Summary report generation method and device, equipment, medium and product

Abstract

The embodiment of the invention provides a summary report generation method and device, equipment, a medium and a product, and relates to the field of report compiling. The method comprises the steps of analyzing first report data and second report data to obtain a first report and a second report respectively, screening items with the same item name through a large language model based on a first prompt text to obtain a first item group with new content progress, screening items, with the item names different from those in the second report, in the first report as a second item group, based on the second prompt text and the third prompt text, the first entry group and the second entry group are processed through a large language model, first summary content and second summary content are obtained, and therefore a summary report is generated. Through the mode, the summary report for comparing the contents in the two report data is generated, the progress condition before and after the project is reflected, the report is more targeted, the practicability of the report is higher, the time of the user is saved, the user can more intuitively know the contents in the two reports, and the use experience of the user is improved.

Description

Technical Field

[0001] This application relates to the field of report compilation, and particularly to a summary report generation method, device, equipment, medium and product. Background Art

[0002] With the increasingly widespread use of large language models, report writing can also be generated through large language models. In related technologies, for report types of reports, it is usually limited to summarizing the content of a single report. For some reports related to the progress of projects in actual applications, it is often necessary to compare the changes before and after the project. By comparing the content in two reports, the progress of the project is summarized. However, in related technologies, there is no practical summary report generation method that can compare the data in two reports and generate a summary report.

[0003] Therefore, how to generate a comparative summary report based on two reports through a large language model is an urgent problem to be solved. Summary of the Invention

[0004] Based on the above technical problems, the embodiments of this application provide a summary report generation method, device, equipment, medium and product, aiming to generate a comparative summary report based on two reports through a large language model.

[0005] In the first aspect of the embodiments of this application, a summary report generation method is provided. The method includes:

[0006] Analyze the first report data and the second report data to obtain the first report and the second report respectively. The generation time of the first report data is later than the generation time of the second report data;

[0007] Based on the first prompt text, use the large language model to screen the entries with the same project name to obtain the first entry group with new progress in the entry content. The first prompt text at least includes: the entry information in the first report, the entry information in the second report, and there is new progress in the entries with the same project name in the first report compared with those in the second report;

[0008] Screen each entry in the first report whose project name is different from that in the second report, and determine it as the second entry group;

[0009] Based on the second prompt text, use the large language model to process the first entry group to obtain the first summary content. The second prompt text at least includes: the entry information of the first entry group; the project progress;

[0010] Process the second entry group through the large language model based on the third prompt text to obtain a second summary content, where the third prompt text at least includes: entry information of the second entry group; project progress;

[0011] Generate a summary report according to the first summary content and the second summary content.

[0012] Optionally, the method further includes:

[0013] Generate a report template through the large language model based on the fourth prompt text, where the fourth prompt text at least includes: project category; summary content template for the project category;

[0014] Add the first summary content and the second summary content to the report template according to the project category to generate the summary report.

[0015] Optionally, the method further includes:

[0016] When the project categories of each entry in the first entry group and the second entry group are known, cluster the entries in the first entry group and the second entry group according to the project category of each entry to obtain a clustering result;

[0017] When the project categories of each entry in the first entry group and the second entry group are unknown, obtain a preset category as the project category;

[0018] Classify the entries with unknown project categories in the first entry group and the second entry group through the large language model based on the fifth prompt text, where the fifth prompt text at least includes: project category and entries with unknown project categories;

[0019] Cluster the entries in the first entry group and the second entry group according to the project category of each entry to obtain a clustering result;

[0020] Generate a summary report according to the clustering result and the summary content template corresponding to each project category.

[0021] Optionally, screen the entries with the same project name through the large language model based on the first prompt text to obtain a first entry group with new progress in the entry content. The method includes:

[0022] Obtain each entry in the first report and the second report;

[0023] Compare the project names of each entry in the first report with the project names of each entry in the second report;

[0024] Screen each entry with the same item name in the first report and the second report;

[0025] Based on the first prompt text, use the large language model to screen the entries with the same item name, and obtain the first group of entries with new progress in the entry content.

[0026] Optionally, compare the item names of each entry in the first report with the item names of each entry in the second report; screen each entry with the same item name in the first report and the second report, and the method includes:

[0027] Establish a database for the parsed first report and the second report;

[0028] Retrieve the item name in the database, and obtain each entry in the first report and the second report that contains the same item name.

[0029] Optionally, parse the first report data and the second report data, and the method includes:

[0030] Parse the first report data and the second report data into a target format, where the target format is list text, and the list text includes at least one entry, and each entry includes at least an item name.

[0031] The second aspect of the embodiments of the present application provides a summary report generation device, and the device includes:

[0032] A data parsing module, configured to parse the first report data and the second report data respectively to obtain the first report and the second report, and the generation time of the first report data is later than the generation time of the second report data;

[0033] A first group of entry determination module, configured to screen the entries with the same item name based on the first prompt text through the large language model to obtain the first group of entries with new progress in the entry content; the first prompt text at least includes: the entry information in the first report, the entry information in the second report, and there is new progress compared with the entries with the same item name in the second report in the first report;

[0034] A second group of entry determination module, configured to screen each entry in the first report with a different item name from that in the second report, and determine it as the second group of entries;

[0035] A first summary content determination module, configured to process the first group of entries through the large language model based on the second prompt text to obtain the first summary content, and the second prompt text at least includes: the entry information of the first group of entries; the project progress;

[0036] The second summary content determination module is used to process the second entry group through the large language model based on the third prompt text to obtain the second summary content. The third prompt text at least includes: the entry information of the second entry group; the project progress;

[0037] The summary report generation module is used to generate a summary report according to the first summary content and the second summary content.

[0038] Optionally, the summary report generation device further includes:

[0039] The report template generation module is used to generate a report template through the large language model based on the fourth prompt text. The fourth prompt text at least includes: the project category; the summary content template of the project category;

[0040] The summary report generation module is used to add the first summary content and the second summary content to the report template according to the project category to generate the summary report.

[0041] Optionally, the summary report generation device further includes:

[0042] The first clustering module is used to cluster the entries in the first entry group and the second entry group according to the project category of each entry when the project category of each entry in the first entry group and the second entry group is known, to obtain a clustering result;

[0043] The preset category acquisition module is used to acquire a preset category as the project category when the project category of each entry in the first entry group and the second entry group is unknown;

[0044] The entry classification module is used to classify the entries with unknown project categories in the first entry group and the second entry group through the large language model based on the fifth prompt text. The fifth prompt text at least includes: the project category and the entries with unknown project categories;

[0045] The second clustering module is used to cluster the entries in the first entry group and the second entry group according to the project category of each entry, to obtain a clustering result;

[0046] The summary report generation module generates a summary report according to the clustering result and the summary content template corresponding to each project category.

[0047] Optionally, the first entry group determination module includes:

[0048] The entry acquisition sub-module is used to acquire each entry in the first report and the second report;

[0049] A project name comparison sub-module for comparing the project names of each entry in the first report with the project names of each entry in the second report;

[0050] An entry screening sub-module for screening each entry with the same project name in the first report and the second report;

[0051] A first entry group acquisition sub-module for screening the entries with the same project name through the large language model based on the first prompt text to obtain a first entry group with new progress in the entry content.

[0052] Optionally, the entry screening sub-module includes:

[0053] A database establishment unit for establishing a database of the parsed first report and the second report;

[0054] A project name retrieval unit for retrieving the project name in the database to obtain each entry in the first report and the second report containing the same project name.

[0055] Optionally, the data parsing module further includes:

[0056] A list text acquisition sub-module for parsing the first report data and the second report data into a target format, the target format being list text, and the list text including at least one entry, and each entry including at least a project name.

[0057] In the third aspect of the embodiments of the present application, an electronic device is provided. The electronic device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, the summary report generation method in the first aspect of the embodiments of the present application is implemented.

[0058] In the fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the summary report generation method in the first aspect of the embodiments of the present application is implemented.

[0059] In the fifth aspect of the embodiments of the present application, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the summary report generation method in the first aspect of the embodiments of the present application is implemented.

[0060] Through the summary report generation method of the embodiments of the present application, by parsing the report data generated at two different times, the first report and the second report are obtained. Then, through the large language model, the items in the two reports are screened according to the prompt text, obtaining the first item group composed of the same items and the second item group composed of the items that only appear in the report with a later time. Then, the large language model summarizes the progress of the first item group and the second item group, and finally generates a summary report.

[0061] In the present application, on the basis of using the large language model to generate a summary report, taking two pieces of report data for the same part of content but with different generation times as input, by setting the prompt text for the large language model, the comparison of the content in the two pieces of report data is realized. It can not only reflect the progress of the projects involved in the report data between the two reports, but also the screened items are more targeted, and the key projects with progress and new projects can be reflected in the report, making the report more practical, saving the user's time for comparing the two reports, facilitating the user to more intuitively understand the content of the two reports, and improving the user experience. Brief Description of the Drawings

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required to be used in the description of the embodiments of the present application. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0063] Figure 1 It is a flowchart of a summary report generation method proposed in an embodiment of the present application;

[0064] Figure 2 It is a structural block diagram of a summary report generation device provided in an embodiment of the present application;

[0065] Figure 3 It is a schematic diagram of an electronic device shown in an embodiment of the present application. Detailed Description of the Embodiments

[0066] The following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0067] In the accompanying drawings, for the sake of clarity, the sizes of the constituent elements, the thicknesses of the layers, or the areas may sometimes be exaggerated. Therefore, any implementation of the present application is not necessarily limited to the sizes shown in the figures, and the shapes and sizes of the components in the drawings do not reflect the true proportions. In addition, the drawings schematically show ideal examples, and any implementation of the present application is not limited to the shapes or values shown in the drawings, etc.

[0068] Please refer to Figure 1 , Figure 1 which is a flowchart of a method for generating a summary report of the present application. As Figure 1 shown, the method may include steps S101 to S106:

[0069] Step S101: Analyze the first report data and the second report data to obtain the first report and the second report respectively, where the generation time of the first report data is later than the generation time of the second report data;

[0070] Step S102: Screen the entries with the same project name through a large language model based on the first prompt text to obtain the first entry group with new progress in the entry content; the first prompt text at least includes: the entry information in the first report, the entry information in the second report, and there is new progress compared with the entries with the same project name in the second report in the first report;

[0071] Step S103: Screen each entry in the first report whose project name is different from that in the second report and determine it as the second entry group;

[0072] Step S104: Process the first entry group through the large language model based on the second prompt text to obtain the first summary content, where the second prompt text at least includes: the entry information of the first entry group; the project progress;

[0073] Step S105: Process the second entry group through the large language model based on the third prompt text to obtain the second summary content, where the third prompt text at least includes: the entry information of the second entry group; the project progress;

[0074] Step S106: Generate a summary report according to the first summary content and the second summary content.

[0075] In the process of using a large language model for report generation, usually only the content set of one report can be parsed and processed to summarize the content in one report. However, in the process of production and life, a summary report often needs to compare historical data to find the development trend from the comparison, which is more beneficial to summarize experience and lessons and obtain more obvious and targeted data.

[0076] Therefore, a method is proposed in this application that can efficiently and conveniently generate a comparative summary report based on the data of two reports. Using the report data with relevant content and different time sequences as input, the large language model filters the entries with the same project name based on the first prompt text. The filtered entries are those that exist in both reports, which are determined as the first entry group. The entries in this group appear in pairs, and in the two reports, the content of the entries has progressed. There is also a part of the entries that exist in the later report data but not in the earlier report data, which form the second entry group. As new entries, their progress also needs to be counted. Then, based on the second prompt text and the third prompt text, the large language model summarizes the content in the first entry group and the second entry group to generate a summary report with a comparison relationship between before and after.

[0077] Step S101: Parse the first report data and the second report data to obtain the first report and the second report respectively. The generation time of the first report data is later than the generation time of the second report data.

[0078] In the embodiments of this application, first, two report data at different time points are used as input. After obtaining the first report data and the second report data, the original report data needs to be parsed and converted into a format that can be processed subsequently to obtain the corresponding first report and second report. And the generation time of the first report data is later than that of the second report data, so the content in the obtained first report is newer than the content in the second report.

[0079] Step S102: Filter the entries with the same project name based on the first prompt text through the large language model to obtain the first entry group with new progress in the entry content; the first prompt text at least includes: the entry information in the first report, the entry information in the second report, and the entries in the first report have new progress compared with the entries with the same project name in the second report.

[0080] In the embodiments of the present application, after obtaining the two reports that need to be summarized, they need to be processed by a large language model. Since the purpose of the present application is to generate a summary report comparing the contents of the two reports, first, the contents of the same project at different times in the two reports need to be screened out for easy comparison. Specifically, the entries can be screened by the large language model based on the first prompt text, where the first prompt text at least includes the entry information in the first report and the entry information in the second report, and there is new progress in the entry in the first report compared to the entry in the second report. The entry information contains the project name, and the project name is a fixed name for each entry and does not change in the entry at different times. For example, "Gymnasium construction project", "Subway Line 8 construction project", etc. Therefore, in the first report and the second report, according to the project name, the entries to which the same project belongs in the data of the two reports at different times can be determined. By screening out such two entries, the progress of the project before and after can be compared. Here, we screen out the entries with the same project name and new progress in the first report compared to the second report to form the first entry group.

[0081] For example, when screening the first entry group through the large language model, the following prompt words can be used:

[0082] You are a professional report writer. Given the descriptions of an entry in the new and old reports respectively, determine whether there is new progress in the content of this entry in the new report compared to the old report, and summarize the new completion situation. If there is, return "There is new progress", if not, return "None".

[0083] The new report is as follows:

[0084] ${Element information of a certain entry in the new report}

[0085] The old report is as follows:

[0086] ${Element information of the corresponding entry in the old report}

[0087] Step S103: Screen out each entry in the first report whose project name is different from that in the second report, and determine it as the second entry group.

[0088] In the embodiment of the present application, after comparing and obtaining entries whose project names are the same in the first report and the second report, it is also necessary to compare and obtain entries with different project names, especially entries that exist in the first report but not in the second report, indicating that the project is a newly added project after the second report is generated and before the first report is generated. For comparing the progress in the two reports, it is important to summarize the progress between the same projects, but the addition of new projects should not be ignored. Therefore, after screening and obtaining that the project names in the first report are different from the entries in the second report, these entries are confirmed as the second entry group.

[0089] Step S104: Based on the second prompt text, the first item group is processed by the large language model to obtain a first summary content, wherein the second prompt text at least includes: item information of the first item group; and project progress.

[0090] In an embodiment of the present application, after determining which items have progress, it is also necessary to convert the progress of the project into a report format. For this purpose, the present application processes each item in the first item group based on the second prompt text through a large language model, wherein the main content of the second prompt text should be a summary of the progress of the project based on the element information of the two items before and after the same project, thereby obtaining the first summary content of the first item group processed by the large language model as part of the input for generating a summary report.

[0091] For example, when the first summary content is obtained through processing with a large language model, the following prompt words can be used:

[0092] You are a professional report writer. You are given a description of an item in the new and old statements, summarizing the new developments of the item.

[0093] The new report is as follows:

[0094] ${New report item element information}

[0095] The old report is as follows:

[0096] ${Corresponding item element information of old report}

[0097] Step S105: Based on the third prompt text, the second item group is processed by the large language model to obtain a second summary content, wherein the third prompt text at least includes: item information of the second item group; and project progress.

[0098] In the embodiments of the present application, while summarizing the projects with progress before and after, it is also necessary to summarize the progress of the newly added projects during the process from the time of addition to the moment of generating the first report data. Therefore, the large language model is also used to process the entries in the second entry group based on the third prompt text, where the main content of the third prompt text should be to summarize the main progress of the entries in the second entry group, that is, to summarize the progress of the newly launched projects.

[0099] For example, when obtaining the second summary content through the processing of the large language model, the following prompt words can be used:

[0100] You are a professional report writer. Given the description of an entry in a report, summarize the main progress of this entry.

[0101] The description is as follows:

[0102] ${Element information corresponding to this entry}

[0103] Step S106: Generate a summary report according to the first summary content and the second summary content.

[0104] In the embodiments of the present application, the first summary content is generated for the projects with progress and the second summary content is generated for the newly launched projects based on the large language model. Among them, the first summary content and the second summary content cover the progress of all projects in the first report data and the second report data. The summary report generated in this way can effectively reflect the situation changes of each project before and after the two reports.

[0105] Combined with the above embodiments, in one implementation manner, the present application also provides a method for generating a summary report, which specifically includes the following content:

[0106] First, generate a report template through the large language model based on the fourth prompt text, where the fourth prompt text at least includes: project category; project category summary content template.

[0107] In the embodiments of the present application, since a single report may contain various types of items, in order to make the generated summary report well - organized, the large - language model can also be used to classify each item. Specifically, based on the fourth prompt text, a report template is generated through the large - language model. The fourth prompt text includes at least the item category and the summary content template for the item category. The item category can be classified according to the classification in the entry content of the report data. If there is no classification in the entry content, it can also be determined according to the user's usage requirements. The summary content template for the item category is a template divided according to the item category. For example, if the summary report contains entries of the engineering project category and entries of the policy implementation category, then it can be divided into a summary content template for the engineering project category and a summary content template for the policy implementation category. The summary content templates for each category are input into the large - language model as the report template used to generate the summary report.

[0108] For example, when generating a report template through the large - language model, the following prompt words can be used:

[0109] I. Item category

[0110] ${Summary content of item category}

[0111] II. Policy category

[0112] ${Summary content of policy category}

[0113] …

[0114] Then, the first summary content and the second summary content are added to the report template according to the item category to generate the summary report.

[0115] In the embodiments of the present application, after obtaining the first summary content regarding the first entry group and the second summary content regarding the second entry group through the large - language model in the above process, the summary content can be filled into the summary report template generated in the large - language model according to the item category to which each entry belongs. The multiple summary contents under the same category are spliced together to obtain all the summary contents under that category. The summary contents of each category are spliced together to finally generate a summary report comparing the two time - based reports.

[0116] Combining the above embodiments, in one implementation, the present application also provides a method for generating a summary report, which specifically includes the following content:

[0117] First, when the item categories of each entry in the first entry group and the second entry group are known, the entries in the first entry group and the second entry group are clustered according to the item categories of each entry to obtain a clustering result.

[0118] In the embodiments of the present application, when grouping each entry in the entry group as described above, there are two cases. One is that the item category information of the entry is included in each entry, that is, the item categories of each entry are known; the other is that the item category information of the entry is not included in each entry. When the item categories of each entry are known, as described above, each entry in the first entry group and the second entry group can be clustered, and each entry can be divided into different groups according to the item category.

[0119] Then, when the item categories of each entry in the first entry group and the second entry group are unknown, a preset category is obtained as the item category.

[0120] In the embodiments of the present application, when the item category information is not included in each entry, the category of each entry cannot be directly obtained from the report data. Therefore, these entries are classified according to a preset classification. Based on the relevant information involved in each entry, they are divided into several preset categories, and the preset categories are used as the classification targets for each entry in the first entry group and the second entry group.

[0121] Next, for the entries with unknown item categories in the first entry group and the second entry group, they are classified by the large language model based on the fifth prompt text, and the fifth prompt text at least includes: item category and entries with unknown item categories.

[0122] In the embodiments of the present application, when classifying entries that do not contain item category information, multiple preset classifications are first obtained, and classification also needs to be performed through the large language model according to the summary content of each entry. Specifically, the obtained preset classifications are input into the large language model together with the summary content of each entry. Based on the fifth prompt text, the large language model analyzes the correlation between the content of each entry and the preset classifications, and assigns each entry to the preset classification with the highest expected correlation, thereby achieving the purpose of classifying entries with unknown item categories.

[0123] For example, when classifying entries with unknown item categories through the large language model, the following prompt words can be used:

[0124] You are a professional report writer. Classify the following content, and the categories included are: ${preset classification}.

[0125] Content to be classified:

[0126] ${summary content}

[0127] Finally, according to the item categories of each entry, the entries in the first entry group and the second entry group are clustered to obtain a clustering result;

[0128] Generate a summary report based on the clustering results and the summary content templates corresponding to each project category.

[0129] In the embodiments of the present application, it is also necessary to cluster each entry with a completed category. According to the project types of each entry, multiple entries of the same category are clustered into one category. According to different categories, combined with the summary report template for classifying categories described above, the entries of the same category are written into the summary report template corresponding to the category, so as to obtain a final summary report reflecting the project progress classified by category.

[0130] Combined with the above embodiments, in one implementation, the present application also provides a method for generating a summary report. Based on the first prompt text, the large language model is used to screen the entries with the same project name, and the first entry group with new progress in the entry content is obtained, which specifically includes the following content:

[0131] First, obtain each entry in the first report and the second report.

[0132] In the embodiments of the present application, in the process of obtaining the first entry group described above, the following content is also included. First, obtain each entry in the first report and the second report. Since it is necessary to compare each entry in the first report and the second report to check whether there is an entry in the second report with the same project name as in the first report, it is necessary to obtain all the projects in the first report and the second report for comparison.

[0133] Then, compare the project names of each entry in the first report with the project names of each entry in the second report.

[0134] In the embodiments of the present application, after obtaining all the entries in the two reports, it is necessary to compare the project names of each entry in the first report. The comparison is for two purposes. One is to compare whether there is an entry in the second report with the same project name as in the first report. If so, it means that it is the report data of the same project at different times, and further comparison is needed to check whether there is progress in the project corresponding to the entry. The other is to screen out the entries that exist in the first report but do not have the same project name in the second report. These entries are new entries and their project progress after being added needs to be reflected in the summary report.

[0135] Next, screen out each entry with the same project name in the first report and the second report.

[0136] In the embodiments of the present application, for the entries with the same item names in the first report and the second report, they need to be regarded as a pair of entries, and further confirm whether there is progress. For the entries with different item names in the first report and the second report, there are two cases. One is as described above, which exists in the first report but not in the second report, so it is the entry corresponding to the newly added item. If it exists in the second report but not in the first report, it is considered as an item that has been completed and will not be considered here.

[0137] Finally, based on the first prompt text, the large language model is used to screen the entries with the same item names, and the first entry group with new progress in the entry content is obtained.

[0138] In the embodiments of the present application, after obtaining the entry pairs in the first entry group and the second entry group with the same item names in the above screening, the large language model can be used to screen them. Specifically, based on the first prompt text, the large language model can be informed of the descriptions of a certain entry in the new and old reports respectively, and then judge whether there is new progress for this entry. According to the judgment result, it is determined whether this item needs to be written into the summary report.

[0139] Combined with the above embodiments, in one implementation manner, the item names of each entry in the first report are compared with the item names of each entry in the second report; screening each entry with the same item name in the first report and the second report specifically includes the following content:

[0140] First, establish a database for the parsed first report and the second report.

[0141] In the embodiments of the present application, for the screening of whether there are entries with the same item names as described above, first, a database needs to be established for the first report parsed from the first report data and the second report parsed from the second report data. This database can obtain all the element information in the entry corresponding to the item name by searching for the item name. After establishing the database, it is possible to view and retrieve each entry in the first report and the second report in an organized manner.

[0142] Then, search for the item name in the database to obtain each entry in the first report and the second report that contains the same item name.

[0143] In the embodiment of the present application, after establishing the databases of the first report and the second report, the item names in the first report and the second report can be retrieved, and each item with the same item name in the first report and the second report can be screened out to form a pair of data pairs, which will be used to input into the large language model for project progress judgment later. Since the item name is unique and different projects have different item names, and for the same project, the item name remains the same at different times, using the item name as the retrieval target can clearly find the same items in the two reports.

[0144] Combined with the above embodiments, in one implementation, parsing the first report data and the second report data specifically includes the following content:

[0145] Parse the first report data and the second report data into a target format, where the target format is a list text, and the list text includes at least one item, and each item includes at least the item name.

[0146] In the embodiment of the present application, the obtained report data can be in various formats. However, in order to facilitate establishing the database and inputting into the large language model for the parsed first report and the second report, in one embodiment, the first report data and the second report data can be parsed into a target format, that is, a list text, which includes at least one item, and each item can contain multiple item contents, including at least the unique item name. In this way, it is convenient for users to classify and screen the report data and other operations, improving the efficiency of generating the summary report. Specifically, the target format can be the json (JavaScript Object Notation, JS key-value pair data) format, where the json structure is a list, and each element in the list is an item, and the key and value represent the specific content in the item.

[0147] Based on the same design concept, an embodiment of the present application provides a summary report generation device. Refer to Figure 2 , Figure 2 is the structural block diagram of the summary report generation device provided by an embodiment of the present application. As Figure 2 shown, the device includes:

[0148] A data parsing module, configured to parse the first report data and the second report data to obtain the first report and the second report respectively, where the generation time of the first report data is later than the generation time of the second report data;

[0149] The first entry group determination module is used to screen the entries with the same project name based on the first prompt text through a large language model, and obtain the first entry group with new progress in the entry content; the first prompt text at least includes: the entry information in the first report, the entry information in the second report, and the new progress compared with the entries with the same project name in the first report and the second report.

[0150] The second entry group determination module is used to screen each entry in the first report whose project name is different from that in the second report, and determine it as the second entry group.

[0151] The first summary content determination module is used to process the first entry group through the large language model based on the second prompt text, and obtain the first summary content. The second prompt text at least includes: the entry information of the first entry group; the project progress.

[0152] The second summary content determination module is used to process the second entry group through the large language model based on the third prompt text, and obtain the second summary content. The third prompt text at least includes: the entry information of the second entry group; the project progress.

[0153] The summary report generation module is used to generate a summary report according to the first summary content and the second summary content.

[0154] Optionally, the summary report generation device further includes:

[0155] The report template generation module is used to generate a report template through the large language model based on the fourth prompt text. The fourth prompt text at least includes: the project category; the summary content template of the project category.

[0156] The summary report generation module is used to add the first summary content and the second summary content to the report template according to the project category, and generate the summary report.

[0157] Optionally, the summary report generation device further includes:

[0158] The first clustering module is used to cluster the entries in the first entry group and the second entry group according to the project category of each entry when the project category of each entry in the first entry group and the second entry group is known, and obtain a clustering result.

[0159] The preset category acquisition module is used to obtain a preset category as the project category when the project category of each entry in the first entry group and the second entry group is unknown.

[0160] An item classification module, which is used to classify the items with unknown item categories in the first item group and the second item group based on the fifth prompt text through the large language model. The fifth prompt text at least includes: item categories and items with unknown item categories;

[0161] A second clustering module, which is used to cluster the items in the first item group and the second item group according to the item categories of each item to obtain a clustering result;

[0162] A summary report generation module, which generates a summary report according to the clustering result and the summary content template corresponding to each item category.

[0163] Optionally, the first item group determination module includes:

[0164] An item acquisition sub-module, which is used to acquire each item in the first report and the second report;

[0165] An item name comparison sub-module, which is used to compare the item names of each item in the first report with the item names of each item in the second report;

[0166] An item screening sub-module, which is used to screen each item with the same item name in the first report and the second report;

[0167] A first item group acquisition sub-module, which is used to screen the items with the same item name based on the first prompt text through the large language model to obtain the first item group with new progress in item content.

[0168] Optionally, the item screening sub-module includes:

[0169] A database establishment unit, which is used to establish a database for the parsed first report and the second report;

[0170] An item name retrieval unit, which is used to retrieve the item name in the database to obtain each item in the first report and the second report that contains the same item name.

[0171] Optionally, the data parsing module further includes:

[0172] A list text acquisition sub-module, which is used to parse the first report data and the second report data into a target format. The target format is list text, and the list text includes at least one item, and each item includes at least an item name.

[0173] Based on the same inventive concept, another embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the summary report generation method described in any of the above embodiments of the present application are implemented.

[0174] Based on the same inventive concept, another embodiment of the present application provides a computer program product, including a computer program / instructions. When the computer program / instructions are executed by a processor, the steps in the summary report generation method described in any of the above embodiments of the present application are implemented.

[0175] Based on the same inventive concept, another embodiment of the present application provides an electronic device, as Figure 3 shown. Figure 3 FIG. is a schematic diagram of an electronic device shown in an embodiment of the present application. The electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes, the steps in the summary report generation method described in any of the above embodiments of the present application are implemented.

[0176] For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiment.

[0177] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other.

[0178] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a device, or a computer program product. Therefore, the embodiments of the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0179] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing terminal devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing terminal devices generate for implementing the processes in Figure 1one or more processes and / or blocks Figure 1 means for the functions specified in one or more blocks

[0180] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device that implements the processes Figure 1 one or more processes and / or blocks Figure 1 the functions specified in one or more blocks

[0181] These computer program instructions may also be loaded onto a computer or other programmable data processing terminal device, such that a series of operation steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable terminal device provide steps for implementing the functions specified in Figure 1 one or more processes and / or blocks Figure 1 one or more blocks

[0182] Although the preferred embodiments of the embodiments of the present application have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0183] Finally, it should also be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or terminal device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising the element.

[0184] The above has introduced in detail a summary report generation method, apparatus, device, medium and product provided by the present application. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A summary report generation method, characterized in that: include: Parsing the first report data and the second report data to obtain a first report and a second report respectively, wherein the generation time of the first report data is later than the generation time of the second report data; Based on the first prompt text, the entries with the same project name are screened by a large language model to obtain a first group of entries with new progress in the entry content; the first prompt text at least includes: the entry information in the first report, the entry information in the second report, and the entry with the same project name in the first report has new progress compared with the entry in the second report; Filter the items in the first report whose project names are different from those in the second report to determine them as a second item group; The first item group is processed by the large language model based on the second prompt text to obtain a first summary content, wherein the second prompt text at least includes: item information of the first item group; project progress; The second item group is processed by the large language model based on the third prompt text to obtain a second summary content, wherein the third prompt text at least includes: item information of the second item group; project progress; A summary report is generated according to the first summary content and the second summary content.

2. The summary report generation method according to claim 1, characterized in that: The method further comprises: Generate a report template based on the fourth prompt text through the large language model, the fourth prompt text at least including: project category; project category summary content template; The first summary content and the second summary content are added to the report template according to the project category to generate the summary report.

3. The summary report generation method according to claim 2, characterized in that: The method further comprises: When the item category of each item in the first item group and the second item group is known, clustering the items in the first item group and the second item group according to the item category of each item to obtain a clustering result; When the item category of each item in the first item group and the second item group is unknown, obtaining a preset category as the item category; Classifying the items of unknown project category in the first item group and the second item group by the large language model based on a fifth prompt text, the fifth prompt text at least including: an item category and an item of unknown project category; Clustering the items in the first item group and the second item group according to the item category of each item to obtain a clustering result; A summary report is generated based on the clustering results and the summary content templates corresponding to each project category.

4. The summary report generation method according to claim 1, characterized in that: Based on the first prompt text, the entries with the same project name are screened through a large language model to obtain the first group of entries with new progress in the entry content, including: Obtaining each entry in the first report and the second report; comparing the item names of each item in the first report with the item names of each item in the second report; Filtering each entry in the first report and the second report having the same project name; Based on the first prompt text, the entries with the same project name are screened by the large language model to obtain a first entry group with new progress in entry content.

5. The summary report generation method according to claim 4, characterized in that: comparing the item names of each item in the first report with the item names of each item in the second report; Filtering the items in the first report and the second report that have the same project name, including: Establishing a database with the parsed first report and the second report; The project name is searched in the database, and each entry in the first report and the second report containing the same project name is obtained.

6. The summary report generation method according to any one of claims 1 to 5, characterized in that: Parsing the first report data and the second report data includes: The first report data and the second report data are parsed into a target format, wherein the target format is a list text, wherein the list text includes at least one entry, and each entry includes at least a project name.

7. A summary report generating device, characterized in that: The device comprises: A data parsing module, used for parsing first report data and second report data to obtain a first report and a second report respectively, wherein the generation time of the first report data is later than the generation time of the second report data; A first entry group determination module is used to filter entries with the same project name through a large language model based on a first prompt text to obtain a first entry group with new progress in entry content; the first prompt text at least includes: entry information in the first report, entry information in the second report, and new progress in the first report compared with the entry with the same project name in the second report; A second item group determination module, used for screening items in the first report whose project names are different from those in the second report, and determining them as a second item group; A first summary content determination module is used to process the first item group through the large language model based on a second prompt text to obtain a first summary content, wherein the second prompt text at least includes: item information of the first item group; project progress; A second summary content determination module is used to process the second item group through the large language model based on a third prompt text to obtain a second summary content, wherein the third prompt text at least includes: item information of the second item group; project progress; A summary report generating module is used to generate a summary report according to the first summary content and the second summary content.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the computer program is executed by the processor, the summary report generating method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the summary report generating method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the summary report generating method according to any one of claims 1 to 6 is implemented.

Images