Method, device, storage medium and electronic device for retrieval

Abstract

The embodiment of the specification discloses a method, device, storage medium and electronic equipment for retrieval, first obtaining a retrieval requirement description and a first data source input by a user, inputting the retrieval requirement description and the first data source into a target intelligent agent; then generating corresponding retrieval tool code according to the retrieval requirement description and the first data source through the target intelligent agent; finally, generating corresponding customized data retrieval tool according to the retrieval tool code through the target intelligent agent, and outputting the customized data retrieval tool, wherein the customized data retrieval tool is used for retrieving information to be retrieved input by the user in a second data source and outputting corresponding retrieval results, and the first data source and the second data source correspond to the same format and structure.

Description

Technical Field

[0001] This invention relates to computer technology, and more particularly to a method, apparatus, storage medium, and electronic device for retrieval. Background Technology

[0002] With the widespread application of large-scale language models across various fields, the text content they generate is often based on massive training data, resulting in significant text illusion. Statistics show that 57.46% of annotation tasks related to large models involve text authenticity verification. This verification process serves a dual purpose: evaluating the overall accuracy of the content generated by the large model; and providing validated, high-quality data for model training to reduce the rate of text illusion. Currently, annotators face problems such as slow search speeds and difficulty in performing authenticity verification tasks, leading to low efficiency. Summary of the Invention

[0003] The purpose of the embodiments in this specification is to provide a method, apparatus, storage medium, and electronic device for retrieval.

[0004] This specification provides a retrieval method, proposing an intelligent search approach based on intelligent agents. This method can solve the challenges of manual search in both limited-source and open-source environments, thereby improving search efficiency. By constructing a text authenticity verification process through an intelligent search agent, the manual annotation process is simplified from "reading-searching-comparing-generating answers" to only "reading + verification." The agent automatically completes cross-platform information retrieval. Simultaneously, based on the code generation capabilities of a large language model, users can dynamically generate customized retrieval tools through natural language descriptions. This reduces algorithm development cycles, improves annotation efficiency, and allows non-technical personnel to create tools independently. The method systematically solves the dual pain points of low search efficiency and high tool development costs in traditional processes. The method includes: Obtain the user's input description of the search request and the first data source, and input the description of the search request and the first data source into the target intelligent agent; The target intelligent agent generates corresponding retrieval tool code based on the retrieval requirement description and the first data source. The target intelligent agent generates a corresponding customized data retrieval tool based on the retrieval tool code, and outputs the customized data retrieval tool. The customized data retrieval tool is used to retrieve the information to be retrieved by the user in the second data source and output the corresponding retrieval results. The first data source and the second data source have the same format and structure.

[0005] Further, the step of generating corresponding retrieval tool code by the target intelligent agent based on the retrieval requirement description and the first data source includes: The target intelligent agent identifies key elements of the retrieval requirement description and the first data source to obtain at least one key element. The target intelligent agent generates corresponding retrieval tool code based on at least one key element.

[0006] Furthermore, the at least one key element includes at least one of the following: Data source format; Data source structure; Search criteria; Filtering rules; Output format requirements.

[0007] Furthermore, the retrieval tool code includes first code information for implementing the retrieval algorithm.

[0008] Furthermore, the retrieval tool code also includes second code information for data loading and / or data preprocessing.

[0009] Furthermore, the retrieval tool code also includes third code information for post-processing of retrieval results.

[0010] This specification also provides an embodiment of a retrieval method, the method comprising: Obtain the search query information input by the user, and input the search query information into the target intelligent agent; The target intelligent agent identifies the user intent corresponding to the search query information, and generates new target query information based on the identification result, wherein the target query information covers the user intent; The target intelligent agent inputs the target query information into a preset search engine for retrieval, obtains the corresponding retrieval results, and outputs the retrieval results.

[0011] Furthermore, the step of generating new target query information based on the recognition result includes: Based on the recognition results, the retrieval query information is broken down into multiple sub-intents; New target query information is generated based on the multiple sub-intents.

[0012] Furthermore, the step of generating new target query information based on the recognition result includes: Based on the identification results, at least one vaguely descriptive text is determined from the retrieval query information; The retrieval query information is rewritten based on at least one descriptive fuzzy text to generate new target query information.

[0013] Further, the step of inputting the target query information into a preset search engine through the target intelligent agent to retrieve the corresponding search results includes: The target intelligent agent inputs the target query information and the corresponding retrieval constraint information into a preset search engine for retrieval, and obtains the corresponding retrieval results.

[0014] Furthermore, the method also includes: Obtain the retrieval constraint information input by the user in response to the retrieval query information, and input the retrieval constraint information into the target intelligent agent.

[0015] Furthermore, the method also includes: The target agent determines the retrieval constraint information corresponding to the target query information based on the recognition results.

[0016] Furthermore, the output of the search results includes: The target intelligent agent performs webpage source detection on the search results, and if the detection is successful, the search results are output.

[0017] Furthermore, the output of the search results includes: The target agent performs a cross-comparison of the search results, and if the comparison is successful, the search results are output.

[0018] This specification also provides an embodiment of a retrieval device, comprising: The first input module is used to obtain the user's input description of the search request and the first data source, and input the search request description and the first data source into the target intelligent agent; The generation module is used to generate corresponding retrieval tool code based on the retrieval requirement description and the first data source through the target intelligent agent; The first output module is used to generate a corresponding customized data retrieval tool based on the retrieval tool code through the target intelligent agent, and output the customized data retrieval tool. The customized data retrieval tool is used to retrieve the information to be retrieved by the user in the second data source and output the corresponding retrieval results. The first data source and the second data source have the same format and structure.

[0019] This specification also provides an embodiment of a retrieval device, comprising: The second input module is used to obtain the search query information input by the user and input the search query information into the target intelligent agent; The identification module is used to identify the user intent corresponding to the search query information through the target intelligent agent, and generate new target query information based on the identification result, wherein the target query information covers the user intent; The second output module is used to input the target query information into a preset search engine through the target intelligent agent for retrieval, obtain the corresponding search results, and output the search results.

[0020] This specification also provides a storage medium storing a computer program adapted to be loaded by a processor and to execute the steps of the method described above.

[0021] This specification also provides an electronic device, including a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to execute the steps of the method described above.

[0022] This specification also provides a computer program product having at least one instruction stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0023] According to the embodiments of this specification, an intelligent search method based on intelligent agents is proposed, which can solve the problem of manual search in both limited-source and open-source environments, thereby improving search efficiency. By constructing a text authenticity verification process through intelligent search agents, the manual annotation process is simplified from "reading-searching-comparing-generating answers" to only "reading + verification". The intelligent agent automatically completes cross-platform information retrieval. At the same time, based on the code generation capability of a large language model, users can dynamically generate customized search tools through natural language descriptions, which can reduce the algorithm development cycle, improve annotation efficiency, and allow non-technical personnel to create tools independently. This systematically solves the dual pain points of low search efficiency and high tool development costs in traditional processes. Attached Figure Description

[0024] Figure 1 A flowchart illustrating a retrieval method provided in an embodiment of this specification; Figure 2 A flowchart illustrating a retrieval method provided in an embodiment of this specification; Figure 3 A flowchart illustrating a method for generating a customized data retrieval tool, provided as an embodiment of this specification; Figure 4 A schematic diagram of a retrieval device provided in an embodiment of this specification; Figure 5 A schematic diagram of a retrieval device provided in an embodiment of this specification; Figure 6This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this specification will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of them. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0026] Please see Figure 1 This is a flowchart illustrating a retrieval method provided in an embodiment of this specification. In this embodiment, the retrieval method is applied to a retrieval device (hereinafter referred to as a "first retrieval device") or an electronic device equipped with a first retrieval device. The following will focus on... Figure 1 The process shown will be described in detail. The retrieval method may specifically include the following steps: S102, obtain the user's input description of the search request and the first data source, and input the description of the search request and the first data source into the target intelligent agent.

[0027] In some embodiments, the retrieval requirement description is a retrieval requirement described in natural language. Natural language refers to a symbolic system that humans naturally form in social interactions, evolve with culture, and use to communicate thoughts and emotions. It is distinct from computer programming languages, mathematical / logical formal languages, and artificial languages. Natural language is a language that humans are born with, learn, and evolve, including but not limited to spoken language, written language, and sign language. This example embodiment does not make any special limitations on this. The retrieval requirement is the specific information content, scope, type, and quality requirements that a user (e.g., a classmate) expects to obtain from a fixed information source in order to solve a problem or complete a task. It is the starting point of the retrieval behavior and is used to locate relevant content from the fixed information source. In some embodiments, the first data source is a fixed information source, that is, the retrieval scope is limited to a pre-provided structured data source. For example, the first data source can be an Excel spreadsheet or a database. This example embodiment does not make any special limitations on this. In some embodiments, the retrieval requirement description and the first data source are input into a target intelligent agent. An intelligent agent refers to an artificial intelligence agent system that can autonomously perform specific tasks.

[0028] S104, the target intelligent agent generates corresponding retrieval tool code based on the retrieval requirement description and the first data source.

[0029] In some embodiments, the target agent is a code agent, which is an AI system based on a large model and possessing autonomous planning, tool invocation, code execution, and iterative optimization capabilities. It can translate natural language requirements into executable development actions and autonomously complete end-to-end software development tasks from requirements analysis to code implementation, debugging, testing, and refactoring. In some embodiments, the target agent generates complete retrieval tool code by parsing and understanding the format and structure of the retrieval requirement description and the first data source, thus constructing a complete retrieval and usage logic chain.

[0030] S106, the target intelligent agent generates a corresponding customized data retrieval tool according to the retrieval tool code, and outputs the customized data retrieval tool. The customized data retrieval tool is used to retrieve the information to be retrieved by the user in the second data source and output the corresponding retrieval results. The first data source and the second data source have the same format and structure.

[0031] In some embodiments, the target agent generates a corresponding, directly usable, customized data retrieval tool based on the retrieval tool code and outputs the customized data retrieval tool. In some embodiments, users can use this customized data retrieval tool to retrieve user-input information from a second data source (i.e., fixed-source retrieval, where the retrieval scope is limited to a pre-provided structured data source) and output the corresponding retrieval results. The retrieval results are used to characterize the authenticity of the information to be retrieved. The first data source and the second data source correspond to the same format and structure; for example, the first data source and the second data source are the same data source, or they are two different data sources corresponding to the same format and structure. In some embodiments, the target agent, based on automatically generated code (i.e., the retrieval tool code), performs function encapsulation, parameter solidification, exception handling, and UI / command-based packaging to finally output an independent, usable tool—the customized data retrieval tool—that users can run without programming, modification, or copying. In some embodiments, users need to review and confirm the retrieval results output by the customized data retrieval tool. According to the embodiments of this specification, a customized data retrieval tool is automatically generated by the target intelligent agent, simplifying the annotation process of "reading -> searching -> comparing -> generating answers" to only the "reading" step. The rest is automatically completed by the target intelligent agent, with the annotation personnel only needing to confirm and verify. Based on the code generation and understanding capabilities of the target intelligent agent, and with a fixed information source, the embodiment of this specification can dynamically generate customized data retrieval tools adapted to different tasks based on natural language descriptions.

[0032] In some embodiments, generating corresponding retrieval tool code by the target intelligent agent based on the retrieval requirement description and the first data source includes: identifying key elements of the retrieval requirement description and the first data source by the target intelligent agent to obtain at least one key element; and generating corresponding retrieval tool code by the target intelligent agent based on the at least one key element. In some embodiments, identifying key elements of the retrieval requirement description and the first data source by the target intelligent agent to obtain at least one key element in the retrieval requirement description and the first data source, the key element refers to a structured core information unit that is necessary for completing the retrieval task and can be directly used by subsequent modules, located and extracted by the target intelligent agent from the retrieval requirement description and the first data source. In some embodiments, the target intelligent agent generates corresponding retrieval tool code based on at least one key element, for example, by taking the extracted structured key element as input, automatically selecting a code template, filling in variables, and concatenating logic according to the retrieval task intent, and finally generating directly executable code.

[0033] In some embodiments, the at least one key element includes at least one of the following: data source format; data source structure; search criteria; filtering rules; and output format requirements. In some embodiments, the data source format refers to the file type, encoding standard, and syntax rules adopted by the data source, determining how the data in the data source is read, written, and parsed; it is the external manifestation of the data source. In some embodiments, the data source structure refers to the logical organization of the data in the data source, including but not limited to field names, field types, hierarchical relationships, primary / foreign key associations, array nesting, table row and column structures, etc., which are not specifically limited in this example embodiment. In some embodiments, search criteria refer to semantic matching or keyword matching rules proposed by the user for conducting a search, used to locate the information to be searched in a fixed information source; it is the core intent expression of the search task. In some embodiments, filtering rules are constraints for secondary filtering of the retrieved data, typically based on precise field values, ranges, and Boolean logic, used to exclude irrelevant or invalid data and improve the accuracy of the results. In some embodiments, output format requirements are constraints on the presentation format, field subsets, structural specifications, and display methods of the final results after search filtering, ensuring that the results can be directly read by users, parsed by the system, or used by downstream tools.

[0034] In some embodiments, the retrieval tool code includes first code information for implementing the retrieval algorithm. In some embodiments, the retrieval algorithm refers to a type of calculation method that, given a data source and information to be retrieved, automatically locates, filters, and sorts retrieval results from the data source based on the information to be retrieved according to preset matching rules, similarity calculation methods, or sorting strategies, and returns the retrieval results.

[0035] In some embodiments, the retrieval tool code further includes second code information for data loading and / or data preprocessing. In some embodiments, data loading refers to the process of reading the raw data from an external storage medium into memory or a computing environment, according to the format and structure of a specified data source, before executing a retrieval task, to form a data object accessible to the model. In some embodiments, data preprocessing refers to a series of standardized operations performed after data loading, including cleaning, transformation, normalization, structuring, and enhancement of the raw data from the data source to eliminate noise, redundancy, and format inconsistencies, thereby improving data quality and retrieval efficiency.

[0036] In some embodiments, the retrieval tool code further includes third-party code information for post-processing of retrieval results. In some embodiments, post-processing of retrieval results refers to a series of processing operations performed on the retrieval results before outputting them to the user, including normalization, filtering, fusion, rearrangement, deduplication, cropping, and format conversion, aiming to improve the accuracy, usability, compliance, and adaptability of the results.

[0037] Please see Figure 2 This is a flowchart illustrating a retrieval method provided in an embodiment of this specification. In this embodiment, the retrieval method is applied to a retrieval device (hereinafter referred to as a "second retrieval device") or an electronic device equipped with a second retrieval device. The following will focus on... Figure 2 The process shown will be described in detail. The retrieval method may specifically include the following steps: S202, obtain the search query information input by the user, and input the search query information into the target intelligent agent.

[0038] In some embodiments, the retrieval query information is a retrieval request submitted by a user or intelligent agent to a retrieval system (e.g., a search engine). It serves as the sole input for the retrieval system to perform data matching, recall, and ranking, clarifying what content needs to be retrieved. The retrieval query information is the initial input unit for the retrieval task and can be natural language text, keywords, semantic vectors, structured conditions, or combinations thereof. The retrieval system, by parsing the intent, key elements, retrieval conditions, and constraints of the retrieval query information, drives the retrieval algorithm to locate and return relevant data from the data source. In some embodiments, the retrieval query information includes information to be retrieved.

[0039] S204, the target intelligent agent identifies the user intent corresponding to the search query information, and generates new target query information based on the identification result, wherein the target query information covers the user intent.

[0040] In some embodiments, the target agent identifies the user intent corresponding to the retrieval query information. For example, after receiving the retrieval query information, the target agent automatically parses and structures the identification results, including the user's true retrieval target, core needs, constraints, and expected output format, through semantic understanding, key element extraction, intent classification, and slot filling. In some embodiments, new target query information is generated based on the identification results. For example, the target agent uses the parsed identification results, including the user's true retrieval target, core needs, constraints, and expected output format, to perform semantic reconstruction, element completion, terminology standardization, and search-friendly processing on the original retrieval query information, generating new target query information that is more accurate, more complete in elements, less ambiguous, and adapted to the retrieval algorithm, thereby improving the recall and relevance of the retrieval system. The target query information covers all user intents of the original retrieval query information and the intents are clear. The target query information includes the information to be retrieved.

[0041] S206, the target intelligent agent inputs the target query information into a preset search engine for retrieval, obtains the corresponding retrieval results, and outputs the retrieval results.

[0042] In some embodiments, the target agent inputs the target query information into a preset search engine, which then searches for the information to be retrieved (i.e., open-source retrieval, a retrieval conducted in the open environment of the Internet without a fixed search source, typically requiring verification of the information's authenticity across multiple websites), obtains the corresponding search results, and outputs the search results to the user. The search results are used to characterize the authenticity of the information to be retrieved. In some embodiments, the user needs to review and confirm the search results output by the target agent.

[0043] In some embodiments, generating new target query information based on the recognition result includes: splitting the retrieval query information into multiple sub-intents based on the recognition result; and generating new target query information based on the multiple sub-intents. In some embodiments, where the original retrieval query information contains multiple intents, the target agent, through semantic parsing and intent recognition, decomposes the retrieval query information containing multiple intents into two or more independent, semantically complete, and individually executable sub-intents. Each sub-intent corresponds to an independent retrieval target, and they are not included or substituted for each other. Then, the target agent performs semantic normalization, element completion, and retrieval-friendly reconstruction on each independent sub-intent, generating one or more accurate new target query information that can be directly used for retrieval for each sub-intent, ultimately forming multiple new target query information.

[0044] In some embodiments, generating new target query information based on the recognition result includes: determining at least one vaguely descriptive text in the retrieval query information based on the recognition result; and rewriting the retrieval query information based on the at least one vaguely descriptive text to generate new target query information. In some embodiments, the original retrieval query information contains vague descriptions. After the target agent completes intent recognition of the original retrieval query information, it locates at least one text fragment in the retrieval query information that is vaguely descriptive, semantically ambiguous, unclear in reference, or broad in expression, i.e., vaguely descriptive text, based on the recognition result. Using this vaguely descriptive text as the optimization object, the original retrieval query information is rewritten through semantic clarification, reference resolution, element completion, scope limitation, and terminology standardization, ultimately generating new target query information that is clear, unambiguous, and has higher retrieval accuracy.

[0045] In some embodiments, the step of inputting the target query information into a preset search engine through the target intelligent agent to retrieve the corresponding search results includes: inputting the target query information and corresponding search constraint information into the preset search engine through the target intelligent agent to retrieve the corresponding search results. In some embodiments, search constraint information refers to restrictive conditions used to limit the search scope, filter results, exclude irrelevant content, and ensure that the results meet user requirements. Search constraint information includes, but is not limited to, search scope constraints and source confidence constraints, etc., which are not specifically limited in this example embodiment. In some embodiments, the target intelligent agent inputs the target query information and corresponding search constraint information into the preset search engine, so that the search engine retrieves the information to be retrieved based on the search constraint information and obtains the corresponding search results.

[0046] In some embodiments, the method further includes: obtaining retrieval constraint information input by the user for the retrieval query information, and inputting the retrieval constraint information into the target intelligent agent. In some embodiments, the retrieval constraint information input by the user for the original retrieval query information is obtained first, and then the retrieval constraint information is also input into the target intelligent agent. This can be done by inputting the retrieval query information and the retrieval constraint information into the target intelligent agent together, or by inputting the retrieval query information and the retrieval constraint information into the target intelligent agent sequentially.

[0047] In some embodiments, the method further includes: determining retrieval constraint information corresponding to the target query information by the target agent based on the recognition result. In some embodiments, after performing intent recognition on the original retrieval query information, restrictive conditions for limiting the retrieval scope, filtering results, and excluding irrelevant data are selected, extracted, and normalized from the identified structured information, i.e., the recognition result, to form clear, unambiguous, and directly usable structured retrieval constraint information.

[0048] In some embodiments, outputting the search results includes: performing webpage source detection on the search results by the target AI agent; if the detection passes, the search results are output. In some embodiments, the target AI agent performs webpage source detection on the search results; if the source confidence of the webpage source corresponding to the search results is greater than or equal to a preset threshold, or if the type of the webpage source corresponding to the search results belongs to a preset type (e.g., a specific organization's official website type, a leading company in the industry type, etc.), then the detection is determined to have passed, and the search results are only output if the detection passes.

[0049] In some embodiments, outputting the search results includes: cross-comparing the search results by the target agent; if the comparison passes, the search results are output. In some embodiments, the target agent performs cross-comparison of the search results. Cross-comparison refers to comparing whether multiple search results from multiple different web page sources are consistent with the information to be searched. If they are consistent, the comparison is determined to pass, and the search results are only output if the comparison passes.

[0050] Figure 3 This is a flowchart illustrating a method for generating a customized data retrieval tool, as provided in an embodiment of this specification.

[0051] like Figure 3 As shown, the first step is user input, where users input natural language descriptions of their needs. Operations personnel then describe their needs in natural language and upload an Excel spreadsheet. The second step is intelligent parsing, where a large model centered on a code agent deeply understands the needs and the spreadsheet structure. The third step is automated development, where AI (artificial intelligence) automatically writes code to build a complete retrieval and usage logic chain. The fourth step is tool delivery, generating a customizable data retrieval tool that can be used directly. This enables rapid, coding-free customization of data retrieval tools, improving operational efficiency.

[0052] Figure 4 This is a schematic diagram of a retrieval device provided in an embodiment of this specification. The retrieval device (hereinafter referred to as "first retrieval device 1") can be implemented as all or part of an electronic device through software, hardware, or a combination of both. According to some embodiments, the first retrieval device 1 includes a first input module 11, a generation module 12, and a first output module 13.

[0053] The first input module 11 is used to obtain the user's input description of the search request and the first data source, and input the search request description and the first data source into the target intelligent agent; The generation module 12 is used to generate corresponding retrieval tool code based on the retrieval requirement description and the first data source through the target intelligent agent; The first output module 13 is used to generate a corresponding customized data retrieval tool through the target intelligent agent according to the retrieval tool code, and output the customized data retrieval tool. The customized data retrieval tool is used to retrieve the information to be retrieved by the user in the second data source and output the corresponding retrieval results. The first data source and the second data source have the same format and structure.

[0054] In some embodiments, the generation module 12 is configured to: identify key elements of the retrieval requirement description and the first data source through the target intelligent agent to obtain at least one key element; and generate corresponding retrieval tool code through the target intelligent agent based on the at least one key element.

[0055] In some embodiments, the at least one key element includes at least one of the following: data source format; data source structure; retrieval criteria; filtering rules; and output format requirements.

[0056] In some embodiments, the retrieval tool code includes first code information for implementing the retrieval algorithm.

[0057] In some embodiments, the retrieval tool code further includes second code information for data loading and / or data preprocessing.

[0058] In some embodiments, the retrieval tool code may further include third code information for post-processing of retrieval results.

[0059] The above-described apparatus embodiments correspond to the aforementioned method embodiments. For detailed descriptions, please refer to the description in the method embodiments section; further details will not be repeated here. The apparatus embodiments are derived from the corresponding method embodiments and have the same technical effects. For detailed descriptions, please refer to the corresponding method embodiments.

[0060] Figure 5 This is a schematic diagram of a retrieval device provided in an embodiment of this specification. This retrieval device (hereinafter referred to as "second retrieval device 2") can be implemented as all or part of an electronic device through software, hardware, or a combination of both. According to some embodiments, the second retrieval device 2 includes a second input module 21, an identification module 22, and a second output module 23. The second input module 21 is used to obtain retrieval query information input by a user and input the retrieval query information into a target intelligent agent; the identification module 22 is used to identify the user intent corresponding to the retrieval query information through the target intelligent agent, and generate new target query information based on the identification result, wherein the target query information covers the user intent; the second output module 23 is used to input the target query information into a preset search engine through the target intelligent agent for retrieval, obtain the corresponding retrieval results, and output the retrieval results.

[0061] In some embodiments, generating new target query information based on the identification result includes: splitting the retrieval query information into multiple sub-intents based on the identification result; and generating new target query information based on the multiple sub-intents.

[0062] In some embodiments, generating new target query information based on the identification result includes: determining at least one descriptive fuzzy text in the retrieval query information based on the identification result; rewriting the retrieval query information based on the at least one descriptive fuzzy text to generate new target query information.

[0063] In some embodiments, the step of inputting the target query information into a preset search engine through the target intelligent agent to retrieve the corresponding search results includes: inputting the target query information and the corresponding search constraint information into a preset search engine through the target intelligent agent to retrieve the corresponding search results.

[0064] In some embodiments, the second retrieval device 2 is further configured to: obtain retrieval constraint information input by the user for the retrieval query information, and input the retrieval constraint information into the target intelligent agent.

[0065] In some embodiments, the second retrieval device 2 is further configured to: determine the retrieval constraint information corresponding to the target query information by the target agent based on the recognition result.

[0066] In some embodiments, outputting the search results includes: performing webpage source detection on the search results through the target agent, and outputting the search results if the detection is successful.

[0067] In some embodiments, outputting the search results includes: performing a cross-comparison of the search results by the target agent, and outputting the search results if the comparison is successful.

[0068] The above-described apparatus embodiments correspond to the aforementioned method embodiments. For detailed descriptions, please refer to the description in the method embodiments section; further details will not be repeated here. The apparatus embodiments are derived from the corresponding method embodiments and have the same technical effects. For detailed descriptions, please refer to the corresponding method embodiments.

[0069] This specification also provides a computer storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in this specification.

[0070] This specification also provides a computer program product that stores at least one instruction, which is loaded by the processor and executes the method described in this specification embodiment.

[0071] This specification also provides an electronic device, including a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and execute the method described in the embodiments of this specification.

[0072] The embodiments in this specification also provide Figure 6 The diagram shows the structure of the electronic device. Figure 6 At the hardware level, the electronic device includes a processor, internal bus, network interface, memory, and non-volatile memory, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then runs it to implement the above method.

[0073] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0074] Those skilled in the art will understand that embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this specification may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0075] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0076] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0077] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0078] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0079] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0080] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0081] The above description is merely an embodiment of this specification and is not intended to limit this specification. Various modifications and variations can be made to this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of the claims of this specification.

Claims

1. A method for retrieval, comprising: Obtain the user's input description of the search request and the first data source, and input the description of the search request and the first data source into the target intelligent agent; The target intelligent agent generates corresponding retrieval tool code based on the retrieval requirement description and the first data source. The target intelligent agent generates a corresponding customized data retrieval tool based on the retrieval tool code, and outputs the customized data retrieval tool. The customized data retrieval tool is used to retrieve the information to be retrieved by the user in the second data source and output the corresponding retrieval results. The first data source and the second data source have the same format and structure.

2. The method according to claim 1, wherein generating corresponding retrieval tool code by the target agent based on the retrieval requirement description and the first data source includes: The target intelligent agent identifies key elements of the retrieval requirement description and the first data source to obtain at least one key element. The target intelligent agent generates corresponding retrieval tool code based on at least one key element.

3. The method according to claim 2, wherein the at least one key element comprises at least one of the following: Data source format; Data source structure; Search criteria; Filtering rules; Output format requirements.

4. The method according to claim 1, wherein the retrieval tool code includes first code information for implementing the retrieval algorithm.

5. The method according to claim 4, wherein the retrieval tool code further includes second code information for data loading and / or data preprocessing.

6. The method according to claim 4 or 5, wherein the retrieval tool code further includes third code information for post-processing of retrieval results.

7. A method for retrieval, comprising: Obtain the search query information input by the user, and input the search query information into the target intelligent agent; The target intelligent agent identifies the user intent corresponding to the search query information, and generates new target query information based on the identification result, wherein the target query information covers the user intent; The target intelligent agent inputs the target query information into a preset search engine for retrieval, obtains the corresponding retrieval results, and outputs the retrieval results.

8. The method according to claim 7, wherein generating new target query information based on the identification result includes: Based on the recognition results, the retrieval query information is broken down into multiple sub-intents; New target query information is generated based on the multiple sub-intents.

9. The method according to claim 7, wherein generating new target query information based on the identification result includes: Based on the identification results, at least one vaguely descriptive text is determined from the retrieval query information; The retrieval query information is rewritten based on at least one descriptive fuzzy text to generate new target query information.

10. The method according to claim 7, wherein the step of inputting the target query information into a preset search engine through the target intelligent agent to retrieve the corresponding search results includes: The target intelligent agent inputs the target query information and the corresponding retrieval constraint information into a preset search engine for retrieval, and obtains the corresponding retrieval results.

11. The method of claim 10, further comprising: Obtain the retrieval constraint information input by the user in response to the retrieval query information, and input the retrieval constraint information into the target intelligent agent.

12. The method of claim 10, further comprising: The target agent determines the retrieval constraint information corresponding to the target query information based on the recognition results.

13. The method according to claim 7, wherein outputting the search result includes: The target intelligent agent performs webpage source detection on the search results, and if the detection is successful, the search results are output.

14. The method according to claim 7 or 13, wherein outputting the search result includes: The target agent performs a cross-comparison of the search results, and if the comparison is successful, the search results are output.

15. An apparatus for retrieval, comprising: The first input module is used to obtain the user's input description of the search request and the first data source, and input the search request description and the first data source into the target intelligent agent; The generation module is used to generate corresponding retrieval tool code based on the retrieval requirement description and the first data source through the target intelligent agent; The first output module is used to generate a corresponding customized data retrieval tool based on the retrieval tool code through the target intelligent agent, and output the customized data retrieval tool. The customized data retrieval tool is used to retrieve the information to be retrieved by the user in the second data source and output the corresponding retrieval results. The first data source and the second data source have the same format and structure.

16. An apparatus for retrieval, comprising: The second input module is used to obtain the search query information input by the user and input the search query information into the target intelligent agent; The identification module is used to identify the user intent corresponding to the search query information through the target intelligent agent, and generate new target query information based on the identification result, wherein the target query information covers the user intent; The second output module is used to input the target query information into a preset search engine through the target intelligent agent for retrieval, obtain the corresponding search results, and output the search results.

17. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 14.

18. An electronic device, characterized in that, include: A processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to execute the steps of the method as claimed in any one of claims 1 to 14.

19. A computer program product having at least one instruction stored thereon, characterized in that, When the at least one instruction is executed by the processor, it implements the steps of the method according to any one of claims 1 to 14.

Images