Production data processing method and device, storage medium and program product

Abstract

The invention provides a production data processing method and device, a storage medium and a program product, and is applied to the technical field of computers. The method comprises the steps of calling a large model to obtain target report information based on an obtained to-be-replied question; the large model is used for extracting entity information from a to-be-replied question, performing recall based on the entity information to obtain a recall result, and searching target report information based on the recall result; obtaining target production index information associated with the target report information; determining target production index data corresponding to the target production index information based on the index data; the index data is obtained by processing production data in a plurality of data sources; and generating a report according to the target report information and the target production index data, and outputting the report as a reply of the to-be-replied question. According to the method, the technical effect of improving the efficiency of generating the complex report is achieved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a production data processing method, device, storage medium and program product. Background Art

[0002] The report includes charts and tables. The report can include production data from all aspects, and the production situation can be fully and intuitively reflected through the report.

[0003] In the related technology, professionals are required to summarize, analyze and calculate various data involved in production and then draw them to obtain a report, which results in low efficiency in generating reports. Summary of the invention

[0004] The present application provides a production data processing method, device, storage medium and program product to achieve the technical effect of improving the efficiency of generating complex production data reports.

[0005] In a first aspect, the present application provides a production data processing method, comprising:

[0006] Based on the obtained questions to be answered, the big model is called to obtain the target report information; the big model is used to extract entity information from the questions to be answered, recall based on the entity information, obtain the recall results, and search for the target report information based on the recall results; obtain the target production indicator information associated with the target report information; determine the target production indicator data corresponding to the target production indicator information based on the production indicator data; the production indicator data is obtained by processing the production data in multiple data sources; generate a production data report based on the target report information and the target production indicator data, and output the production data report as a response to the questions to be answered.

[0007] In a second aspect, the present application provides a production data processing device, comprising:

[0008] The recall module is used to call the big model to obtain the target report information based on the questions to be answered. The big model is used to extract entity information from the questions to be answered, perform recall based on the entity information, obtain the recall result, and search for the target report information based on the recall result.

[0009] An acquisition module, used for acquiring target production index information associated with target report information;

[0010] A determination module, used to determine target production index data corresponding to target production index information based on production index data; the production index data is obtained by processing production data from multiple data sources;

[0011] The output module is used to generate a production data report based on the target report information and the target production indicator data, and output the production data report as a response to the question to be answered.

[0012] In a third aspect, the present application further provides a computer device, the computer comprising: a processor, and a memory communicatively connected to the processor;

[0013] The memory stores computer-executable instructions;

[0014] The above method is implemented when the processor executes the computer-executable instructions stored in the memory.

[0015] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to implement the above method when executed by a processor.

[0016] In a fifth aspect, the present application also provides a computer program product, comprising computer execution instructions, which implement the above method when executed by a processor.

[0017] The present application provides a production data processing method, device, computer equipment, storage medium and program product. The method includes: through the powerful processing power of the big model, the entity extraction of the question to be answered is carried out, and the recall is carried out based on the extracted entity information, so that the target report information that matches the question to be answered can be obtained, and the quality of the target report information is improved; because the target report information is associated with the target production indicator information, the target production indicator information required for generating the report can be obtained based on the target report information, and the target production indicator data corresponding to the target production indicator information can be obtained from the production indicator data, that is, the target report information is recalled through the big model, and the associated target production indicator information is subsequently determined, and then the target production indicator data is obtained based on the target production indicator information. In this way, the big model does not need to directly process a large number of The production data is analyzed and processed, which reduces the amount of data and calculation involved in obtaining the target production index data, and greatly improves the efficiency of obtaining the target production index data; since the production index data is obtained by processing the production data of multiple data sources, the target production index data can come from different data sources, and there is no need to obtain the production data required for generating reports from different data sources respectively, which improves the convenience of obtaining data from different data sources and also improves the efficiency of finding the target production index data; in addition, the above process does not require human participation, and can automatically process the questions to be answered and output reports that are consistent with the questions to be answered, effectively improving the efficiency of generating complex reports. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0019] Figure 1 A schematic diagram of an application scenario of a production data processing method provided in an embodiment;

[0020] Figure 2 A flowchart of a production data processing method provided in one embodiment;

[0021] Figure 3 A flowchart for constructing a report template in one embodiment;

[0022] Figure 4 A flowchart of saving production index information, production index data and index generation rules to an index service platform in one embodiment;

[0023] Figure 5 A schematic diagram of a process of querying a production data report through an intelligent question-answering task in one embodiment;

[0024] Figure 6 A schematic diagram of the structure of a production data processing device provided in one embodiment;

[0025] Figure 7 A hardware structure diagram of a computer device provided in one embodiment.

[0026] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0027] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0028] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0029] The production data processing method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the terminal 102 communicates with the server 104 through a network. The data storage system can store data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers.

[0030] Terminal 102 can call the big model to obtain target report information based on the obtained questions to be answered; the big model is used to extract entity information from the questions to be answered, recall based on the entity information, obtain recall results, and search for target report information based on the recall results. Terminal 102 can obtain target production indicator information associated with the target report information; terminal 102 can determine the target production indicator data corresponding to the target production indicator information based on the indicator data; the indicator data is obtained by processing production data from multiple data sources; terminal 102 can generate a report based on the target report information and the target production indicator data, and output the report as a response to the questions to be answered.

[0031] The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0032] In some embodiments, Figure 2 As shown, a production data processing method is provided, which can be applied to a system composed of a terminal and a server, or to a terminal. Figure 1 The following is an example of a terminal in the example, including:

[0033] S202, based on the obtained questions to be answered, call the big model to obtain the target report information; the big model is used to extract entity information from the questions to be answered, perform recall based on the entity information, obtain the recall result, and search for the target report information based on the recall result.

[0034] Among them, the questions to be answered are questions input by the user; in actual applications, the terminal can display an intelligent question and answer page, which includes a question input control. The user can input the questions to be answered through the question input control, and the terminal obtains the questions to be answered in the question input control; in this embodiment of the application, the questions to be answered can be questions used to query reports.

[0035] The big model is a machine learning model with a large number of parameters and computing resources. The big model shows strong processing capabilities in processing large-scale data and complex tasks. In the embodiment of the present application, the big model can be used to process complex question-and-answer tasks related to report generation. In practical applications, the big model can be implemented based on the GPT model. After fine-tuning the GPT model according to the report generation task, the big model used to process complex question-and-answer tasks related to report generation in the embodiment of the present application can be obtained. GPT (Generative Pre-trained Transformer) is a natural language processing model based on deep learning. The big model can also be implemented based on the Zhimei Master big model. After fine-tuning the Zhimei Master big model according to the report generation task, the big model used to process complex question-and-answer tasks related to report generation in the embodiment of the present application can be obtained.

[0036] Entity information may include but is not limited to: reporting entity, indicator entity and time period entity; for example, the question to be answered is: please output the production report of commodity A in month M, then the entity information may include: commodity A, production report and month M, among which commodity A belongs to the indicator entity, the production report is the reporting entity, and month M is the time period entity.

[0037] The target report information is information of a report used to reply to a question to be replied; the target report information may include a name of the report and may also include a description of the report; recall refers to an information retrieval operation.

[0038] The recall result may include the first report information, or may be empty (ie, no information is recalled); the first report information may be information of a report related to the question to be answered.

[0039] In some embodiments, extracting entity information from the question to be answered includes: segmenting the question to be answered based on a production terminology vocabulary to obtain segmented words to be processed; and extracting entities based on the segmented words to be processed to obtain entity information.

[0040] The production terminology lexicon is pre-set, and the production terms in the production terminology lexicon may include proper nouns related to production data, and may also include industry jargon.

[0041] Specifically, the big model performs word segmentation on the question to be answered based on the production terminology vocabulary to obtain the word segmentation to be processed; when the question to be answered includes production terminology, the production terminology can be identified, and then the word segmentation to be processed includes the production terminology; when the question to be answered does not include production terminology, the word segmentation to be processed is obtained by word segmentation based on general field knowledge; the big model performs entity extraction based on the word segmentation to be processed to obtain entity information; when the word segmentation to be processed includes production terminology, the entity information can be the entity information corresponding to the production.

[0042] It should be noted that if the questions to be answered are not segmented based on the production terminology lexicon, the production terms related to the report in the questions to be answered may be divided into different words based on the general knowledge segmentation, resulting in a low correlation between the recall results and the meaning of the questions to be answered; for example, the questions to be answered include the daily production report of product A, where "production daily report" is a production term, and the entity information corresponding to the "production daily report" can be recalled later; if the words are not segmented based on the production terminology lexicon, two segmentations of "production" and "daily report" may be obtained, and the recall will be carried out based on the entity information corresponding to "production" and "daily report", resulting in the recall results being inconsistent with the questions to be answered.

[0043] In the above embodiment, the question to be answered is segmented based on the production terminology lexicon, so that the extracted entity information meets the user's intention, thereby improving the relevance between the subsequently determined recall result and the question to be answered.

[0044] Specifically, the terminal can obtain questions to be answered input by the user. For example, the terminal displays an intelligent question and answer page, and intelligently obtains questions to be answered input by the user through the intelligent question and answer page. After obtaining the questions to be answered, the big model is called to input the questions to be answered into the big model. The big model is used to extract entities from the questions to be answered, and at least one entity information among the report entity, indicator entity or time period entity is obtained. The big model is used to perform recall based on the extracted entity information to obtain a recall result.

[0045] In one possible implementation, when the recall result includes the first report information and the first report information does not meet the report generation conditions, the large model can further search for target report information based on the first report information; in this implementation, the relevance of the target report information to the question to be answered is higher than the relevance of the first report information to the question to be answered; it should be noted that searching for target report information based on the recall result is also a recall process.

[0046] In another possible implementation, the recall result includes the first report information, and when the first report information meets the report generation condition, the first report information can be used as the target report information.

[0047] In another possible implementation, the recall result is empty, and the large model can perform a vector-level search based on the question to be answered to obtain the target report information.

[0048] It should be noted that the large model recalls based on the extracted entity information, which is a recall process; searching for the target report information based on the recall results may not require a recall process, or may include at least one recall process. Therefore, the process of determining the target report information based on the questions to be answered includes at least one recall.

[0049] S204, obtaining target production indicator information associated with the target report information.

[0050] Among them, the target production indicator information includes the name of the production indicator and may also include a description of the production indicator; the production indicator corresponding to the target production indicator information may be a component of the report corresponding to the target report information; for example, if the report corresponding to the target report information is a report form, then the production indicator corresponding to the target production indicator information may be an attribute in the report form.

[0051] Specifically, after the terminal obtains the target report information, it can search for the target production indicator information associated with the target report information in the relational database.

[0052] S206, determining target production index data corresponding to the target production index information based on the production index data; the production index data is obtained by processing production data from multiple data sources.

[0053] Among them, the production index data is used to reflect the data of various indicators of production; illustratively, various indicators of production can be raw material cost, output, yield, yield rate, sales volume, etc.

[0054] Multiple data sources are different data sources. In actual applications, in order to meet production needs, it is often necessary to set up multiple different data sources to collect data on multiple aspects of production. For example, in the production of a certain product, the raw material data source collects the relevant data for generating the required raw materials, the output data source collects the output data for producing a certain product, and the order data source collects the relevant data for selling a certain product.

[0055] Production indicator data is obtained by integrating, cleaning and converting production data from multiple data sources, that is, production indicator data is integrated and standardized production data. When querying production-related data, you can directly conduct a comprehensive query through the production indicator data without having to enter different data sources for query separately; when analyzing production-related data, you can also directly conduct in-depth analysis of production data from different data sources through production indicator data, without having to pull production data from different data sources and then analyze it through complex analysis rules; therefore, processing production data from different data sources into indicator data breaks the data silos formed by multiple different data sources, making it more convenient to query and analyze production data, and improving the accuracy and efficiency of querying and analyzing production data.

[0056] Specifically, the production index data can be stored in the index service platform; the terminal can obtain the target production index data from the index service platform according to the target production index information, and the terminal can also obtain the production index data corresponding to the target production index information from the index service platform, and then calculate the production index data to obtain the target production index data.

[0057] S208, generating a production data report based on the target report information and the target production index data, and outputting the production data report as a response to the question to be answered.

[0058] Specifically, the terminal can obtain a target report template corresponding to the target report information, and the target report template includes at least one production indicator attribute. For each production indicator attribute, the terminal obtains the target production indicator sub-data corresponding to the production indicator attribute in the target production indicator data, generates a production data report according to the target report template, the indicator attribute and the corresponding target production indicator sub-data, and outputs the production data report as a response to the question to be answered.

[0059] In some embodiments, after obtaining the input question to be answered, it also includes: displaying a report type selection window; obtaining the report type based on the report type selection window; the report type includes a report type or a chart type; generating a production data report based on the target report information and the target production indicator data, including: when the report type is a report type, generating a production data report based on the target report and the target production indicator data corresponding to the target report information; when the report type is a chart type, generating a production data chart based on the target report and the target production indicator data corresponding to the target report information.

[0060] Specifically, when the terminal obtains the questions to be answered input by the user through the smart question and answer page, the terminal can display a report type selection window in the smart question and answer page, and the report type selection window can be displayed in the smart question and answer page in the form of a floating page.

[0061] The user selects a report type in the report type selection window. When the terminal generates a production data report, it can obtain a target report template belonging to the report type according to the target report information. For example, if the report type is a report type, the target report template corresponding to the target report information can be obtained. For another example, if the report type is a chart type, the target chart template corresponding to the target report information can be obtained.

[0062] When the report type is a report type, the terminal generates a production data report according to the target report template, indicator attributes and corresponding target production indicator sub-data of the report type, and outputs the production data report as a response to the question to be answered.

[0063] When the report type is a chart type, the terminal generates a production data chart according to the target report template, indicator attributes and corresponding target production indicator sub-data of the chart type, and outputs the production data chart as a response to the question to be answered.

[0064] In some embodiments, Figure 3 As shown, the production data processing method also includes: obtaining preset report information, obtaining preset production indicator information corresponding to the preset report information, and for each report type in multiple report types, constructing a report template under the report type based on the preset report information and the preset production indicator information.

[0065] Among them, based on the preset report information and preset production indicator information, constructing report templates under different report types may include constructing report templates under the report type based on the preset report information and preset production indicator information, and may also include: constructing report templates under the chart type based on the preset report information and preset production indicator information.

[0066] The above production data processing method, through the powerful processing ability of the big model, extracts entities for the questions to be answered, and recalls them based on the extracted entity information, so as to obtain the target report information that matches the questions to be answered, thereby improving the quality of the target report information; since the target report information is associated with the target production indicator information, the target production indicator information required for generating the report can be obtained based on the target report information, and the target production indicator data corresponding to the target production indicator information can be obtained from the production indicator data, that is, the target report information is recalled through the big model, and the associated target production indicator information is subsequently determined, and then the target production indicator data is obtained based on the target production indicator information. In this way, the big model does not need to directly A large amount of production data is analyzed and processed, which reduces the amount of data and calculation involved in obtaining the target production index data, and greatly improves the efficiency of obtaining the target production index data; since the production index data is obtained by processing the production data of multiple data sources, the target production index data can come from different data sources, and there is no need to obtain the production data required for generating reports from different data sources respectively, which improves the convenience of obtaining data from different data sources and also improves the efficiency of finding the target production index data; in addition, the above process does not require human participation, and can automatically process the questions to be answered and output reports that are consistent with the questions to be answered, effectively improving the efficiency of generating complex reports.

[0067] In some embodiments, searching for target report information based on the recall results includes: when the first report information included in the recall results meets the report generation conditions, using the first report information as the target report information; when the first report information does not meet the report generation conditions, performing at least one recall based on the first report information and the question to be answered to obtain the target report information; when the recall result is empty, extracting the question feature vector of the question to be answered, and performing at least one recall based on the question feature vector to obtain the target report information.

[0068] Among them, the first report information meets the report generation conditions, which can indicate that the amount of first report information is small and the differences between the first report information are small; the first report information meets the report generation conditions, which can reflect that the first recall has obtained better results.

[0069] The first report information does not meet the report generation condition, which may mean that the amount of first report information is large, or the differences between the first report information are large.

[0070] The recall result is empty, which means that the first recall performed by the large model did not obtain any results.

[0071] Specifically, when the recall result includes the first report information, it is determined whether the first report information meets the report generation conditions. If the first report information meets the report generation conditions, the first report information is used as the target report information, that is, the target report information that matches the question to be answered is obtained through one recall.

[0072] If the first report information does not meet the report generation conditions, the first report information and the question to be answered will be recalled at least once through the big model to obtain the target report information; in this way, the big model can use more information for in-depth information retrieval based on the previous recall to obtain the target report information that matches the question to be answered.

[0073] When the recall result is empty, the question feature vector of the question to be answered is extracted through the large model, and feature-level recall is performed based on the question feature vector to obtain the target report information; when the recall result is empty, it means that the report information matching the question to be answered cannot be recalled based on the entity information of the question to be answered, and a deep feature-level search can be performed on the question to be answered to recall the target report information that matches the question to be answered at the semantic feature level.

[0074] In the above embodiment, depending on whether the first report information included in the recall result meets the report generation conditions, the first report information can be used as the target report information, or a deep recall can be performed based on the first report information and the question to be answered to obtain the target report information. When the recall result is empty, a deep recall at the feature level can be performed based on the question to be answered to obtain the target report information, so that the target report information is consistent with the question to be answered, thereby improving the correlation between the target report information and the question to be answered, and thereby improving the quality of subsequently generated reports.

[0075] In some embodiments, the report generation conditions include a preset quantity condition and a preset difference condition; when the first report information included in the recall result satisfies the report generation conditions, the first report information is used as the target report information, including: when the quantity of the first report information included in the recall result satisfies the preset quantity condition, based on the first similarity between the first report information and the question to be answered, determining a first similarity difference value; when the first similarity difference value satisfies the preset difference condition, the first report information is used as the target report information.

[0076] Specifically, the quantity of the first report information satisfies the preset quantity condition, which may be that the quantity of the first report information belongs to the preset quantity interval; the quantity belonging to the preset quantity interval is less than the quantity not belonging to the preset quantity interval, so the quantity of the first report information satisfies the preset quantity condition, which means that the quantity of the first report information is small. The specific value of the preset quantity interval can be set according to actual needs, and the embodiments of the present application do not limit this.

[0077] The first report feature vector of the first report information and the question feature vector of the question to be answered can be extracted through the large model, and the first similarity is calculated based on the first report feature vector and the question feature vector; when the number of first reports is at least two, the variance is calculated based on at least two first similarities to obtain a first similarity difference value, and when the first similarity difference value belongs to a preset difference interval, it is determined that the first similarity difference value meets the preset difference condition, and when the first similarity difference value does not belong to the preset difference interval, it is determined that the first similarity difference value does not meet the preset difference condition. The preset difference interval can be [0, 1.5]; the specific value of the preset difference interval can be set according to actual needs, and the embodiment of the present application does not limit this.

[0078] In some cases, when the number of first report information is one, it is determined whether the first similarity belongs to a preset similarity interval. If so, the first report information is used as the target report information. The preset similarity interval can be [0.75, 1]; the specific value of the preset similarity interval can also be set according to actual needs, and the embodiment of the present application does not limit this.

[0079] In practical applications, ElasticSearch can be used to extract the first report feature vector of the first report information, extract the question feature vector of the question to be answered, and calculate the first similarity; ElasticSearch is an open source distributed search engine and analysis engine, and Elasticsearch can perform retrieval based on the similarity between feature vectors.

[0080] It should be noted that the first similarity belongs to the preset similarity interval, indicating that the similarity between the first report information and the question to be answered is high; the first similarity variance belongs to the preset difference interval, indicating that the difference between multiple first report information is small, that is, the difference between multiple first report information obtained by the large model is small, and the recall result is relatively stable.

[0081] When the first similarity satisfies a preset difference condition, the first report information is used as target report information.

[0082] In some embodiments, when the first similarity does not meet the preset difference condition, at least one recall is performed based on the first report information and the question to be answered to obtain the target report information; that is, when the similarity between the first report information and the question to be answered is small, or the difference between multiple first report information is large, it means that the correlation between the first report information and the question to be answered is low, and deep recall can be performed on the question to be answered and the first report information to obtain the target report information that matches the question to be answered.

[0083] In the above embodiment, the first report information that meets the preset quantity condition and the preset difference condition is used as the target report information, so that the target report information is consistent with the question to be answered, thereby improving the quality of the target report information.

[0084] In some embodiments, the report generation conditions include a preset quantity condition and a preset difference condition; when the first report information does not meet the report generation conditions, at least one recall is performed based on the first report information and the questions to be answered to obtain target report information, including: when the quantity of the first report information does not meet the preset quantity condition, the first report information is batched to obtain multiple report information batches; a recall is performed based on the questions to be answered and multiple report information batches to obtain second report information; and the target report information is determined based on the second report information and the questions to be answered.

[0085] Specifically, the number of first report information does not meet the preset quantity condition, indicating that the number of first report information is large; in order to ensure the quality and efficiency of the output report, it is necessary to limit the number of target report information. When the number of first report information is large, the first report information needs to be further screened.

[0086] The batch number can be selected in the preset quantity range, and the first report information can be divided into batches according to the batch number, that is, the number of multiple report information batches obtained is the batch number; for example, the number of first report information is 6, and the batch number is 2, then the multiple report information batches are 2 report information batches, and each report information batch includes 3 first report information.

[0087] An agent can be used to submit report information batches to the big model for recall. For each report information batch, the first report information and questions to be answered of the report information batch are recalled through the big model to obtain the second report information corresponding to the report information batch. The big model determines the target report information based on the second report information and questions to be answered corresponding to multiple report information batches. Among them, the agent is an intelligent entity built based on the big model and can handle tasks such as natural language understanding, dialogue generation, and information retrieval.

[0088] In the above embodiment, when the first report information does not meet the preset quantity condition, the first report information is divided into batches, and the report information batches are recalled to quickly reduce the quantity and obtain the second report information that meets the preset quantity condition. The target report information is then determined based on the second report information, so that the target report information is consistent with the question to be answered, thereby improving the quality of the target report information.

[0089] In some embodiments, target report information is determined based on the second report information and the question to be answered, including: determining a second similarity difference value based on the second similarity between the second report information and the question to be answered; if the second similarity difference value meets a preset difference condition, using the second report information as the target report information; if the second similarity difference value does not meet the preset difference condition, recalling based on the question to be answered and the second report information to obtain the target report information.

[0090] Specifically, the number of second report information is the same as the number of report information batches, that is, the number of second report information is multiple; corresponding to each second report information, the second report feature vector of the second report information is extracted through the large model, and the second similarity is calculated according to the second report feature vector and the problem feature vector; according to the second similarities corresponding to each of the multiple second report information, the second similarity variance is calculated, and the second similarity variance is the second similarity difference value.

[0091] When the second similarity difference value belongs to the preset difference interval, it is determined that the second similarity difference value satisfies the preset difference condition, and then the second report information is used as the target report information.

[0092] When the second similarity difference value belongs to the preset difference interval, it is determined that the second similarity difference value satisfies the preset difference condition, and the second report information is used as the target report information; when the second similarity difference value does not belong to the preset difference interval, it is determined that the second similarity difference value does not satisfy the preset difference condition, and the second report information and the questions to be answered are recalled through the large model to obtain the target report information.

[0093] In practical applications, the second report feature vector of the second report information can be obtained through ElasticSearch, and the second similarity can be calculated.

[0094] In the above embodiment, when the second report information meets the preset difference condition, the second report information is used as the target report information. When the second report information does not meet the preset difference condition, questions to be answered are added for deep recall to obtain the target report information, so that the target report information is consistent with the questions to be answered, thereby improving the quality of the target report information.

[0095] In some embodiments, at least one recall is performed based on the problem feature vector to obtain target report information, including: obtaining a candidate report feature vector in a vector library; performing a recall based on the problem feature vector and the candidate report feature vector to obtain a target report feature vector; obtaining third report information corresponding to the target report feature vector; and determining the target report information based on the third report information.

[0096] Among them, the vector library stores the candidate report feature vectors of all candidate report information; it should be noted that the first report information and the second report information recalled in the foregoing text can be retrieved from the candidate report information.

[0097] The vector library also stores candidate report information corresponding to the candidate report feature vector.

[0098] Specifically, when the recall result is empty, obtain the candidate report feature vector in the vector library, input the candidate report feature vector and the question feature vector into the large model, obtain the target report feature vector through the large model recall, and obtain the third report information corresponding to the target report feature vector in the vector library; when the third report information meets the report generation conditions, use the third report information as the target report information; when the third report information does not meet the report generation conditions, recall is performed based on the third report information and the question to be answered to obtain the target report information.

[0099] In some embodiments, determining target report information based on third report information includes: selecting fourth report information that meets preset difference conditions in the third report information based on a third similarity between the third report information and the question to be answered; and recalling based on the fourth report information and the question to be answered to obtain target report information that meets preset quantity conditions.

[0100] Specifically, when the third report information does not meet the report generation conditions (preset quantity conditions and preset difference conditions), the third similarity between each third report information and the question to be answered is calculated, and the third report information is arranged in order of the third similarity from large to small to obtain a report information sequence, and the first N third report information arranged in the report information sequence are selected, and the third similarity variance is determined based on the third similarities corresponding to the N third report information. If the third similarity variance meets the preset difference condition, the N third report information is used as the fourth report information; wherein N is a positive integer, and the value of N can be set according to actual needs; if the third similarity variance does not meet the preset difference condition, N is reduced, for example, the value of N is updated by N=N-1, and the step of selecting the first N third report information arranged in the report information sequence is continued until the fourth report information that meets the preset difference condition is obtained.

[0101] When the number of the fourth report information does not meet the preset quantity conditions, a recall is performed based on the fourth report information and the questions to be answered. If the number of report information obtained by the recall meets the preset quantity conditions, the report information obtained by the recall is used as the target report information. If the number of report information obtained by the recall does not meet the preset quantity conditions, the target report information that meets the preset quantity is selected from the recalled report information.

[0102] When the quantity of the fourth report information meets a preset quantity condition, the fourth report information may be used as the target report information.

[0103] In the above embodiment, when the recall result is empty, feature-level recall can be performed in the candidate report feature vectors stored in the vector library based on the question feature vector of the question to be answered. Through semantic-level recall, the target report information related to the question to be answered can be effectively recalled, thereby improving the quality of the target report information.

[0104] In some embodiments, searching for target report information based on the recall results includes: when the recall results include candidate production indicator information, searching for recall production indicator information based on the candidate production indicator information, and using default report information corresponding to the recall production indicator information as the target report information; obtaining target production indicator information associated with the target report information, including: using the recall production indicator information as the target production indicator information.

[0105] Specifically, in practical applications, the questions to be answered may be used to query some production indicator information instead of generating a specific report. As a result, the large model fails to retrieve the report information related to the questions to be answered, but retrieves the production indicator information related to the questions to be answered.

[0106] When the recall result includes candidate production indicator information, the indicator similarity between the candidate production indicator information and the question to be answered can be calculated. If the variance between the indicator similarities meets the preset difference condition, the candidate production indicator information is used as the recall indicator information. If the variance between the indicator similarities does not meet the preset difference condition, the large model can recall based on the candidate production indicator information and the question to be answered to obtain the recalled production indicator information.

[0107] The default report information corresponding to the recalled production index information may be, based on the quantity of the recalled production index information, obtaining the default report information whose index quantity is the quantity of the recalled production index information, that is, the number of indicators included in the default report information, that is, the number of the recalled production index information.

[0108] The default report information is used as the target report information, and the recalled production indicator information is used as the target production indicator information. Then, the target production indicator data corresponding to the recalled indicator information can be obtained, and a production data report can be generated based on the target production indicator data corresponding to the default report information (target report information) and the recalled indicator information (target production indicator information). In this way, when the user intention corresponding to the question to be answered is a query indicator, a production data report can be generated based on the recalled indicator information and the default report information, so that the generated production data report is consistent with the question to be answered, thereby improving the quality of the production data report.

[0109] In some embodiments, the target production indicator information includes at least one target sub-information; determining the target production indicator data corresponding to the target production indicator information based on the production indicator data, including: for each target sub-information, obtaining the target production indicator data corresponding to the target sub-information in the production indicator data of the indicator service platform; if the target production indicator data corresponding to the target sub-information is not obtained, obtaining the target generation rule corresponding to the target sub-information in the indicator generation rule; obtaining the indicator sub-data corresponding to the target generation rule in the production indicator data, and generating the target production indicator data corresponding to the target sub-information based on the target generation rule and the indicator sub-data.

[0110] Among them, the indicator service platform includes production indicator data obtained by processing production data from multiple data sources; it can be understood that the indicator service platform is connected to multiple data sources, and the production data from multiple data sources are integrated and converted to the indicator service platform. The indicator service platform can be used to search for production data from multiple data sources without having to enter different data sources separately, and data from all aspects involved in production can be found.

[0111] The target generation rule includes a mapping relationship between initial production index information and preset production index information. The target generation rule can be used to process production index data corresponding to the initial production index information to generate production index data corresponding to the preset production index information.

[0112] Specifically, for each target sub-information, the target production indicator data corresponding to the target sub-information is obtained from all production indicator data on the indicator service platform. If the target production indicator data corresponding to the target sub-information is not obtained in the production indicator data, the target generation rule corresponding to the target sub-information is searched, and the target sub-information is the preset indicator information corresponding to the found target generation rule; the indicator sub-data corresponding to the initial production indicator information in the target generation rule is obtained from the production indicator data, and the indicator sub-data corresponding to the initial production indicator information is calculated according to the target generation rule to obtain the target production indicator data corresponding to the target sub-information.

[0113] In the above embodiment, the target production index data can be obtained in the index service platform, and the required target production index data can also be generated through the index generation rules; the production data of multiple data sources can be searched through the index service platform without entering different data sources separately, and data on all aspects involved in production can be found, thereby improving the efficiency of obtaining target production index data.

[0114] In some embodiments, based on the obtained questions to be answered, before calling the big model to obtain the target report information, it also includes: obtaining a preset first production indicator information; searching for production data that matches the first production indicator information; performing indicator processing on the production data that matches the first production indicator information to obtain production indicator data corresponding to the first production indicator information; associating the first production indicator information and the production indicator data and saving them to the indicator service platform; obtaining a preset second production indicator information and an indicator generation rule corresponding to the second production indicator information; associating the second production indicator information and the indicator generation rule and saving them to the indicator service platform.

[0115] The production index data corresponding to the first production index information can be obtained by converting the production data, and the production index data corresponding to the second production index information can be obtained by processing the production data according to the index generation rule.

[0116] Specifically, exemplarily, Figure 4As shown, after obtaining the preset first production indicator information, the production data matching the first production indicator information is searched in the data source, and the production data matching the first production indicator information is processed into indicators to obtain the production indicator data corresponding to the first production indicator information; wherein the indicator processing may include: data cleaning and data format conversion processing; the first production indicator information and the production indicator data are associated and saved to the indicator service platform, so that the production indicator data required for generating a production data report can be obtained based on the associated relationship later; after obtaining the preset second production indicator information and the indicator generation rule corresponding to the second production indicator information, the second production indicator information and the indicator generation rule are saved to the indicator service platform.

[0117] In some embodiments, the update status of production data of a data source can be detected. When the production data corresponding to the first production indicator information in the data source is updated, the indicator service platform can pull the updated production data and perform indicator processing on the updated production data to synchronously update the production indicator data corresponding to the first production indicator information.

[0118] In actual applications, the first production indicator information, the second production indicator information and the indicator generation rules can be implemented respectively through scripting languages, and the association relationship between the report information and the production indicator information can also be implemented through scripting languages. When it is necessary to modify the first production indicator information, the second production indicator information, the indicator generation rules, the report information and the association relationship between the production indicator information, the modification can be achieved by modifying the script language. Further, when the user's production situation is updated, the script language involved in the update can be modified, that is, the synchronous update of production indicators, reports and production conditions can be achieved. There is no need to rewrite the report generation logic according to the updated production situation. When the production situation is updated, the efficiency of synchronously updating production indicators and reports is improved.

[0119] In some embodiments, searching for production data that matches the first production index information includes: integrating data from multiple data sources to obtain integrated production data; and searching for production data that matches the first production index information in the integrated production data.

[0120] Specifically, data integration of multiple data sources may be performed by integrating production data of multiple data sources according to time periods to obtain integrated production data that is unified on a time scale. The integrated production data may be stored in a database, and production data matching the first production indicator information may be subsequently searched in the database.

[0121] By integrating data from multiple data sources, the integrated production data can reflect data from all aspects of production. Subsequently, the required production data can be found based on the integrated production data and processed into production indicator data. Through data integration, the data silos formed by multiple different data sources are broken, and production data can be queried and analyzed more conveniently, which improves the accuracy and efficiency of querying and analyzing production data.

[0122] It should be noted that the production indicator data corresponding to the first production indicator information may be the production data that is required for generating a production data report and is not frequently updated, and the indicator data corresponding to the second production indicator information may be the production data that is required for generating a production data report and is frequently updated, such as production indicator data that need to be statistically updated every day. If such production data is directly stored on the indicator service platform, it may cause greater data processing pressure on the indicator service platform; associating the first production indicator information and the corresponding production indicator data and saving them on the indicator service platform, and associating the second production indicator information and the corresponding indicator generation rules and saving them on the indicator service platform can reduce the load on the indicator service platform in data conversion and storage, and ensure the query efficiency of the indicator service platform.

[0123] In some embodiments, the production data processing method can be applied to scenarios where reports are queried through intelligent question-answering tasks, such as Figure 5 shown.

[0124] The user enters the question to be answered on the displayed question and answer page. The terminal obtains the input question to be answered on the question and answer page, calls the big model, segmentes the question to be answered based on the production terminology vocabulary, obtains the segmented words to be processed, extracts entities based on the segmented words to be processed, obtains entity information, and recalls based on the entity information through the big model to obtain the recall result.

[0125] When the first report information included in the recall result satisfies the report generation condition, the first report information is used as the target report information.

[0126] In the case where the recall result includes the first report information, the quantity of the first report information is obtained, and whether the preset quantity condition is met is determined based on the quantity of the first report information; if the preset quantity condition is not met, the first report information is batched to obtain multiple report information batches, and a recall is performed based on the questions to be answered and the multiple report information batches to obtain the second report information; if the preset quantity condition is met, the first report information is used as the second report information; based on the second report information, it is determined whether the preset difference condition is met; if it is met, the second report information is used as the target report information; if it is met, the second report information is used as the target report information; if it is not met, a recall is performed based on the second report information and the questions to be answered to obtain the target report information.

[0127] When the recall result is empty, extract the question feature vector of the question to be answered, obtain the candidate report feature vector in the vector library, perform recall based on the question feature vector and the report feature vector, and obtain the third report information through feature-level recall; based on the similarity between the third report information and the question to be answered, select the fourth report information that meets the preset difference condition from the third report information; perform selective recall through the fourth report information and the question to be answered, and obtain the target report information that meets the preset quantity condition.

[0128] Obtain target production indicator information associated with the target report information, determine the target production indicator data corresponding to the target production indicator information based on the production indicator data, generate a production data report based on the target report information and the target production indicator data, and output the production data report as a response to the question to be answered.

[0129] Through the above-mentioned production data processing method, through the powerful processing ability of the big model, the entity extraction of the question to be answered is carried out, and the recall is carried out based on the extracted entity information, so that the target report information consistent with the question to be answered can be obtained, and the quality of the target report information is improved; because the target report information is related to the target production indicator information, the target production indicator information required for generating the report can be obtained according to the target report information, and the target production indicator data corresponding to the target production indicator information can be obtained in the production indicator data, that is, the target report information is recalled through the big model, and the associated target production indicator information is subsequently determined, and then the target production indicator data is obtained based on the target production indicator information. In this way, the big model does not need to directly Analyzing and processing a large amount of production data reduces the amount of data and calculation involved in obtaining the target production index data, greatly improving the efficiency of obtaining the target production index data; since the production index data is obtained by processing the production data of multiple data sources, the target production index data can come from different data sources, and there is no need to obtain the production data required for generating reports from different data sources respectively, which improves the convenience of obtaining data from different data sources and also improves the efficiency of finding the target production index data; in addition, the above process does not require human participation, and can automatically process questions to be answered and output reports that are consistent with the questions to be answered, effectively improving the efficiency of generating complex reports.

[0130] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0131] Based on the same inventive concept, an embodiment of the present application also provides a production data processing device. The implementation solution for solving the problem provided by the device is similar to the implementation solution recorded in the above method. Therefore, the specific limitations in one or more production data processing device embodiments provided below can refer to the limitations on the production data processing method above and will not be repeated here.

[0132] In one embodiment, Figure 6 As shown, a production data processing device is provided, comprising:

[0133] The recall module 601 is used to call the big model to obtain target report information based on the obtained questions to be answered; the big model is used to extract entity information from the questions to be answered, perform recall based on the entity information, obtain recall results, and search for target report information based on the recall results;

[0134] An acquisition module 602 is used to acquire target production index information associated with target report information;

[0135] Determination module 603, used to determine target production index data corresponding to target production index information based on production index data; the production index data is obtained by processing production data from multiple data sources;

[0136] The output module 604 is used to generate a production data report according to the target report information and the target production index data, and output the report as a response to the question to be answered.

[0137] In some embodiments, the recall module 601 is also used to, if the first report information included in the recall result satisfies the report generation condition, use the first report information as the target report information; if the first report information does not satisfy the report generation condition, perform at least one recall based on the first report information and the question to be answered to obtain the target report information; if the recall result is empty, extract the question feature vector of the question to be answered, perform at least one recall based on the question feature vector to obtain the target report information.

[0138] In some embodiments, the report generation conditions include a preset quantity condition; the recall module 601 is also used to batch the first report information to obtain multiple batches of report information when the quantity of the first report information does not meet the preset quantity condition; recall based on the questions to be answered and the multiple batches of report information to obtain second report information; determine the target report information based on the second report information and the questions to be answered.

[0139] In some embodiments, the recall module 601 is also used to obtain candidate report feature vectors in the vector library; perform recall based on the problem feature vector and the candidate report feature vector to obtain a target report feature vector; obtain third report information corresponding to the target report feature vector; and determine the target report information based on the third report information.

[0140] In some embodiments, the target production indicator information includes at least one target sub-information, and the acquisition module 602 is also used to obtain, for each target sub-information, the target production indicator data corresponding to the target sub-information in the production indicator data of the indicator service platform; if the target production indicator data corresponding to the target sub-information is not obtained, obtain the target generation rule corresponding to the target sub-information in the indicator generation rule; obtain the indicator sub-data corresponding to the target generation rule in the production indicator data, and generate the target production indicator data corresponding to the target sub-information based on the target generation rule and the indicator sub-data.

[0141] In some embodiments, the production data processing device also includes: an indicator service platform processing module, which is used to obtain a preset first production indicator information; search for production data that matches the first production indicator information; perform indicator processing on the production data that matches the first production indicator information to obtain production indicator data corresponding to the first production indicator information; associate the first production indicator information and the production indicator data and save them to the indicator service platform; obtain the preset second production indicator information and the indicator generation rule corresponding to the second production indicator information; associate the second production indicator information and the indicator generation rule and save them to the indicator service platform.

[0142] Each module in the above production data processing device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0143] Figure 7 The schematic diagram of the structure of (device subject) provided in this application. Figure 7As shown, the computer device 70 provided in this embodiment includes: at least one processor 701 and a memory 702. Optionally, the device 70 also includes a communication component 703. The processor 701, the memory 702 and the communication component 703 are connected via a bus 704.

[0144] In a specific implementation process, at least one processor 701 executes the computer-executable instructions stored in the memory 702, so that at least one processor 701 executes the above method.

[0145] The specific implementation process of the processor 701 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.

[0146] In the above embodiments, it should be understood that the processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the invention can be directly implemented as a hardware processor, or can be implemented by a combination of hardware and software modules in the processor.

[0147] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (NVM), such as at least one disk storage.

[0148] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of this application is not limited to only one bus or one type of bus.

[0149] The embodiment of the present application further provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, each step of the method in the above embodiment is implemented.

[0150] The embodiment of the present application also provides a computer program product, including computer executable instructions, which implement the various steps of the method in the above embodiment when executed by a processor.

[0151] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0152] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0153] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0154] The division of units is only a logical function division, and there may be other divisions 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 mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0155] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0156] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0157] 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, 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. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0158] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0159] Finally, it should be noted that those skilled in the art will readily conceive of other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention, which follow the general principles of the present invention and include common knowledge or customary technical means in the art not disclosed by the present invention, are not limited to the precise structure described above and shown in the drawings, and may be modified and changed in various ways without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A production data processing method, characterized in that: include: Based on the questions to be answered, the big model is called to obtain the target report information; The large model is used to extract entity information from the question to be answered, perform recall based on the entity information, obtain a recall result, and search for the target report information based on the recall result; Acquiring target production indicator information associated with the target report information; Determining target production index data corresponding to the target production index information based on the production index data; the production index data is obtained by processing production data from multiple data sources; A production data report is generated according to the target report information and the target production index data, and the production data report is output as a response to the question to be answered.

2. The method according to claim 1, characterized in that The searching for the target report information based on the recall result includes: In a case where the first report information included in the recall result meets the report generation condition, using the first report information as the target report information; In the case that the first report information does not satisfy the report generation condition, performing at least one recall based on the first report information and the question to be answered to obtain the target report information; When the recall result is empty, the question feature vector of the question to be answered is extracted, and at least one recall is performed based on the question feature vector to obtain the target report information.

3. The method according to claim 2, characterized in that The report generation conditions include a preset quantity condition; In the case where the first report information does not meet the report generation condition, performing at least one recall based on the first report information and the question to be answered to obtain the target report information includes: When the quantity of the first report information does not satisfy the preset quantity condition, dividing the first report information into batches to obtain a plurality of report information batches; Performing a recall based on the questions to be answered and the multiple report information batches to obtain second report information; The target report information is determined based on the second report information and the question to be answered.

4. The method according to claim 2, characterized in that: The performing at least one recall based on the problem feature vector to obtain the target report information includes: Obtain candidate report feature vectors in the vector library; Perform recall based on the question feature vector and the candidate report feature vector to obtain a target report feature vector; Obtaining third report information corresponding to the target report feature vector; The target report information is determined based on the third report information.

5. The method according to any one of claims 1 to 4, characterized in that The target production index information includes at least one target sub-information; the determining the target production index data corresponding to the target production index information based on the production index data includes: For each target sub-information, obtain the target production indicator data corresponding to the target sub-information from the production indicator data of the indicator service platform; In the case where the target production indicator data corresponding to the target sub-information is not obtained, obtaining the target generation rule corresponding to the target sub-information in the indicator generation rule; The indicator sub-data corresponding to the target generation rule is obtained from the production indicator data, and the target production indicator data corresponding to the target sub-information is generated according to the target generation rule and the indicator sub-data.

6. The method according to claim 5, characterized in that Before calling the large model to obtain target report information based on the obtained questions to be answered, the method further includes: Obtaining preset first production indicator information; Searching for production data matching the first production indicator information; Performing index processing on the production data matching the first production index information to obtain production index data corresponding to the first production index information; Associating the first production index information with the production index data and saving them to the index service platform; Obtaining preset second production index information and an index generation rule corresponding to the second production index information; The second production index information and the index generation rule are associated and saved in the index service platform.

7. A production data processing device, characterized in that: The device comprises: A recall module is used to call the big model to obtain target report information based on the obtained questions to be answered; the big model is used to extract entity information from the questions to be answered, perform recall based on the entity information, obtain recall results, and search for the target report information based on the recall results; An acquisition module, used for acquiring target production index information associated with the target report information; A determination module, configured to determine target production index data corresponding to the target production index information based on production index data; the production index data is obtained by processing production data from multiple data sources; An output module is used to generate a production data report based on the target report information and the target production indicator data, and output the report as a response to the question to be answered.

8. A computer device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

10. A computer program product, characterized in that The method comprises computer-executable instructions, which implement the method according to any one of claims 1 to 6 when executed by a processor.

Images