computer application systems. SQL language can not only be used independently in terminals, but also as a sub-language to provide effective assistance for other program designs. In this program application, SQL can optimize program functions together with other program languages, thereby providing users with more comprehensive information.

[0047] The SQL language used by MySQL is the most commonly used standardized language for accessing databases.

[0048] JSON (JavaScript Object Notation) is a lightweight data exchange format. It is based on a subset of ECMAScript (the JS specification developed by the European Computer Manufacturers Association) and uses a text format that is completely independent of the programming language to store and represent data. The simplicity and clear hierarchical structure make JSON an ideal data exchange language. It is easy for people to read and write, and it is also easy for machines to parse and generate, and it effectively improves network transmission efficiency.

[0049] Trino is an open source distributed SQL query engine. The purpose is to enable fast and easy querying of large data sets distributed across one or more heterogeneous data sources. Trino can not only query data lakes stored on different storage systems and in open source columnar storage data file formats in Apache Hive or Iceberg format. Trino can also perform federated queries to query tables in different data sources. The Trino query engine mainly provides a table-based abstraction of the underlying data source through each connector. As long as the data types available to Trino can be used to represent data in tables, columns, and rows, a connector can be created and the query engine can use the data for query processing. Currently supported connectors include: Hive, Iceberg, MySQL, PostgreSQL, Oracle, SQLServer, ClickHouse, and MongoDB.

[0050] Reference Figure 1 , shows a flowchart of a method for generating chart information provided in an embodiment of the present invention, which may specifically include the following steps:

[0051] Step 101, determining a first data table for a first event and a second data table for a second event; the first data table and the second data table have the same aggregation condition identifier;

[0052] Step 102, sending a format command to the natural language processing model; the format command is used to control the natural language processing model to return first response data for the first data table and the second data table in a data readable format of the data query engine;

[0053] Step 103, sending the database statement in the first response data to the data query engine; the data query engine is configured to, in response to receiving the database statement, obtain a multi-source query result for the database statement from the multi-source database, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier;

[0054] Step 104: Generate first chart information using the multi-source query result after performing the aggregation operation and the chart type in the first response data.

[0055] In a specific implementation, the embodiments of the present invention can be applied to a cloud server. The cloud server can be configured with a corresponding display device. The display device can be used to display chart information. The display device can exchange data with the cloud server through a network. Exemplarily, the display device can be a web device used by operation and maintenance personnel, or it can be a vehicle-mounted computer. When the display device is a vehicle-mounted computer, the vehicle can be provided with a vehicle-mounted computer and a vehicle-mounted display. The vehicle can exchange data with the cloud server through the vehicle-mounted computer, and the vehicle-mounted display can be used to present data sent by the cloud server.

[0056] Taking the on-board computer as a display terminal device as an example, the on-board computer can be used to receive the user's natural language instructions. For example, the user can input a voice instruction through a sound pickup device connected to the on-board computer. When the on-board computer receives the voice instruction, the voice instruction can be converted into a text instruction through a voice recognition program loaded on the on-board computer, and the text instruction is sent to the cloud server through the on-board computer; for another example, the on-board computer is configured with a text input program, and the user can input a text instruction through an on-board display that interacts with the on-board computer for data, and the text instruction is sent to the cloud server through the on-board computer; for another example, the on-board computer is configured with a mobile terminal device capable of data interaction, and the mobile terminal device can include but is not limited to mobile phones, smart glasses, tablet computers, etc. The user can issue voice or text instructions through the mobile terminal device, the mobile terminal device can send the text instruction to the on-board computer, and the on-board computer can forward the text instruction to the cloud server. If it is a voice instruction, the mobile terminal device can first convert it into a text instruction and send the text instruction to the on-board computer. Of course, the voice instruction can also be directly sent to the on-board computer, and the on-board computer converts the voice instruction into a text instruction.

[0057] The cloud server can be configured with a corresponding natural language processing model, a data query engine and a multi-source database. Exemplarily, the natural language processing model can be ChatGPT4.0, the data query engine can be Trino, and the multi-source database can include Hive, Iceberg, MySQL, PostgreSQL, Oracle, SQLServer, ClickHouse and MongoDB databases supported by Trino. The natural language processing model can process the natural language instructions when the cloud server receives the natural language instructions sent by the display device. The data query engine can query the data corresponding to the multiple events contained in the natural language instructions from the multi-source database corresponding to the cloud server.

[0058] In actual applications, since the data query engine needs to query the data corresponding to the event in the multi-source database, the language type of the data sent to the multi-source database needs to conform to the database type of the multi-source database, and the format of the data must also be a data-readable format readable by the relational database, that is, the data query engine has the corresponding database type and data-readable format. For example, when the multi-source database includes the relational database MySQL, the database type corresponding to the data query engine can be SQL, and the data-readable format can be JSON.

[0059] Of course, the above examples are only for illustration, and those skilled in the art may determine the query results from any other database through a data query engine, and other databases may also be adapted to other corresponding database languages ​​and data readable formats, which are not limited in this embodiment of the present invention.

[0060] In actual applications, if the user wants to present the data in the form of a chart when acquiring data, the application needs to preset the chart type first to generate a chart corresponding to the data type. However, when developing the program, the R&D personnel cannot determine which chart type the user wants to present the data. Therefore, once the chart type is determined, if the user wants to change the chart type, the program needs to be re-released, which not only increases the production cost, but also fails to achieve the purpose of flexibly displaying related charts based on user input.

[0061] In a specific implementation, the embodiment of the present invention may receive a chart generation instruction sent by a display terminal device, where the chart generation instruction has a corresponding first event and a second event.

[0062] Exemplarily, a chart generation instruction may represent a user's need to generate a chart, which may be used to express the first event and the second event in the instruction. For example, when the chart generation instruction input by the user is, "Please use a line graph to represent the average daily driving time of vehicles within 7 days", the "vehicle charging status" of the first event associated with the "average daily driving time" and the "vehicle driving status" of the second event associated with the "average daily driving time" may be determined. Similarly, regional road conditions, oil reserves and oil prices at gas stations within a preset range, etc. may all be used as events.

[0063] In the embodiment of the present invention, a first data table for the first event and a second data table for the second event may also be determined, wherein the first data table is a data table corresponding to the first event, and the second data table is a data table corresponding to the second event. In a specific implementation, the data table may be presented as a DDL table, where DDL is the full name of Data Definition Languages, which is used to define database objects (databases, tables, fields); and finally, the DML language, which is used to add, delete, and modify data records in tables in the database.

[0064] Exemplarily, assuming that the chart generation instruction sent by the user includes a first event "charging status of the vehicle" and a second event "driving status of the vehicle", the DDL table may include a first data table and a second data table, wherein the first data table may be represented by Table 1 and the second data table may be represented by Table 2.

[0065] Table 1:

[0066]

[0067]

[0068] Table 2:

[0069] Fields type describe vin String 'Vehicle VIN number' start_time String 'Start time, string, format: yyyy-MM-dd' end_time String 'End time' trip_score String Driving score syscode String 'Organization code' update_time DateTime 'Data update time'

[0070] The cloud server of an embodiment of the present invention can send format commands to the natural language processing model. The format commands can be presented in natural language. The format commands can be used to control the natural language processing model to return first response data for the data table in a data-readable format based on the natural language processing model proxy interface configured on the cloud server, and the first response data may include database statements and chart types that conform to the database type.

[0071] For example, the format command can be "You are a data analyst, the answer is converted to JSON format and returned, and make sure your answer complies with Trino's SQL specification." By sending the format command to the natural language processing model GTP, the JSON format template returned by the natural language processing model GTP can be defined to facilitate logical processing in the program. The comments after the JSON field can define the behavior of the natural language processing model GPT. At the same time, the returned content contains SQL statements.

[0072] Exemplarily, the rules for returning the first response data in JSON format are as follows:

[0073] {

[0074] "text":"", / / Generated text content. If the user enters instructions such as generate SQL, display SQL, and show SQL text, the SQL text will be concatenated with the existing text content in the form of a code block, but the SQL field and its content need to be retained!

[0075] "sql":"", / / Generated SQL statement

[0076] "chartType":"", / / Input chart type, line chart is represented by line, pie chart is represented by pie, bar chart is represented by bar, table is represented by table, and heat map is represented by heatmap;

[0077] "sqlComments":{"":""} / / Use each selected field in the sql field as the key, and wrap the comment of the field as the value into the field

[0078] }

[0079] Assume that Trino has configured two data sources, namely postgresql and clickhouse, where the scheme of clickhouse defaults and the scheme of postgresql defaults to postgresql. If the chart generation instruction is "please use a line chart to represent the average daily driving time of all vehicles in 7 days", the server can respond to receiving the chart generation instruction, determine the "vehicle charging status" of the first event associated with the "average daily driving time" and the second event "vehicle driving status" associated with the "average daily driving time", and send the first data table (Table 1) corresponding to the first event, the second data table (Table 2) corresponding to the second event and the format command to the natural language processing model. After the natural language processing model processes the data, the first response data returned can be:

[0080] {

[0081] "text":"The following is the query result of the average daily driving time of all vehicles in the past 7 days:",

[0082] "sql":"SELECT t1.<colum_1> , t2.<colum_3> FROM(SELECT<colum_1> ,<colum_2> FROM<catalog_1> .<schema_1> .<table_1> )AS t1JOIN(SELECT<colum_3> ,<colum_4> FROM<catalog_2> .<schema_2> .<table_2> )AS t2 ON t1.id=t2.id;",

[0083] "chartType":"line",

[0084] "sqlComments":{

[0085] "date":"date",

[0086] "average_daily_driving_duration":"Average daily driving duration"

[0087] }

[0088] }

[0089] The SQL statement includes: The SQL statement generated by Trino is different from the ordinary SQL statement. The condition after FROM is different. <colum>For specific field names, placeholders <catalog>Indicates the database type (MySQL, postgresql, etc.), placeholder <schema>Generally, it is public and needs to be dynamically spliced ​​according to the actual situation. Placeholder

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096] is the table name; the generated SQL should not use * to represent all columns, but list the specific column field names; "chartType":"line" can be used to express that the chart type is "line chart". It should be noted that the natural language processing model determines the chart type through the information in the chart generation instruction, or infers the chart type through calculation based on the event type, or determines the chart type based on the user's historical behavior. This is not limited in the embodiments of the present invention. After the natural language processing model GPT4.0 generates the first response data and returns the first response data to the cloud server, the cloud server can send the database statement in the first response data to the data query engine Trino. The data query engine Trino can execute the SQL query statement through the corresponding multi-source database connector connector, determine the multi-source query results from the multi-source database, and perform aggregation operations on the multi-source query results based on the aggregation condition identifier. For example, both Table 1 and Table 2 contain "Vehicle VIN Code". Assuming that the vehicle VIN code is used as the aggregation condition identifier, the result of using the VIN code as the aggregation condition is shown in Table 3. Table 3: Fields type describe vin String 'Vehicle VIN number' start_time String 'Start time, string, format: yyyy-MM-dd' end_time String 'End time' trip_score String Driving score direction String 'Charging direction, in or out' charge_type String 'Charging type, fast or slow' start_soc Float64 'Start SOC' end_soc Float64 'End of SOC' syscode String 'Organization code' update_time DateTime 'Data update time'

[0097] After Table 3 is determined, the required data items can be counted from different databases based on Table 3 and finally displayed as relevant charts.

[0098] The multi-source query result after executing the aggregation operation is returned to the cloud server, and the cloud server can generate first chart information based on the multi-source query result after executing the aggregation operation and the chart type, and send the first chart information to the display terminal device; the display terminal device can be configured to display the first chart information in response to receiving the first chart information, and the first chart information includes at least graphic and text data for the first event and the second event.

[0099] According to an embodiment of the present invention, a first data table for a first event and a second data table for a second event are determined; the first data table and the second data table have the same aggregation condition identifier; a format command is sent to the natural language processing model; the format command is used to control the natural language processing model to return first response data for the first data table and the second data table in a data readable format of the data query engine; a database statement in the first response data is sent to the data query engine; the data query engine is configured to, in response to receiving the database statement, obtain a multi-source query result for the database statement from the multi-source database, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier; first chart information is generated using the multi-source query result after the aggregation operation and the chart type in the first response data, thereby achieving the acquisition of query results for different events from multiple data sources, automatically determining the chart type, and automatically generating chart information based on the chart type, thereby improving the efficiency of generating chart information.

[0100] Based on the above embodiment, a variant embodiment of the above embodiment is proposed. It should be noted that in order to make the description concise, only the differences from the above embodiment are described in the variant embodiment.

[0101] In an optional embodiment of the present invention, the data query engine has a corresponding database type and further includes:

[0102] Searching for a target field in the first data table and the second data table that has an association relationship with the database type;

[0103] Determine a processing strategy for the target field;

[0104] A target field processing command is sent to the natural language processing model to control the natural language processing model to generate a database statement for the target field based on the target field processing strategy when generating the first response data.

[0105] In practical applications, before the natural language processing model outputs the results, defining the processing method for specific sub-fields of the DDL table structure can effectively improve the accuracy of the final generated SQL statement.

[0106] Exemplarily, assuming that the chart generation instruction sent by the user includes the first event "charging status of the vehicle" and the second event "driving status of the vehicle", the DDL table may include a first data table and a second data table, wherein the first data table can be represented by Table 1 and the second data table can be represented by Table 2.

[0107] Table 1:

[0108] Fields type describe vin String 'Vehicle VIN number' start_time String 'Start time, string, format: yyyy-MM-dd' end_time String 'End time' direction String 'Charging direction, in or out' charge_type String 'Charging type, fast or slow' start_soc Float64 'Start SOC' end_soc Float64 'End of SOC' dt String 'Date string, format: yyyy-MM-dd' syscode String 'Organization code' update_time DateTime 'Data update time'

[0109] Table 2:

[0110]

[0111]

[0112] The target field in the first data table and the second data table that is associated with the database type is the start_time field; before the natural language processing model outputs the response data for the charging event table, a processing strategy can be determined for the start_time field. When the start_time field is processed as a certain time, the toDateTime() function needs to be called first, and then the toHour() function needs to be called, and a target field processing command is generated based on the strategy. The target field processing command can be a text used to represent the strategy. After receiving the processing strategy for the start_time field, the natural language processing model can first call the toDateTime() function, and then call the toHour() function to generate an SQL statement for the start_time field.

[0113] Of course, the above examples are only for illustration, and those skilled in the art may determine any other field associated with the database type as the target field, and the embodiments of the present invention are not limited to this.

[0114] In an embodiment of the present invention, target fields that are associated with the database type are searched in the first data table and the second data table; a processing strategy for the target field is determined; and a target field processing command is sent to the natural language processing model to control the natural language processing model to generate a database statement for the target field based on the target field processing strategy when generating the first response data, thereby defining a processing method for fields in the data table structure that are associated with the database type, and effectively improving the accuracy of the final generated database statement.

[0115] In an optional embodiment of the present invention, the statement format type for the database statement can also be determined; and a statement format type control command for characterizing the statement format type is sent to the natural language processing model to control the natural language processing model to generate a database statement based on the statement format type when generating the first response data.

[0116] Exemplarily, the format type control command is, when the alias in the SQL statement is Chinese, using the symbol (`) to wrap the SQL statement, and the natural language processing model generates a database statement based on the format type control command.

[0117] The embodiment of the present invention determines the statement format type for the database statement; sends a statement format type control command for characterizing the statement format type to the natural language processing model to control the natural language processing model to generate the database statement based on the statement format type when generating the first response data, thereby defining the format of the database statement containing Chinese in the response data returned by the natural language processing model, and effectively improving the accuracy of the final generated database statement.

[0118] In an optional embodiment of the present invention, it also includes:

[0119] determining a limit number of results for the multi-source query;

[0120] A quantity control command for representing the limited quantity is sent to the natural language processing model to control the natural language processing model to generate the first response data so that the first response data includes a database statement for limiting the number of generated multi-source query results.

[0121] Exemplarily, assuming that there is a need to limit the number of SQL query results generated from multiple data sources, that is, the generated SQL query results include at most 100 related records, and the quantity control command is a natural statement used to characterize "the generated SQL query results include at most 100 related records". The natural language processing model will generate an SQL statement based on the quantity control command to characterize "controlling multi-source query results within 100".

[0122] Of course, the above examples are only for illustration, and those skilled in the art may use any number to limit the number of generated database statements according to actual conditions, and the embodiments of the present invention are not limited to this.

[0123] According to an embodiment of the present invention, a limit number for the multi-source query results is determined; a quantity control command for representing the limit number is sent to the natural language processing model to control the natural language processing model when generating the first response data, so that the first response data includes a database statement for limiting the number of generated multi-source query results, thereby achieving control over the number of multi-source query results, thereby reducing the amount of computation required for the generation process of the first chart information and improving the efficiency of generating the first chart information.

[0124] In an optional embodiment of the present invention, the cloud server is configured with a corresponding display terminal device, and the display terminal device is used to generate and send a chart generation instruction to the cloud server, further comprising:

[0125] Determining whether the chart generation instruction includes a chart type instruction;

[0126] If so, the chart type is determined using the chart type instruction.

[0127] In a specific implementation, a user may generate a chart generation instruction through a display terminal device of an embodiment of the present invention, and the chart generation instruction may be sent to a cloud server through the display terminal device. After receiving the chart generation instruction, the cloud server may determine whether the chart generation instruction includes a chart type instruction. If it is determined that the chart generation instruction includes a chart type instruction, it may be considered that the user has determined the chart type, and therefore, the chart type instruction may be used to determine the chart type.

[0128] In an optional embodiment of the present invention, the step of using the chart type instruction to determine the chart type includes:

[0129] The chart type instruction is forwarded to the natural language processing model to control the natural language processing model to determine a chart type based on the chart type instruction, and to generate the first response data based on the chart type and the database statement.

[0130] In a specific implementation, when a user issues a chart generation instruction through a display device, the chart generation instruction includes a chart type instruction to control the natural language processing model to determine the chart type based on the chart type instruction, and then generate first response data based on the chart type and the database statement, and send the first response data to the cloud server. The cloud server can then generate first chart information based on the chart type, and forward the first chart information to the display device so that the first chart information is presented in the display device, thereby making the chart type of the first chart information correspond to the chart type instruction issued by the user, thereby improving the efficiency of generating the first chart information while also improving the user experience.

[0131] For example, assume that Trino has configured two data sources, namely postgresql and clickhouse, where the scheme of clickhouse defaults and the scheme of postgresql defaults to postgresql. If the chart generation instruction is "please use a line chart to represent the average daily driving time of all vehicles in 7 days", the server can respond to receiving the chart generation instruction, determine the "vehicle charging status" of the first event associated with the "average daily driving time" and the second event "vehicle driving status" associated with the "average daily driving time", and send the first data table (Table 1) corresponding to the first event, the second data table (Table 2) corresponding to the second event and the format command to the natural language processing model. After the natural language processing model processes the data, the first response data returned can be:

[0132] {

[0133] "text":"The following is the query result of the average daily driving time of all vehicles in the past 7 days:",

[0134] "sql":"SELECT t1.<colum_1> , t2.<colum_3> FROM(SELECT<colum_1> ,<colum_2> FROM<catalog_1> .<schema_1> .<table_1> )AS t1 JOIN(SELECT<colum_3> ,<colum_4> FROM<catalog_2> .<schema_2> .<table_2> )AS t2 ON t1.id=t2.id;",

[0135] "chartType":"line",

[0136] "sqlComments":{

[0137] "date":"date",

[0138] "average_daily_driving_duration":"Average daily driving duration"

[0139] }

[0140] }

[0141] The SQL statement includes: The SQL statement generated by Trino is different from the ordinary SQL statement. The condition after FROM is different. <colum>For specific field names, placeholders <catalog>Indicates the database type (MySQL, postgresql, etc.), placeholder <schema>Generally, it is public and needs to be dynamically spliced ​​according to the actual situation. Placeholder

[0142]

[0143]

[0144]

[0145]

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152]

[0153]

[0154]

[0155]

[0156]

[0157]

[0158]

[0159]

[0160]

[0161]

[0162]

[0163]

[0164]

[0165]

[0166]

[0167]

[0168]

[0169] Figure 2 Figure 2

[0170]

[0171]

[0172]

[0173]

[0174]

[0175] is the table name; the generated SQL should not use * to represent all columns, but needs to list the specific column field names; "chartType":"line", can be used to express that the chart type is "line chart". For the chart type instruction of "line chart", it is inferred that adapting to the line chart is preferred, so the first response data may include the SQL statement "chartType":"line". In an optional embodiment of the present invention, if it is determined that the chart generation instruction does not contain a chart type instruction, the natural language processing model is controlled to determine the historical chart information corresponding to the first event and the second event, and the historical chart type corresponding to the historical chart information, and the natural language processing model is controlled to generate the second response data based on the historical chart type and the database statement; the natural language processing model is used to send the second response data to the cloud server; the second chart information is generated using the query result and the historical chart type. In a specific implementation, after receiving a chart generation instruction, the cloud server can determine whether the chart generation instruction contains a chart type instruction. If it is determined that the chart generation instruction does not contain a chart type instruction, it can be considered that the user has not selected a chart type. At this time, the natural language processing model can be controlled to determine the historical chart information corresponding to the first event and the second event, and the historical chart type corresponding to the historical chart information, and the natural language processing model can be controlled to generate second response data based on the historical chart type and the database statement; the natural language processing model can also send the second response data to the cloud server after generating the second response data; the cloud server can use the query results and the historical chart type to generate the second chart information, thereby achieving that even if the user has not selected a chart type, the cloud server can generate the second chart information based on the user's historical behavior, thereby further improving the efficiency of generating chart information. In an optional embodiment of the present invention, the vehicle is used to generate a regeneration instruction for the first chart information, and also includes: in response to receiving the regeneration instruction sent by the vehicle, controlling the natural language processing model to determine the historical chart information corresponding to the first event and the second event, and the historical chart type corresponding to the historical chart information; the natural language processing model is used to generate third response data based on the historical chart type and the database statement, and send the third response data to the cloud server; generate third chart information using the query result and the historical chart type; send the third chart information to the display terminal device; the display terminal device is configured to present the third chart information.In actual applications, the user may not determine the chart type in the chart generation instruction, so the natural language processing model can determine the chart type for different events based on the preset reasoning capability. The first response data corresponding to the chart type may include chart information for expressing the chart type. A line chart is represented by line, a pie chart is represented by pie, a column is represented by bar, a table is represented by table, and a heat map is represented by heatmap. For example, assume that Trino has configured two data sources, namely postgresql and clickhouse, where the scheme of clickhouse defaults and the scheme of postgresql defaults to postgresql. If the chart generation instruction is "please use a graph to represent the average daily driving time of all vehicles in the past 7 days", the server can respond to receiving the chart generation instruction, determine the "vehicle charging status" of the first event associated with the "average daily driving time" and the second event "vehicle driving status" associated with the "average daily driving time", and send the first data table (Table 1) corresponding to the first event, the second data table (Table 2) corresponding to the second event and the format command to the natural language processing model. The chart generation instruction only specifies the use of a chart, but does not specify the chart type. However, after the natural language processing model processes the data, the first response data returned may be: {"text":"The following is the query result of the average daily driving time of all vehicles in the past 7 days:", "sql":"SELECT t1.<colum_1> , t2.<colum_3> FROM(SELECT<colum_1> ,<colum_2> FROM<catalog_1> .<schema_1> .<table_1> )AS t1 JOIN(SELECT<colum_3> ,<colum_4> FROM<catalog_2> .<schema_2> .<table_2> )AS t2 ON t1.id=t2.id;","chartType":"line","sqlComments":{"date":"Date","average_daily_driving_duration":"Daily average driving time"}}It can be seen from the first response data that the natural language processing model GPT4.0 can, based on its own reasoning ability, infer that adapting the line chart is the preferred option for the first event "vehicle charging status" and the second event "vehicle driving status". Therefore, the first response data may include the SQL statement "chartType":"line". Even if the user does not specify the chart type in the chart generation instruction, the cloud server can still generate the first chart information based on the query result and the chart type, and send the first chart information to the display device, so that the display device presents the first chart information to the user.However, in actual applications, the user may be dissatisfied with the presentation of the first chart information. For example, the user believes that it is unreasonable to use a line graph to express the average daily driving time of all vehicles in 7 days. In this case, after the first chart information is presented on the vehicle display, a regeneration instruction for the first chart information can be generated by the display device. The display device can then send the regeneration instruction to the cloud server. The cloud server can be configured to, in response to receiving the regeneration instruction sent by the display device, control the natural language processing model to determine the historical chart information corresponding to the first event and the second event. The historical chart information can be chart information generated by the user for the event in the past history. The information may have a corresponding historical chart type, and the historical chart type may be used to express the user's response to the event. In the past history, what chart type was used to express the chart information corresponding to the event. For example, the current first chart information uses a line chart to display the average daily driving time, while the historical chart information uses a bar chart to represent the average daily driving time. The natural language processing model can be used to generate third response data based on the bar chart and the database statement, and send the third response data to the cloud server; the server can use the historical chart type (bar chart) and the database statement to generate the third chart information, and send the third chart information to the display terminal device so that the display terminal device can present the third chart information. It should be noted that the third response data is different from the first response data and the second response data. It is the response data generated for the regeneration instruction. The third response data may include a historical chart type and a database statement; and the third chart information is the chart information generated based on the historical chart type. Of course, the above examples are only examples. The historical chart information can be determined according to the actual situation, and may not be limited to line charts, pie charts, bars, tables, heat maps, etc., and the embodiments of the present invention are not limited to this. In an embodiment of the present invention, in response to receiving the regeneration instruction sent by the vehicle, the natural language processing model is controlled to determine the historical chart information corresponding to the first event and the second event, and the historical chart type corresponding to the historical chart information; the natural language processing model is used to generate third response data based on the historical chart type and the database statement, and send the third response data to the cloud server; the third chart information is generated using the query result and the historical chart type, which not only improves the efficiency of generating chart information, but also improves the user experience of viewing chart information. In order to enable those skilled in the art to better understand the embodiment of the present invention, a complete example is used to illustrate the embodiment of the present invention. Reference is another timing diagram of the process of expressing chart information generation provided in the embodiment of the present invention. The specific implementation method includes the following steps: S1. Pre-configure the corresponding GTP4.0 related promote to build the cornerstone of the key functions of the system.(a). Use Markdown syntax to describe field information and reduce token consumption when calling the GPT4 interface. In order to improve the GPT recognition rate, the comment must be detailed and semantically clear. Example: Now database 1 has the following data table, charging event table (Table 1) Table 1:. Fields type describe vin String 'Vehicle VIN number' start_time String 'Start time, string, format: yyyy-MM-dd' end_time String 'End time' direction String 'Charging direction, in or out' charge_type String 'Charging type, fast or slow' start_soc Float64 'Start SOC' end_soc Float64 'End of SOC' dt String 'Date string, format: yyyy-MM-dd' syscode String 'Organization code' update_time DateTime 'Data update time'

[0176] Database 2 has the following data tables: Driving event table (Table 2)

[0177] Table 2:

[0178] Fields type describe vin String 'Vehicle VIN number' start_time String 'Start time, string, format: yyyy-MM-dd' end_time String 'End time' trip_score String Driving score syscode String 'Organization code' update_time DateTime 'Data update time'

[0179] Then, assuming that the result of using VIN code as the aggregation condition is as shown in Table 3 below, the required data items can be counted from different databases and finally displayed as relevant charts.

[0180] Table 3:

[0181] Fields type describe vin String 'Vehicle VIN number' start_time String 'Start time, string, format: yyyy-MM-dd' end_time String 'End time' trip_score String Driving score direction String 'Charging direction, in or out' charge_type String 'Charging type, fast or slow' start_soc Float64 'Start SOC' end_soc Float64 'End of SOC' syscode String 'Organization code' update_time DateTime 'Data update time'

[0182] (b) Pre-configure the role of GPT, agree on the return of GPT, and meet relevant specifications

[0183] Example: You are a SQL data analyst, and the answer is converted to JSON format and returned to ensure that your answer complies with Trino's SQL specifications;

[0184] (c) Trino has configured two data sources, namely postgresql and clickhouse. The default scheme of clickhouse is default, and the default scheme of postgresql is postgresql.

[0185] (d) Define the inference process of the chart type, which determines the final chart type displayed on the front-end page.

[0186] Example 1: The returned results for the predefined chart types are: line for line chart, pie for pie chart, bar for bar chart, table for table, and heatmap for heatmap.

[0187] Example 2: When the user enters a new answer, infer the corresponding chart type based on the user's last input; if the chart type is not explicitly entered, automatically infer the most suitable chart type based on the input;

[0188] (e) Define the JSON format template returned by GPT to facilitate logical processing in the program. The comments after the JSON field can define the behavior of GPT. Try to make the semantics clear and the description simple.

[0189] Example: JSON format for reply text processing, the rules are as follows

[0190] {

[0191] "text":"", / / Generated text content. If the user enters instructions such as generate SQL, display SQL, and show SQL text, the SQL text will be concatenated with the existing text content in the form of a code block, but the SQL field and its content need to be retained!

[0192] "sql":"", / / Generated SQL statement

[0193] "chartType":"", / / Input chart type, line chart is represented by line, pie chart is represented by pie, bar chart is represented by bar, table is represented by table, and heat map is represented by heatmap;

[0194] "sqlComments":{"":""} / / Use each selected field in the sql field as the key, and wrap the comment of the field as the value into the field

[0195] }

[0196] (f) GPT returns an example result, which can make the result more accurate and controllable.

[0197] Example:

[0198] If the user inputs: Please use a line graph to represent the average daily driving time of all vehicles in the past 7 days.

[0199] Then the JSON format data you need to be returned is:

[0200] {

[0201] "text":"The following is the query result of the average daily driving time of all vehicles in the past 7 days:",

[0202] "sql":"SELECT t1.<colum_1> , t2.<colum_3> FROM(SELECT<colum_1> ,<colum_2> FROM<catalog_1> .<schema_1> .<table_1> )AS t1 JOIN(SELECT<colum_3> ,<colum_4> FROM<catalog_2> .<schema_2> .<table_2> )AS t2 ON t1.id=t2.id;",

[0203] "chartType":"bar",

[0204] "sqlComments":{

[0205] "date":"date",

[0206] "average_daily_driving_duration":"Average daily driving duration"

[0207] }

[0208] }

[0209] For the SQL statement in (f): The conditions after FROM in the generated Trino SQL statement are different from those in the ordinary SQL statement. <colum>For specific field names, placeholders <catalog>Indicates the database type (MySQL, postgresql, etc.), placeholder <schema>Generally, it is public and needs to be dynamically spliced ​​according to the actual situation. Placeholder

[0210]

[0211]

[0212]

[0213]

[0214]

[0215]

[0216]

[0217]

[0218]

[0219]

[0220]

[0221]

[0222]

[0223] Figure 3

[0224]

[0225]

[0226]

[0227]

[0228]

[0229]

[0230]

[0231]

[0232]

[0233]

[0234]

[0235]

[0236]

[0237]

[0238]

[0239]

[0240] is the table name; the generated SQL should not use * to represent all columns, but should list the specific column field names. S2. When querying the table field information, query the database type and the scheme information at the same time. S3. Pass the database type and scheme information of S2 to the GPT-4.0 interface to obtain the returned JSON data. S4. Send the returned SQL statement to the Trino query engine, and the connector of the database corresponding to Trino will execute the SQL query statement, and then Trino will aggregate the query results of multiple data sources and return them to the backend service. S5. Assemble the corresponding data format according to the data result set returned by S3 and the chart type returned by S2. S6. Return the corresponding JSON final result to the web display end. By generating chart information in the above way, the following beneficial effects are achieved. 1. Good flexibility. This method can generate charts of different types and styles according to user input and needs, such as bar charts, line charts, pie charts, scatter charts, etc. Users can use simple natural language descriptions or provide data tables to let GPT-4 help generate appropriate SQL statements. Users can adjust the chart details in real time, and GPT-4 dynamically generates SQL statements to update feedback immediately. 2. Strong adaptability. Corresponding services can be provided according to different time periods to meet different needs. 3. Costs are saved, and there is no need for separate development and release for different types of chart types. 4. Simplify prompt descriptions to reduce token consumption when calling the GPT4 interface. Multiple data sources can be queried, and statistical data for BI reports can be obtained across different business systems. It should be noted that for the method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, certain steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.Referring to, a structural block diagram of a chart information generating device provided in an embodiment of the present invention is shown, which may specifically include the following modules: a data table determining module 301, used to determine a first data table for a first event and a second data table for a second event; the first data table and the second data table have the same aggregation condition identifier; a format command sending module 302, used to send a format command to the natural language processing model; the format command is used to control the natural language processing model to return the first response data for the first data table and the second data table according to the data readable format of the data query engine; a database statement sending module 303, used to send the database statement in the first response data to the data query engine; the data query engine is configured to obtain a multi-source query result for the database statement from the multi-source database in response to receiving the database statement, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier; a first chart information generating module 304, used to generate first chart information using the multi-source query result after the aggregation operation and the chart type in the first response data. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The embodiment of the present invention further provides a vehicle, comprising: one or more processors; and one or more machine-readable media having instructions stored thereon, when executed by the one or more processors, the vehicle executes the method described in the embodiment of the present invention. The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by the processor, each process of the above-mentioned embodiment of the method for generating chart information is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. Among them, the computer-readable storage medium is such as a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are only illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in the field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, all of which belong to the protection of the present invention. Ordinary technicians in the field can realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the embodiments of the present invention can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. It can be clearly understood by those skilled in the art that for the convenience and simplicity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the embodiments provided in the present application, it should be understood that the disclosed device and method can be implemented in other ways. For example, the device embodiment described above is only schematic, for example, the division of the unit is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, each functional unit in each embodiment of the present invention can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, ROM, RAM, disk or optical disk and other media that can store program code. The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims. < / schema> < / catalog> < / colum> < / schema> < / catalog> < / colum> < / schema> < / catalog> < / colum>

Claims

1. A method for generating chart information, characterized in that: The method is applied to a cloud server, which includes a natural language processing model, a data query engine, and a multi-source database, including: Determine a first data table for a first event and a second data table for a second event; the first data table and the second data table have the same aggregation condition identifier; Sending a format command to the natural language processing model; the format command is used to control the natural language processing model to return first response data for the first data table and the second data table in a data readable format of the data query engine; the data query engine has a corresponding database type; Searching for a target field in the first data table and the second data table that has an association relationship with the database type; Determine a processing strategy for the target field; Sending a target field processing command to the natural language processing model to control the natural language processing model to generate a database statement for the target field based on the target field processing strategy when generating the first response data; Sending the database statement in the first response data to the data query engine; the data query engine is configured to, in response to receiving the database statement, obtain a multi-source query result for the database statement from the multi-source database, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier; First chart information is generated by using the multi-source query result after the aggregation operation is performed and the chart type in the first response data.

2. The method according to claim 1, characterized in that Also includes: determining a limit number of results for the multi-source query; A quantity control command for representing the limited quantity is sent to the natural language processing model to control the natural language processing model to generate the first response data so that the first response data includes a database statement for limiting the number of generated multi-source query results.

3. The method according to any one of claims 1 to 2, characterized in that: The cloud server is configured with a corresponding display terminal device, and the display terminal device is used to generate and send a chart generation instruction to the cloud server, and further includes: Determining whether the chart generation instruction includes a chart type instruction; If so, the chart type is determined using the chart type instruction.

4. The method according to claim 3, characterized in that The step of using the chart type instruction to determine the chart type comprises: The chart type instruction is forwarded to the natural language processing model to control the natural language processing model to determine a chart type based on the chart type instruction, and to generate the first response data based on the chart type and the database statement.

5. The method according to claim 3, characterized in that: Also includes: If it is determined that the chart generation instruction does not include a chart type instruction, controlling the natural language processing model to determine the historical chart information corresponding to the first event and the second event, and the historical chart type corresponding to the historical chart information, and controlling the natural language processing model to generate second response data based on the historical chart type and the database statement; The natural language processing model is used to send the second response data to the cloud server; The second chart information is generated using the query result and the historical chart type.

6. The method according to any one of claims 1-2, characterized in that: The vehicle is used to generate a regeneration instruction for the first chart information, further comprising: In response to receiving the regeneration instruction sent by the vehicle, controlling the natural language processing model to determine the historical chart information corresponding to the first event and the second event, and the historical chart type corresponding to the historical chart information; the natural language processing model is used to generate third response data based on the historical chart type and the database statement, and send the third response data to the cloud server; The query result and the historical chart type are used to generate third chart information.

7. A diagram information generating device, characterized in that: The device is applied to a cloud server, which includes a natural language processing model, a data query engine and a multi-source database, including: A data table determination module, used to determine a first data table for a first event and a second data table for a second event; the first data table and the second data table have the same aggregation condition identifier; a format command sending module, used for sending a format command to the natural language processing model; the format command is used for controlling the natural language processing model to return first response data for the first data table and the second data table in a data readable format of the data query engine; the data query engine has a corresponding database type; A module for searching the target field in the first data table and the second data table that is associated with the database type; determining a processing strategy for the target field; and sending a target field processing command to the natural language processing model to control the natural language processing model to generate a database statement for the target field based on the target field processing strategy when generating the first response data; a database statement sending module, configured to send the database statement in the first response data to the data query engine; the data query engine is configured to, in response to receiving the database statement, obtain a multi-source query result for the database statement from the multi-source database, and perform an aggregation operation on the multi-source query result based on the aggregation condition identifier; The first chart information generating module is used to generate first chart information by using the multi-source query result after performing the aggregation operation and the chart type in the first response data.

8. A vehicle, characterized in that: include: one or more processors; and One or more machine-readable media having instructions stored thereon, when executed by the one or more processors, cause the vehicle to perform the method of any one of claims 1-6.

9. A computer-readable storage medium having instructions stored thereon, which, when executed by one or more processors, cause the processors to perform the method according to any one of claims 1 to 6.

Images