Document processing method and device, equipment and medium

Abstract

The embodiment of the invention provides a document processing method and device, equipment and a medium, and relates to the technical field of data processing.The method comprises the steps that subject segmentation is carried out on a to-be-processed document to obtain all subjects included in the to-be-processed document, content segmentation is carried out on each subject, and content slices included in all the subjects are obtained. According to the method, the content slices of the same topic are associated with the topics to which the content slices belong, the topics are summarized to obtain summarized slices, and the summarized slices are subjected to recursion to obtain a processing result representing the hierarchical relationship of the to-be-processed document, so that the accuracy of document processing is ensured.

Claims

1. A document processing method, characterized in that: include: Performing topic segmentation on the document to be processed to obtain the topics included in the document to be processed; Segment the content of each of the topics to obtain content slices included in each of the topics; The content slices of the same topic are associated with the topics to which they belong, and each of the topics is summarized to obtain summary slices, and the processing results representing the hierarchical relationship of the documents to be processed are recursively obtained.

2. The document processing method according to claim 1, characterized in that: The subject segmentation of the document to be processed to obtain the subjects included in the document to be processed includes: Performing regular matching between the document to be processed and each preset regular expression to determine a target regular expression that matches the document to be processed, wherein the regular expression includes a document level rule; The topics included in the document to be processed are segmented based on the document level rules in the target regular expression.

3. The document processing method according to claim 2, characterized in that: The summarizing of each of the topics to obtain summary slices and recursively obtaining processing results representing the hierarchical relationship of the documents to be processed include: Based on the document level rules in the target regular expression, summarize each of the topics to obtain summary slices; A conceptual graph representing the hierarchical relationship of the documents to be processed is recursively obtained based on the summary slices.

4. The document processing method according to claim 3, characterized in that: The method of summarizing each of the topics based on the document hierarchy rules in the target regular expression to obtain summary slices, and recursively obtaining a concept map representing the hierarchical relationship of the document to be processed based on the summary slices includes: Determine the segmentation granularity of each of the topics in the to-be-processed document based on the document-level rule in the target regular expression; Deriving an initial document hierarchical structure based on the segmentation granularity; Adding each of the content slices to each leaf node of the initial document hierarchy structure; Summarize the title and content slice of each of the topics into a first summary slice, add the first summary slice to the Nth-level intermediate node of the initial document hierarchy, and link the Nth-level intermediate node corresponding to each of the topics with the leaf node corresponding to the content slice belonging to the topic; wherein N is an integer greater than or equal to 2; Summarize the first summary slices belonging to the same N-1th level classification into second summary slices, add the second summary slices to the N-1th level intermediate nodes of the initial document hierarchy structure, and link the N-1th level intermediate nodes where the first summary slices are located with the N-1th level intermediate nodes corresponding to the N-1th level classifications to which they belong; Until each summary slice of the NM-th level classification is summarized into an overall summary slice added to the root node of the initial document hierarchy structure, and each NM-th layer intermediate node is linked to the root node; wherein M is an integer greater than or equal to 2.

5. The document processing method according to claim 4, characterized in that: Each summary slice is obtained by calling the large model summary, and the method further includes: Vectorizing the slices in each of the leaf nodes, the intermediate nodes, and the root node; The results of the vectorization processing are stored in a vector library, and the nodes to which the results of the vectorization processing belong in the concept graph are identified.

6. A document processing method, characterized in that: include: Receiving a search request for a document to be processed, wherein the search request includes information to be retrieved; Retrieving a target slice matching the information to be retrieved from the processing result obtained by the document processing method according to any one of claims 1 to 5 and representing the hierarchical relationship of the document to be processed; Based on the matching degree with the information to be retrieved, sorting each of the target slices to obtain a sorting result; A search result matching the information to be searched is determined based on the ranking result.

7. The document processing method according to claim 6, characterized in that: The processing result is a concept map, which includes content slices and summary slices corresponding to the document to be processed; The step of sorting the target slices based on the matching degree with the information to be retrieved to obtain a sorting result includes: When each of the target slices includes a content slice, all the content slices in the target slice are sorted separately to obtain a content slice sorting result; When each of the target slices includes a summary slice, all summary slices in the target slices are sorted separately to obtain a summary slice sorting result; The step of determining the search result matching the information to be searched based on the ranking result includes: Determine a first sub-search result matching the information to be retrieved based on the content slice sorting result; Determine a second sub-search result matching the information to be searched based on the summary slice sorting result; The first sub-search result and the second sub-search result are used as search results that match the information to be searched.

8. A document processing device, characterized in that: include: An information acquisition module is used to divide the document to be processed into topics and obtain the topics included in the document to be processed; The information processing module is used to divide the content of each of the topics separately to obtain content slices included in each of the topics; associate the content slices of the same topic with the topic to which it belongs, summarize each of the topics to obtain summary slices, and recursively obtain processing results that characterize the hierarchical relationship of the documents to be processed.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the document processing method according to any one of claims 1 to 7 when executing the program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a computer program, and when the computer program is executed, the electronic device where the computer-readable storage medium is located is controlled to execute the document processing method according to any one of claims 1 to 7.

Images