Document analysis method, device, computer storage medium and equipment

Abstract

The present invention provides a document analysis method, the method comprising: displaying target document information in a first interface area; displaying a target tag set in a second interface area through a first tree control, the target tag set comprising tag items with a parent-child relationship, the tag items having tag identifiers, the tag identifiers expressing the parent-child relationship of the corresponding tag items, the tag items being displayed by nodes of the first tree control; receiving an indexing instruction, obtaining a first target node pointed to by the indexing instruction, obtaining a first target tag identifier of the tag item corresponding to the first target node; storing the first target tag identifier in an indexing record corresponding to the target document information and the target tag set, the indexing record comprising one or more tag identifiers. The above document analysis method is easy to operate, and because a standardized analysis feature specification is pre-set, the accuracy of the analysis is high.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a document analysis method, device, computer storage medium and equipment. Background Art

[0002] With the development of information technology, more and more documents are transformed from paper to electronic form and then spread through the Internet. In order to deeply analyze the content and value of papers, patents or other documents, it is usually necessary to perform document analysis on them. The common method of document analysis is to add tags to the documents.

[0003] In traditional technology, the analysis method for batches of documents is usually to use Excel tables for analysis, pre-set multiple attribute items or labels, and index the documents one by one. The indexing method is: fill in the attribute items or labels related to the document in the cells of a row corresponding to the document number in the Excel table. However, due to the limitations of the Excel table, for the convenience of statistics, one document usually corresponds to one row, and only one label or one path label can be filled in one row. In addition, if the label item is adjusted, all the documents that have been indexed with the label need to be reviewed, and the re-indexing accompanied by the adjustment of the label item is very inconvenient.

[0004] As for some patent analysis systems, although they provide patent indexing services, the label items used for indexing by these patent analysis systems are usually summarized and input by users based on the subject of the document, or clustered and generated by the AI ​​subject extraction function of the patent analysis system. They are unable to form a set of label items with a certain logical relationship, resulting in low analysis accuracy. Summary of the invention

[0005] Based on this, it is necessary to provide a document analysis method that is easy to operate and improves the accuracy of document analysis in response to the technical problems in the above-mentioned background technology.

[0006] A document analysis method, the method comprising:

[0007] Display target document information in the first interface area;

[0008] Displaying a target tag set in the second interface area through a first tree control, the target tag set includes tag items with a parent-child relationship, the tag items have tag identifiers, the tag identifiers express the parent-child relationship of the corresponding tag items, the tag items are displayed by nodes of the first tree control, and the parent-child relationship of the tag items is consistent with the parent-child relationship of the corresponding nodes on the first tree control;

[0009] Receive an indexing instruction, obtain a first target node pointed to by the indexing instruction, and obtain a first target label identifier of a label item corresponding to the first target node;

[0010] The first target tag identifier is stored in an indexing record corresponding to the target document information and the target tag set, wherein the indexing record includes one or more tag identifiers.

[0011] In one of the embodiments, the indexing record includes one or more sub-partitions, and the indexing instruction includes a sub-partition identifier;

[0012] The storing the first target tag identifier into the indexing record corresponding to the target document information and the target tag set includes:

[0013] Get the target sub-partition identifier corresponding to the indexing instruction;

[0014] An indexing record corresponding to the target document information and the target tag set is obtained, and the first target tag identifier is stored in a sub-partition corresponding to the target sub-partition identifier.

[0015] In one embodiment, the method comprises:

[0016] Displaying the target tag set in the third interface area through a second tree control, and receiving a tag setting instruction for the second tree control;

[0017] A second target node pointed to by the label setting instruction is obtained, and a label item corresponding to the second target node is modified according to the label setting instruction.

[0018] In one embodiment, after modifying the label item corresponding to the second target node according to the label setting instruction, the method further includes:

[0019] When the label identifier of the label item of the second target node is modified, recording the modification record of the label item in the modification order;

[0020] Acquire a target indexing record set corresponding to the target tag set, wherein the target indexing record set includes one or more indexing records;

[0021] The tag identifier in the indexing record in the target indexing record set is modified according to the modification record.

[0022] In one embodiment, the second target node is one or more, the modification record includes one or more, and the modification record includes a pre-modification identifier and a post-modification identifier;

[0023] The step of modifying the label identifier in the indexing record in the target indexing record set according to the modification record comprises:

[0024] Traversing the indexing records in the target indexing record set;

[0025] For the traversed indexing records, the modification records are obtained one by one in sequence, the pre-modification identifier and the post-modification identifier contained in the modification record are obtained, and the pre-modification identifier contained in the traversed indexing record is modified to the post-modification identifier.

[0026] In one embodiment, before modifying the label identifier in the indexing record in the target indexing record set according to the modification record, the method further includes:

[0027] Merge two or more modification records for the same second target node.

[0028] In one embodiment, before modifying the label identifier in the indexing record in the target indexing record set according to the modification record, the method further includes:

[0029] The target tag set and the indexing record corresponding to the target tag set are backed up.

[0030] In one embodiment, modifying the label item corresponding to the second target node according to the label setting instruction includes:

[0031] When the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node does not conform to the parent-child relationship of the second target node in the second tree control, an error prompt is given, or the label identifier is automatically adjusted so that the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node conforms to the parent-child relationship of the second target node in the second tree control.

[0032] In one embodiment, the method further comprises:

[0033] Obtaining an input screening expression, wherein the screening expression includes one or more tag identifiers in the target tag set;

[0034] Obtain an input document information set, and obtain indexing records corresponding to documents in the document information set and a target tag set;

[0035] The documents corresponding to the indexing records whose tag identifiers meet the filter expression are filtered out.

[0036] In addition, it is necessary to provide a document analysis device that is easy to operate and improves the accuracy of document analysis in response to the technical problems in the above-mentioned background technology.

[0037] A document analysis device, comprising:

[0038] A document display module, used to display target document information in the first interface area;

[0039] A label set display module, used for displaying a target label set in a second interface area through a first tree control, wherein the target label set includes label items with a parent-child relationship, the label items have label identifiers, the label identifiers express the parent-child relationship of the corresponding label items, the label items are displayed by nodes of the first tree control, and the parent-child relationship of the label items is consistent with the parent-child relationship of the corresponding nodes on the first tree control;

[0040] A label indexing module, used to receive an indexing instruction, obtain a first target node pointed to by the indexing instruction, and obtain a first target label identifier of a label item corresponding to the first target node;

[0041] The indexing record storage module is used to store the first target tag identifier in an indexing record corresponding to the target document information and the target tag set, wherein the indexing record includes one or more tag identifiers.

[0042] In one embodiment, the indexing record includes one or more sub-partitions, and the indexing instruction includes a sub-partition identifier;

[0043] The indexing record storage module is used to obtain the target sub-partition identifier corresponding to the indexing instruction; obtain the indexing record corresponding to the target document information and the target tag set, and store the first target tag identifier in the sub-partition corresponding to the target sub-partition identifier.

[0044] In one embodiment, the device also includes a label set setting module, which is used to display the target label set through a second tree control in the third interface area, receive a label setting instruction for the second tree control; obtain the second target node pointed to by the label setting instruction, and modify the label item corresponding to the second target node according to the label setting instruction.

[0045] In one embodiment, the device also includes an indexing record change module, which is used to record the modification record of the label item in the modification order when the label identifier of the label item of the second target node is modified; obtain a target indexing record set corresponding to the target label set, the target indexing record set including one or more indexing records; and modify the label identifier in the indexing record in the target indexing record set according to the modification record.

[0046] In one embodiment, the second target node is one or more, the modification record includes one or more, and the modification record includes a pre-modification identifier and a post-modification identifier;

[0047] The indexing record change module is also used to traverse the indexing records in the target indexing record set; for the traversed indexing records, the modification records are obtained one by one in sequence, the pre-modification identifier and the post-modification identifier contained in the modification record are obtained, and the pre-modification identifier contained in the traversed indexing record is modified to the post-modification identifier.

[0048] In one embodiment, the device further includes a modification record merging module, configured to merge two or more modification records for the same second target node.

[0049] In one embodiment, the device further comprises an indexing backup module, which is used to back up the target tag set and the indexing records corresponding to the target tag set.

[0050] In one embodiment, the device also includes a label item verification module, which is used to prompt an error when the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node does not conform to the parent-child relationship of the second target node in the second tree control, or automatically adjust the label identifier so that the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node conforms to the parent-child relationship of the second target node in the second tree control.

[0051] In one embodiment, the device also includes an indexing statistics module, which is used to obtain an input filtering expression, wherein the filtering expression includes one or more tag identifiers in the target tag set; obtain an input document information set, obtain indexing records corresponding to the documents in the document information set and the target tag set; and filter out documents corresponding to the indexing records whose contained tag identifiers meet the filtering expression.

[0052] In addition, it is necessary to provide a computer storage medium that stores computer instructions for executing the above-mentioned document analysis method in response to the technical problems in the above-mentioned background technology, so that the document analysis process is convenient to operate and the accuracy of document analysis is improved.

[0053] A computer storage medium stores one or more computer instructions, wherein the one or more instructions are suitable for being loaded by a processor and executing the aforementioned document analysis method.

[0054] In addition, it is necessary to provide a computer device for executing the above-mentioned document analysis method in response to the technical problems in the above-mentioned background technology, so that the document analysis process is convenient to operate and the accuracy of document analysis is improved.

[0055] A document analysis device comprises a memory and a processor, wherein the memory stores a group of programs, and the processor is used to call the programs stored in the memory to execute the document analysis method as described above.

[0056] The document analysis method, device, computer storage medium and equipment provide a convenient operation mode for document tag indexing analysis. When tagging a document, a tag set is displayed through a tree control. The nodes on the tree control correspond to tag items. The tag items have a parent-child relationship (subordinate relationship) when designed, and the parent-child relationship is expressed through tag identifiers. When the tag items are mapped to the tree control, they are also consistent with the parent-child relationship of the corresponding nodes on the tree control. The user can complete the indexing of the tag item corresponding to the node by simply operating the node on the tree control, and the operation traversal is fast.

[0057] At the same time, since the label identifier of the label item added to a document contains the parent-child relationship information of the label item (the subordinate relationship of the analysis feature classification), the user can understand the hierarchy and subordinate relationship of the label item in the entire label classification through the label identifier of the label item. Therefore, the overall classification information is expressed through only one label identifier, and the amount of information expressed is larger. The user only needs to reasonably design the label classification to obtain the content summary expressed by the target document through the feature information such as the name, description and keywords of the label item indexed for the target document, thereby realizing the analysis of the document. Compared with the method of actively reading the document manually and then writing an analysis summary, the embodiment of the present invention is a reverse analysis method, which is to pre-standardize the analysis features, and then index the features in the pre-defined specifications; therefore, the document analysis process is simpler and the analysis accuracy is more objective and accurate.

[0058] At the same time, since a tag identifier containing a parent-child relationship is set for the tag item, the statistics and screening work is simpler and more efficient. Only a simple string match is needed to filter and count the indexed records accordingly. Compared with the Excel table statistics method in traditional technology, the statistical efficiency is significantly higher and more accurate. After indexing, only a filter expression needs to be entered, and the computer can match and search according to the specific string of the tag identifier in the filter expression, so as to quickly filter out the document list or count the number of documents, making the operation simpler and the execution efficiency higher. In addition, due to the design of the tag identifier, the screening and statistics of the indexed records can be based on a flexible screening expression. The screening expression only needs to contain a specific tag identifier and use a corresponding logical expression to constrain it, so as to complete a variety of document screening and statistical work, making the document analysis work more flexible. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A flowchart of a document analysis method according to an embodiment of the present invention;

[0060] Figure 2 A schematic diagram of a document analysis page in a browser-based document analysis application scenario;

[0061] Figure 3 for Figure 2 The document analysis page of the application scenario shown is a schematic diagram showing a target tag set through a first tree control;

[0062] Figure 4 For Figure 2 A schematic diagram of inputting indexing instructions through a button component of a first tree control on a document analysis page of the application scenario shown;

[0063] Figure 5 A schematic diagram of respectively labeling a main label and a sub-label through a button component of a first tree control on a document analysis page in another document analysis application scenario;

[0064] Figure 6 A schematic diagram of labeling the main label and the secondary label respectively by dragging the nodes of the first tree control on a document analysis page in another document analysis application scenario;

[0065] Figure 7 A schematic diagram of a tag set setting page in one embodiment;

[0066] Figure 8 A schematic diagram of editing a tag item corresponding to a node through an editing window on a tag set setting page in one embodiment;

[0067] Fig. 9 A schematic diagram of a prompt mode when a modified tag identifier fails to pass the edit window verification on a tag set setting page in one embodiment;

[0068] Fig.10 A schematic diagram of a prompt mode for verifying that a modified tag identifier is passed in an edit window on a tag set setting page in an embodiment;

[0069] Fig.11 An example of a technical roadmap graphically displayed based on statistical results on an indexing record statistics page in an embodiment;

[0070] Fig.12 An example of an efficacy matrix diagram graphically displayed according to statistical results on an indexing record statistics page in an embodiment;

[0071] Fig.13 is a schematic diagram of a document analysis device in one embodiment;

[0072] Fig.14The following is an architecture diagram of a computer system for executing the document analysis method in one embodiment. DETAILED DESCRIPTION

[0073] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0074] Unless otherwise specifically and clearly described in the context, the number of elements and components in the present invention may exist in a single form or in a plurality of forms, and the present invention is not limited to this. Although the steps in the present invention are arranged with numbers, they are not used to limit the order of the steps. Unless the order of the steps is clearly stated or the execution of a certain step requires other steps as a basis, the relative order of the steps can be adjusted. It is understood that the term "and / or" used herein relates to and covers any and all possible combinations of one or more of the associated listed items.

[0075] In order to solve the above technical problems, an embodiment of the present invention specifically provides a document analysis method, which can be implemented based on a computer program and run on a computer system, especially a computer system whose processor architecture is based on the von Neumann architecture or the Harvard architecture, but is not limited to these two architectures and can be a computer system of other architectures.

[0076] The computer program is not limited to independent desktop programs, B / S (browser / server) architecture programs or C / S (client / server) architecture programs, and can process UI (User Interface) and some data processing tasks through browser programs, desktop client programs, mobile phone or tablet app applications, store data through back-end server programs, perform data retrieval and other data processing tasks. The above document analysis method can also be run through an independent desktop program, which not only processes UI, but also stores data (for example, integrates a small database for database storage or stores through a file system), performs data retrieval and other data processing tasks.

[0077] The computer system can be a personal computer, tablet computer, laptop computer, diskless workstation, smart phone, server device, etc. In the application scenario based on B / S architecture program or C / S architecture program, the computer system can process UI and perform some data processing tasks through terminal devices such as personal computers, tablet computers, laptop computers, diskless workstations, smart phones, etc., and store data, perform data retrieval and other data processing tasks through server devices.

[0078] In the application scenario of independent desktop programs, the computer system can be a personal computer, tablet computer, laptop computer, diskless workstation or smart phone, which has a database program or a data storage management program similar to a database program (such as a local database or a local file system, etc.) installed on it. The personal computer, tablet computer, laptop computer, diskless workstation or smart phone runs an independent desktop program to handle the UI, store data, perform data retrieval and other data processing tasks.

[0079] The document analysis method displays the target document information to be analyzed on a display interface through a computer program, and displays the target tag set on the display interface. The target tag set is a set consisting of one or more predefined tags, and the tags in the target tag set can express feature information of multiple dimensions. The user can select one or more tags in the target tag set through a specific instruction input method, and annotate them to the target document information, that is, the target document information is tagged. Since the tags in the target tag set express feature information of multiple dimensions when predefined, the one or more tags tagged to the target document information reflect the characteristics of the document, thereby realizing the analysis of the target document.

[0080] Specifically, in one embodiment, Figure 1 As shown, the document analysis method includes:

[0081] Step S202: display target document information in the first interface area.

[0082] refer to Figure 2 As shown, Figure 2 The application scenario of a web application for analyzing patent documents is shown. The web application is based on a B / S architecture. In this application scenario, the browser displays a document analysis page, which is divided into three areas. The area on the left that displays the list is the document list display area, the middle page area is the document information display area, and the right page area is the tag collection display area. A list component is set in the document list display area, which displays the brief information of the document such as the application number (identification information) and application name of multiple patent documents in a patent package, a patent combination or a patent list in the form of a list. One list item corresponds to one patent document. Users can select a document as the target document for analysis by clicking the list item in the list component.

[0083] In this application scenario, when the user clicks a list item to select a document as the target document, the document information display area in the middle displays the detailed information of the document. Figure 2As shown, for example, after the user selects the patent with application number: CN201811202475.3 as the target document, the document information display area can display detailed information of the patent document such as the bibliographic information, legal status information, patent family information, citation data information, claim information, specification information, etc. of the patent CN201811202475.3 through multiple tabs. The user can understand the detailed information of the target patent document by reading the content displayed in the document information display area.

[0084] In this application scenario, the list items in the document list display area on the left and the document information display area in the middle both display document information, and can be equivalent to the first interface area. The list items in the document list display area on the left display brief information of the target document, and the document information display area in the middle displays detailed information of the target document. In other application scenarios, for example, in an application scenario with a smaller mobile APP interface, only the identification (such as patent number) and name of the target document can be displayed on the document analysis page, and the detailed information of the target document needs to be displayed by clicking a link to jump to another page.

[0085] Step S204: display the target tag set in the second interface area through the first tree control.

[0086] Still Figure 2 The application scenarios shown are examples. Figure 2 In the tag set display area of ​​the page area on the right side of the document analysis page shown, the target tag set can be displayed through the first tree control. The target tag set includes tag items with parent-child relationships, the tag items have tag identifiers, the tag identifiers express the parent-child relationships of the corresponding tag items, the tag items are displayed by the nodes of the first tree control, and the parent-child relationships of the tag items are consistent with the parent-child relationships of the corresponding nodes on the first tree control.

[0087] refer to Figure 2 As shown, there is a tag set selector in the tag set display area, and the user selects the tag set loaded in the first tree control by selecting the tag set identifier in the tag set selector. If the user selects the tag set identified as "Analysis Standard A", the first tree control is loaded in the tag set display area, and the tag set of "Analysis Standard A" is expressed by the first tree control.

[0088] In this application scenario, the tag set "Analysis Standard A" used for patent analysis includes three root tags: "Technical Decomposition" (tag identifier T), "Application Classification" (tag identifier A) and "Efficacy Classification" (tag identifier E), and each root tag contains multiple levels of tags. Taking the root tag of "Efficacy Classification" as an example, the root node tag contains three sub-tags: "Physical Efficacy" (tag identifier E.1), "Chemical Efficacy" (tag identifier E.2) and "User Experience" (tag identifier E.3). The sub-tag "Physical Efficacy" includes four sub-tags: "Electrical Efficacy" (tag identifier E.11), "Optical Efficacy" (tag identifier E.12), "Thermodynamic Efficacy" (tag identifier E.13) and "Spatial Efficacy" (tag identifier E.14). The sub-tag "Optical Efficacy" includes 6 sub-tags: "High Brightness" (tag identifier E.121), "High Uniformity" (tag identifier E.122), "High Collimation" (tag identifier E.123), "Speckle Suppression" (tag identifier E.124), "High Color Purity" (tag identifier E.125) and "Others" (tag identifier E.12z).

[0089] The label identifier of the label item expresses the parent-child relationship between the label items. As mentioned above, the label item "physical effect" contains label items of four sub-effect categories: "electrical effect", "optical effect", "spatial effect" and "thermal effect", and the label identifier of the label item "physical effect" is E.1, and the label identifiers of "electrical effect", "optical effect", "spatial effect" and "thermal effect" are E.11, E.12, E.13 and E.14 respectively. All of them start with the label identifier "E.1" of "physical effect", and use the last digits "1", "2", "3" and "4" to distinguish the four sub-label items respectively. Then, through E.11, E.12, E.13 and E.14 can express that their common parent label item is E.1; similarly, for the sub-label item labels: "high brightness" (label identification E.121), "high uniformity" (label identification E.122), "high collimation" (label identification E.123), "speckle suppression" (label identification E.124), "high color purity" (label identification E.125) and "others" (label identification E.12z), the label identifications are E.121, E.122, E.123, E.124, E.125 and E.12z, all starting with E.12. The above 6 sub-labels have a common parent label item E.12, that is, the above 6 sub-label items representing efficacy classification all belong to "optical efficacy".

[0090] The label items and the parent-child relationship of the label items are expressed by the nodes of the first tree control and the relationship between the nodes. Figure 2As shown, the first tree control loaded in the label set display area includes multiple nodes, each of which corresponds to a label item one by one. The parent-child relationship of the nodes on the first tree control in the tree control expresses the parent-child relationship of the corresponding label items. For example, the node expressing the label item "optical efficacy" E.12 has 6 child nodes on the first tree control, and each child node expresses the label items with labels E.121, E.122, E.123, E.124, E.125 and E.12z in turn.

[0091] In this application scenario, the branch nodes belonging to the same root node have the same node color. For example, the branch nodes in "Technical Decomposition" (label T) and its subnodes can all be blue, the branch nodes in "Application Classification" (label A) and its subnodes can all be orange, and the branch nodes in "Efficacy Classification" (label E) and its subnodes can all be green. Correspondingly, the subnodes "Electrical Efficacy", "Optical Efficacy", "Spatial Efficacy" and "Thermal Efficacy" (labels E.11, E.12, E.13 and E.14 respectively) of "Efficacy Classification" (label E) can all be green. This setting can easily prompt users to distinguish the major categories of label classification.

[0092] The sub-nodes of "Optical Efficacy": "High Brightness" (label E.121), "High Uniformity" (label E.122), "High Collimation" (label E.123), "Speckle Suppression" (label E.124), "High Color Purity" (label E.125) and "Others" (label E.12z) are all leaf nodes and can be set to gray to distinguish them from the branch node color, so as to easily prompt the user that the above nodes are leaf nodes.

[0093] In the first tree control, the user can switch between showing / hiding child nodes by clicking on a branch node on the first tree control. Figure 3 As shown in the figure, after the user clicks the node "Optical Effect", the sub-nodes of "Optical Effect" can be expanded or collapsed. When there are many label items, unnecessary nodes can be hidden to facilitate user selection.

[0094] Step S206: receiving an indexing instruction, obtaining a first target node pointed to by the indexing instruction, and obtaining a first target label identifier of a label item corresponding to the first target node.

[0095] In one embodiment, reference Figure 4As shown, a button component is also loaded on the node of the first tree control. One tree control node corresponds to one button component. The triggering operation for the button component is to input an indexing instruction pointing to the tree control node. For example, if the document analyst wants to label the target document with the label "high uniformity", then click the button component behind the node with the label "high uniformity" (i.e. Figure 5 The button “Mark” shown in the figure is pressed, and the indexing instruction pointing to the first target node (labeled “High Uniformity”) is input.

[0096] In this embodiment, a reference record display area is also provided on the browser page. Figure 2 As shown, when the document analyst clicks the button component after the "high uniformity" node, the "high uniformity" label is added to the indexing record display area. In this embodiment, the user can select a button component that is not limited to one tree control node to input the indexing instruction multiple times, thereby indexing multiple label items for the target document. For example, after the user has indexed the "high uniformity" label for the target document, the user can also click the button component after the "high collimation" node to input the indexing instruction again, thereby indexing the "high collimation" label item for the target document, and the "high collimation" label item can be added to the indexing record display area.

[0097] Preferably, a prompt information (tips event) can be added to the label item in the indexing record display area. When the cursor moves to the "high uniformity" label in the indexing record display area, the tips display the label identifier E.122, description information, keywords of the "high uniformity" label, and the label path from the "high uniformity" label to its root label "efficacy classification" through the parent-child relationship:

[0098] “Path: Efficacy classification E-> Physical efficacy E.1-> Optical efficacy E.12-> High uniformity E.122”

[0099] By displaying the indexed label items in the above manner, the user can view the tips information of the label items by moving the cursor to the label items in the indexing record display area to understand the parent label, grandparent label, and even the root label item of the label item, without having to search and view the numerous label items in the first tree control on the right, which is more convenient.

[0100] In one embodiment, when the target tag set is loaded through the first tree control, only the button control can be loaded for the leaf node. In other words, the user is only allowed to index the leaf node for the target document. In this way, the user can be prevented from indexing both the parent node and the child node, which is not conducive to the statistics of the indexing results.

[0101] Step S208: Store the first target tag identifier in an indexing record corresponding to the target document information and the target tag set, wherein the indexing record includes one or more tag identifiers.

[0102] As in the previous application scenario, refer to Figure 2 As shown, a save button is provided on the browser page. When the user clicks the save button, the browser obtains the document identifier of the target document and the tag identifier of the tag item in the index record display area, and then uploads them to the server and stores them in the server's database. If an index record corresponding to the primary key containing the target document identifier already exists, the index record is updated. If not, the corresponding index record is added. In other embodiments, the automatic save function can also be implemented, that is, after the index record display area changes (addition or deletion),

[0103] In an application scenario, as shown in Table 1, an indexing record can be established in the database, and the primary key of the indexing record includes a document identifier: file_id field, and a tag set identifier standard_id field; preferably, in order to distinguish the indexing data of different users for the same document under the same tag set, the primary key of the indexing record also includes a user identifier: user_id field, and the field used to store the indexed tag identifier is the record field. In other embodiments, the primary key of the indexing record may also include other fields for limiting and distinguishing (such as the group identifier in the permission control, etc.), but it must at least include the document identifier and the tag set identifier, which is not limited here.

[0104] As in the previous example, if user wayne completes the indexing of CN201811202475.3, the indexing records are E.122 and E.123, the target document identifier uploaded by the browser: CN201811202475.3, the target tag set identifier: analysis standard A, the user identifier: wayne and the indexed tag identifiers E.121 and E.123 are all stored in the corresponding fields, and the storage records are shown in Table 1:

[0105] Table 1

[0106]

[0107] If the storage record already exists, it will be overwritten. It should be noted that the database structure design described in Table 1 is only used to illustrate the logical process of storing the first target tag identifier in the indexing record corresponding to the target document information and the target tag set, and does not limit the specific database design method.

[0108] In a preferred embodiment, the indexing record includes one or more sub-partitions, and the indexing instruction includes a sub-partition identifier.

[0109] Storing the first target tag identifier in an indexing record corresponding to the target document information and the target tag set includes:

[0110] Obtain a target sub-partition identifier corresponding to the indexing instruction; obtain an indexing record corresponding to the target document information and the target tag set, and store the first target tag identifier in the sub-partition corresponding to the target sub-partition identifier.

[0111] As shown in Table 2, the index record in the database record includes two fields: primary and secondary. The index record display area on the corresponding page also includes two areas. Figure 5 The "primary label" area and "secondary label" area in the record correspond to the primary field and secondary field respectively. The button component on the first tree control also includes two buttons: a "primary" button and a "secondary" button. When the user clicks the "primary" button, the label item of the node of the tree control in front of the button is added to the "primary label" area. When the user clicks the "secondary" button, the label item of the node of the tree control in front of the button is added to the "secondary label" area. After uploading to the server, the server stores one or more label items in the "primary label" area in the primary field of the record, and stores one or more label items in the "secondary label" area in the secondary field of the record.

[0112] Table 2

[0113]

[0114] It should be noted that the database structure design described in Table 2 is only used for the logical method of setting index records in partitions in an instance, and does not limit the specific database design method.

[0115] Setting the indexing record into multiple partitions can facilitate users to conduct multi-dimensional analysis of documents. Taking patent analysis as an example, users can click the "Primary" button to add the label items indexed for the feature analysis of independent claims to the "Primary Label" area, and then store them in the primary field of the indexing record; click the "Secondary" button to add the label items indexed for the feature analysis of dependent claims to the "Secondary Label" area, and then store them in the secondary field of the indexing record.

[0116] refer to Figure 5As shown in Table 2, for the target document CN201811202475.3, the technical effect of the independent claim solution is "high uniformity". Then the user can click the "main" button after the node "high uniformity" E.122 and add "high uniformity" E.122 to the "main label" area. And the dependent claim solution brings an additional technical effect of "high collimation" E.123. Then the user can click the "secondary" button after the node "high collimation" E.123 and add "high collimation" E.123 to the "secondary label" area. After clicking the save button, E.122 and E.123 can be stored in the primary field and secondary field of the indexing record with CN201811202475.3 and analysis standard A as the primary keys respectively.

[0117] In other embodiments, the indexing record can also be set with more than 2 sub-partitions. For example, in a paper analysis application, the indexing record can be set with N (N>2) levels of sub-partitions, and N buttons can be set after the nodes of the tree control, corresponding to the aforementioned N sub-partitions respectively. If the user clicks button i (1 < i < N), the label identifier of the label item of the node of the tree control is added to the field corresponding to sub-partition i for storage.

[0118] In another embodiment, the way to trigger the indexing instruction can also adopt the way of node dragging. Refer to Figure 6 As shown, the nodes on the first tree control can be dragged (drag operation) on the browser page, and a drag-and-drop area (drop area) is also set on the page. When the user drags a node on a tree control to the drag-and-drop area and releases it, the indexing instruction is triggered, and the first target node pointed to by the indexing instruction is the dragged node. Referring to the previous example, this way of triggering the indexing instruction can also be adapted to instances where the indexing record has two or more sub-partitions. As Figure 6 shown, the drag-and-drop area can also be divided into two sub-areas, one area corresponding to the primary field and the other area corresponding to the secondary field. Whichever area the dragged first target node is placed in, the label identifier of the label item of the first target node is stored in that field.

[0119] Since there are many nodes on the first tree control and correspondingly many button components, and each button is small, it is easy to make a wrong click by triggering the indexing instruction through the button components. However, using the way of node dragging to trigger the indexing instruction can avoid this situation and prevent misoperation. It should be noted that the triggering of the indexing instruction by the button components and the dragging of the tree control nodes can exist simultaneously. When loading the tree control, button components can be loaded for the nodes and drag event handling functions can be added at the same time to increase multiple triggering methods for the triggering of the indexing instruction.

[0120] In one embodiment, the user can also edit and modify the target tag set. Specifically, the target tag set can be displayed in the third interface area through the second tree control, and a tag setting instruction for the second tree control can be received; the second target node pointed to by the tag setting instruction can be obtained, and the tag item corresponding to the second target node can be modified according to the tag setting instruction.

[0121] In this embodiment, the target tag set can be displayed through the second tree control on the tag set setting page. Figure 7 As shown, in the application scenario based on the browser B / S architecture, an independent tag set setting page can be provided to modify the tag set. In this page, the stored tag sets are displayed through a drop-down box or a list component. After the user selects the target tag set "Analysis Standard A", the target tag set "Analysis Standard A" is loaded and displayed through the second tree control. The nodes on the second tree control also correspond to the tag items of the target tag set "Analysis Standard A".

[0122] The nodes on the second tree control are also loaded with button components, such as Figure 7 In the example, the button components may include "add", "delete", "combine" and "edit". When the user clicks the button, a label setting instruction is input, and the label setting instruction points to the second target node corresponding to the button, which is the node on the second tree control corresponding to the button. For example, refer to Figure 8 As shown in the figure, when the user clicks the "Edit" button, an editing window will pop up, and the user can edit the tag item in the editing window. In this application scenario, the information of the tag item includes the tag name, tag identifier, description information and keyword set. The user can complete the modification by entering in the corresponding input box.

[0123] In a preferred embodiment, the second tree control also supports node dragging events. The user can drag a node A (regardless of branch nodes or leaf nodes) on the second tree control to another node B (regardless of branch nodes or leaf nodes) and release it, thereby changing node A to have node B as its parent node. The parent label item of the corresponding node A's label item is changed to the label item of node B.

[0124] After the user modifies the nodes on the second tree control, the page can be automatically saved, and the label items corresponding to the modified nodes on the second tree control can be obtained and uploaded to the server for storage. A save button can also be set on the page. After the user modifies the nodes on the second tree control on the page multiple times, the user clicks the save button to obtain the label items corresponding to the modified nodes on the second tree control and upload them to the server for storage. For applications with B / S architecture, it is better to use the save button, which can reduce the number of times the browser accesses the web server and improve efficiency.

[0125] It should be noted that the second interface and the third interface can also be the same page area, and the first tree control and the second tree control can be the same tree control. For example, a tree control edit state switching button can be set on the document analysis page. When the button is clicked to switch to the edit state, a button component for inputting label setting instructions is loaded into the tree control, and the user can edit the label item on the node; when the button is clicked to switch to the indexing state, a button component for inputting indexing instructions is loaded into the tree control, and the user can label the target document by clicking the button.

[0126] That is to say, the embodiments of the present invention do not limit the specific page layout. A first interface area for displaying target document information, a second interface of a first tree control for displaying label items for user selection, and a third interface of a second tree control for displaying label items for user editing are set on one or more display interfaces. These all belong to the information display methods supported by the embodiments of the present invention. The embodiments of the present invention do not limit the above-mentioned first interface area, second interface and third interface to be displayed on the same page, do not limit the above-mentioned second interface and third interface to be the same display area, and do not limit the first tree control and the second tree control to correspond to the same tree control.

[0127] In a preferred embodiment, the modification of the label identifier in the label item may be verified. Specifically, modifying the label item corresponding to the second target node according to the label setting instruction includes:

[0128] Verify the label identifier of the label item corresponding to the second target node, and complete the modification if the parent-child relationship expressed by the label identifier conforms to the parent-child relationship of the second target node in the second tree control;

[0129] When the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node does not conform to the parent-child relationship of the second target node in the second tree control, an error is prompted, or the label identifier is automatically adjusted so that the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node conforms to the parent-child relationship of the second target node in the second tree control.

[0130] In one embodiment, the parent-child relationship expressed by the modified label identifier can be verified by comparing the substring of the preset string bits. For example, the preset string bits can be the first N-1 bits when the label identifier is N bits. If the label identifier of the parent label item is E.1, the label identifier of its child label item is E.1*, and * is a 1-bit character (for example, * is 0-9, any one of az and AZ) and passes the verification. When the label identifier of the label item of the second target node is modified, the label identifier is modified to E.1231, and if the label identifier of the label item of its parent node is E.123, the verification passes. If the label identifier of the label item of its parent node is E.12 or E.124, the verification fails. In other embodiments, the preset string bits can also be the first N-2 bits or N-3 bits, or even Ni bits, when the label identifier is N bits. The larger the distinguishing bit i is, the more character sets that can be provided to the user to set the label identifier, and the more child label items that can be classified by the parent label item.

[0131] In another embodiment, the parent-child relationship expressed by the modified label identifier can be verified by comparing the substring separated by the connector. For example, the connector can be set to "-". When the label identifier of the label item of the second target node is modified, the label identifier is modified to E-123-12-1. The substring separated by the connector is E, 123, 12 and 1. If the label identifier of the label item of its parent node is E-123-12, the verification is passed. If the label identifier of the label item of its parent node is E-121-12 or E-123-13, the verification fails. If the label identifier of the label item of its parent node is E-123-12, then the label identifier of the label item of the second target node can be modified to E-123-12-1 or E-123-12-111 or E-123-12-11d to pass the verification, but the label identifier of the label item of the second target node can be modified to E-123-12- or E-123-12-1- or E-123-12-1-2, then the verification fails. Preferably, the ASCII code of the characters in the label identifier can be limited to 0-9, and only the intervals of az and AZ can pass the verification, and characters such as "-", ".", and " / " are set as connectors.

[0132] In other embodiments, the above two methods can also be combined to verify the parent-child relationship expressed by the modified label identifier. For example, in the previous application scenario, for the nodes on the same second tree control, the parent node E.1 can be verified by comparing the substring of the preset string position. If the first-level child node under the parent node E.12 is selected as the second target node for modification, if the first N-1 bits of the label identifier of the modified second target node are not E.12, the verification fails (for example, E.12s passes the verification, and E.1222 fails the verification). When the user edits the node E.1, the verification can be performed by comparing the substring separated by the connector. After removing the connector ".", if the parent node of E.1 is not E, the verification fails.

[0133] By setting the label identification in the above manner or performing verification or correction when the user customizes the label identification, the label identification of the label item can not only distinguish it from other label items, but also express all label items on the path from the root label item to the label item connected by the parent-child relationship. For example, the aforementioned label item E.121 reflects all label items from the root label item E to the label item E.121: E, E.1, E.12 and E.121, where E is the parent label item of E.1, that is, "physical effect" is the first-level classification of "effect classification"; E.1 is the parent label item of E.12, that is, "optical effect" is the direct sub-classification of "physical effect" and the second-level classification of "effect classification"; E.12 is the parent label item of E.121, that is, "high brightness" is the direct sub-classification of "optical effect" and the third-level classification of "effect classification". This allows document analysts to understand the parent tag item, grandparent tag item, and even the root tag item of a tag item by simply observing the tag identifier of the tag item, thereby understanding the subordinate relationship of the tag item without having to refer to the specific classification of the tag item, which improves convenience. At the same time, for computer programs, the parent tag of a tag item can be obtained through simple string bit operations without having to query the tag set, thereby improving execution efficiency.

[0134] In this embodiment, if the label identifier verification of the label item corresponding to the second target node fails, an error message may be prompted, or the label identifier may be automatically adjusted so that the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node conforms to the parent-child relationship of the second target node in the second tree control.

[0135] For example, in one application scenario, refer to Fig. 9 and Fig.10As shown, the user clicks the "Edit" button of the second target node E.121 to edit it, and the label identifier of its parent node is E.12. If the user changes its label identifier to E.1222 in the "Label Identifier" column, then because its first N-1 digits E.123 do not match the label identifier E.12 of the parent node, the user is prompted to add 1 character based on E.12 to complete the modification.

[0136] In a drag-and-drop editing application scenario, the user drags the second target node E.125 to under the node E.1. Since the first N-1 bits of E.125 are E.12, which does not match E.1, the verification fails. At this time, the last bit 5 of E.125 is taken and spliced ​​with the node E.1 to achieve automatic adjustment, and the label identifier of the second target node is automatically adjusted to E.15. It can be understood that when the preset string bits are the first Ni bits when the label identifier is N bits, the last i bits of the label identifier of the second target node can be obtained and spliced ​​with the label identifier of the parent node of the second target node to perform automatic adjustment. When the label identifier of the label item corresponding to the second target node is modified, a verification is performed, and it is automatically adjusted when the verification fails. The parent-child relationship expressed can be automatically formed to conform to the parent-child relationship of the second target node in the second tree control. The number of user inputs is reduced and the convenience of operation is improved.

[0137] In this embodiment, once the tag set is modified, the historical indexing records will also be affected. For example, if several documents in the historical indexing data have been indexed with the tag E.121, and the user modifies the tag set by moving the tag item E.121 up one classification level to E.14, the tag identifier E.121 is still stored in the historical indexing records and is not synchronously changed to E.14, which results in a change error in the analysis data and inaccurate analysis.

[0138] To solve this problem, in one embodiment, modifying the label item corresponding to the second target node according to the label setting instruction also includes: when the label identifier of the label item of the second target node is modified, recording the modification record of the label item in the modification order.

[0139] The document analysis method can also obtain a target indexing record set corresponding to the target tag set, the target indexing record set including one or more indexing records; and modify the tag identifier in the indexing record in the target indexing record set according to the modification record.

[0140] For example, in an application scenario, still refer to Figure 7As shown, the second target node E.121 under analysis standard A has "Edit" and "Delete" buttons. If you click the "Edit" button, in the pop-up editing window, after changing the label of the second target node E.121 to E.129, a modification record is generated:

[0141] change E.121 to E.129, the identifier before the change is E.121, and the identifier after the change is E.129;

[0142] When the user clicks Save or Auto-Save and sends the modified record to the server, if in this embodiment, the data items stored in the indexed index record in the database are shown in Table 3:

[0143] Table 3

[0144]

[0145] Then, E.121 in the "Indexed Label" field in indexing records 1 and 2 can be replaced with E.129. The Indexed Label field of indexing record 1 is changed to E.129, E.123; the Indexed Label field of indexing record 2 is changed to E.129.

[0146] If the "Delete" button is clicked, the second target node E.121 is removed from the second tree control and a modification record is generated:

[0147] change E.121 to null (or express it as remove E.121), the identifier before the change is E.121, and the identifier after the change is null;

[0148] When the user clicks Save or Auto-Save and sends the modified record to the server, referring to Table 3, E.121 in the "Indexed Label" field in indexing records 1 and 2 can be replaced with empty or deleted; that is, the Indexed Label field of indexing record 1 is changed to E.123; the Indexed Label field of indexing record 2 is changed to empty.

[0149] If the user chooses to merge node E.121 into node E.122, the second target node E.121 is removed from the second tree control and a modification record is generated:

[0150] change E.121 to E.122 (or merge E.121 to E.122), the identifier before the change is E.121, and the identifier after the change is 122;

[0151] When the user clicks Save or Auto-Save and sends the modified record to the server, as shown in Table 1, E.121 in the "Indexed Label" field in indexing record 1-2 can be changed to E.122; that is, the Indexed Label field of indexing record 1 is changed to E.122, E.123; and the Indexed Label field of indexing record 2 is changed to E.122.

[0152] In this embodiment, in order to improve execution efficiency, the user can modify multiple second target nodes and generate multiple modification records, and then click the Save button to upload the modified target tag set expressed by the second tree control and the corresponding modification records to the server. Then, the tag identifier in the indexing record in the target indexing record set is modified according to the modification record.

[0153] Specifically, modifying the label identifier in the indexing record in the target indexing record set according to the modification record may include:

[0154] Traverse the indexing records in the target indexing record set; for the traversed indexing records, sequentially obtain the modification records one by one, obtain the pre-modification identifier and post-modification identifier contained in the modification record, and modify the pre-modification identifier contained in the traversed indexing record to the post-modification identifier.

[0155] In an application scenario, if the user performs the following operations in sequence:

[0156] 1. Click the "Edit" button of node E.121 to change the label from E.121 to E.128

[0157] 2. Click the "Delete" button of node E.123 to delete it

[0158] 3. Drag node E.128 to node E.1 and release it. The label is automatically adjusted to E.18.

[0159] 4. Click the "Save" button

[0160] The above operations generate three modification records in the modification order and upload them to the server:

[0161] r1: change E.121 to E.128

[0162] r2: remove E.123

[0163] r3: change E.128 to E.18

[0164] After receiving the above modification records with order, the server takes the 4 indexing records related to "Analysis Standard A" in Table 3 as an example, and then traverses the above 4 indexing records. For the record with sequence number 1, the indexed label stored is "E.121, E.123", then first replace E.121 with E.128 according to the modification record r1, and the indexed label field data is changed to: "E.128, E.123"; then delete E.122 according to the modification record r2, and the indexed label field data is changed to: "E.128"; then replace E.128 with E.18 according to the modification record r3, and the indexed label field data is changed to: "E.18", and the above 4 indexing records are traversed, and the indexing records after the final change are shown in Table 4:

[0165] Table 4

[0166]

[0167] From this, it can be seen that as the user modifies the label identifier of the label item on the second tree control, the label identifier in the indexing record that has indexed the label item is also changed accordingly. This allows the user to modify the label set at will during the indexing process without affecting the accuracy of the indexing results. There is no need to repeatedly use the modified label items for re-indexing after modifying the label set, which greatly improves the accuracy and efficiency of document analysis.

[0168] In a preferred embodiment, before modifying the label identifier in the indexing record in the target indexing record set according to the modification record, two or more modification records for the same second target node may be merged.

[0169] As in the above example, modification records r1 and r3 are both modifications to the same second target node, that is, the label identifier of the second target node is first changed from E.121 to E.128 through the editor, and then the label identifier of the second target node is changed from E.128 to E.18 through dragging. The modification or change to the same second target node has the following characteristics: that is, in the modification record of the sequential record, there is a pre-change identifier of the later modification record that is the post-change identifier of the previous modification record. Refer to the above example again:

[0170] r1: change E.121 to E.128 is the last modification record, and the mark after the modification is E.128;

[0171] r3: change E.128 to E.18 is the last modification record, and the mark after the change is E.18;

[0172] The way to merge the two is to replace the modified identifier of the previous modification record with the modified identifier of the later modification record, that is, the modification record after r1 and r3 are merged is:

[0173] r1: change E.121 to E.18

[0174] The above merging method can be implemented by traversing and searching for modification records and comparing them pairwise, etc., which is not limited here. The modification records after merging are:

[0175] r1: change E.121 to E.18

[0176] r2: remove E.123

[0177] It can be seen from this that for indexing records, it is only necessary to change the label identification twice in sequence by referring to the modification record to complete the modification of the indexing record. Compared with the method of directly modifying the indexing record according to the modification record without merging (each indexing record needs to be changed three times), the number of times the label identification is changed is greatly reduced, thereby greatly improving the execution efficiency.

[0178] In a preferred embodiment, before modifying the label identifiers in the indexing records in the target indexing record set according to the modification record, the target label set and the indexing records corresponding to the target label set may be backed up.

[0179] As in the above example, the user completes the modification of "Analysis Standard A" loaded in the second tree control on the tag set setting page. After clicking Save, the modified tag set "Analysis Standard A" on the second tree control is uploaded to the server. The server can first back up "Analysis Standard A" and back up the indexing records related to "Analysis Standard A" in the indexing records. For example, a new tag set "Analysis Standard A-backup" can be created, and the tag item information of "Analysis Standard A" can be backed up to "Analysis Standard A-backup". Then, the indexing records with the tag set field of "Analysis Standard A" in the backed-up indexing records can be backed up, and the tag set field in the backed-up indexing records can be modified to "Analysis Standard A-backup" to complete the backup.

[0180] Since the modification of the label set will cause the indexing record to change synchronously, once the modification of the label set is wrong, the indexing record will also be wrong, thus affecting the analysis results. In this case, before making substantial modifications to the label set, back it up and back up the corresponding indexing record synchronously. Even if there is a problem with the modification, the historical label set and indexing record still exist in the backup and can be restored at any time, thus avoiding the risk of analysis data errors caused by the modification of the label set and improving the reliability of the system.

[0181] In one embodiment, after the user has indexed multiple documents and generated multiple indexing records, the user can also perform statistical screening on the indexing records according to the indexed tag identifiers to generate statistical tables and statistical charts. Specifically, the user can obtain an input screening expression, which includes one or more tag identifiers in the target tag set; obtain an input document information set, obtain indexing records corresponding to the document information in the document information set and the target tag set; traverse the indexing records, and screen the indexing records whose tag identifiers meet the screening expression.

[0182] As in the aforementioned application scenario, if the user selects the label set as "Analysis Standard A" on the document analysis page, and uses the label items in "Analysis Standard A" to index several patent documents in Patent Package 1, generating several indexing records, these indexing records can be statistically filtered through multiple dimensions, and the statistical and filtering methods can be defined by the input filtering expression.

[0183] For example, in a patent analysis application scenario, the label items ""T.1, Technical Solution 2"T.2, Technical Solution 3"T.3 and Technical Solution 4"T.4 in "Analysis Standard B" respectively represent four technical solutions in a certain field; the label items "Efficacy 1"E.1, "Efficacy 2"E.2, "Efficacy 3"E.3 and "Efficacy 4"E.4 in "Analysis Standard B" respectively represent four efficacy categories. If the user wants to view patent documents that contain at least Technical Solution 1 and Technical Solution 2 but not Technical Solution 3, and at least Efficacy 1, the label identifiers spliced ​​by logical connectors can be added to the filter expression:

[0184] "T.1 and T.2 and E.1 not T.3"

[0185] Then upload it to the server. The server searches for documents whose indexing records contain at least T.1, T.2 and E.1, but not T.3, among the indexing records that meet "Analysis Standard B" based on the filter expression, and obtains a list of patents that meet the above filter expression.

[0186] For another example, in the above patent analysis application scenario, if the user wants to count the technical development trends of the above four technical solutions in a certain year range, the user can add the time parameters of the start year and the end year, and the label identification parameters of T.1, T.2, T.3, and T.4 in the filter expression and upload it to the server. The server counts the number of indexing records containing each label identification of T.1, T.2, T.3, and T.4, and then counts the distribution of this number in the time dimension of the application year of the patent document corresponding to the indexing record, which can be displayed on the indexing record statistics page as follows: Fig.11The X-axis is time, the Y-axis is quantity, and the legend is a statistical chart of the technical solutions T.1, T.2, T.3, and T.4 reflecting the technical routes of T.1, T.2, T.3, and T.4.

[0187] For another example, in the application scenario of the above patent analysis, if the user hopes to obtain the above four technical solutions and the efficacy matrix diagram of the four efficacy classifications, the user can add the statistical dimension 1 of the label identifiers T.1, T.2, T.3, and T.4 to the screening expression, and add the statistical dimension 2 of the label identifiers E.121, E.122, E.123, and E124, and upload them to the server. The server pairs the label identifiers in the statistical dimension 1 with the label identifiers in the statistical dimension 2 to form sub-expressions such as "T.1 and E.121", "T.1 and E.122"... "T.4 and E.124", and then screens and counts the number of indexing records that meet each of the above sub-expressions in the indexing records, which can be displayed on the indexing record statistics page as shown below. Fig.12 The efficacy matrix is ​​shown.

[0188] It should be noted that the filtering expressions in the above description are only used to illustrate the filtering logic of filtering indexing records and then filtering documents by filtering indexing records, and do not limit the form of filtering expressions. Depending on the storage form of the indexing records (files, streaming data, databases), the expression forms used in different application scenarios may also be different, which are not limited here. And the filtering logic expressed by the filtering expression is not limited to the description of the above application scenarios. According to the user's filtering and statistical needs, as long as the filtering method is based on tag identification, it is a filtering expression supported by the embodiments of the present invention.

[0189] The above document analysis method can also be applied to the analysis of papers, etc. For example, in an application scenario of biological paper analysis, the user can pre-build a label set "biological classification" on the second tree control of the label set setting page, and set each label item that conforms to the biological classification according to the subordinate relationship of "kingdom, phylum, class, order, family, genus, and species" of biological classification, and reasonably set the label identifier of the label item so that the label identifier of the label item expresses the subordinate relationship of "kingdom, phylum, class, order, family, genus, and species". In this application scenario, the label identifier of the label item can be set by using 2 characters to express one level of information for one label item.

[0190] For example, the label identifier K1 can be defined as the label item "Animal Kingdom" (which can be set as the label name of the label item, and the following is similar), K1P1 is the label identifier of the label item "Chordata" (including the label identifier of the parent label item "Animal Kingdom" K1), K1P104 is the label identifier of the label item "Mammalia", K1P10409 is the label identifier of the label item "Carnivora", K1P1040918 is the label identifier of the label item "Felidae", K1P104091807 is the label identifier of the label item "Panthera", K1P10409180703 is the label identifier of the label item "Tiger species", K1P1040915 is the label identifier of the label item "Ursidae", K1P104091509 is the label identifier of the label item "Genus Ailuropoda", and K1P10409150901 is the label identifier of the label item "Genus Ailuropoda". From the settings of the above label identifications, it can be seen that the settings of K1P10409, K1P1040918 and K1P1040915 express the information that "Felidae" and "Ursidae" both belong to "Carnivora", but branches begin to appear under the sub-classification of "Carnivora".

[0191] After the label items are set, users can analyze and index the paper on the document analysis page. If the subject of the analysis is the biological distribution of a certain gene segment, and a biological paper proves that both tigers and giant pandas have this gene segment, the paper can be indexed with the label items K1P10409180703 and K1P10409150901 at the same time. After the papers are batched and analyzed as above, the indexed label identifiers can be counted. If the maximum value of the common prefix of the indexed label identifiers is "K1P10409", it means that the gene segment is a gene unique to the "Carnivora" and may be related to the digestion and decomposition of meat. If the maximum value of the common prefix of the indexed label identifiers is "K1P104", it means that the gene segment is a gene unique to the "Mammalia" and may be related to viviparity.

[0192] It can be seen that after using the above-mentioned document analysis system to analyze the paper, statistical analysis results can be generated through the statistics of label identification, and the above-mentioned document analysis system is easy to operate. After indexing, it can automatically filter and count according to label identification, which greatly improves the efficiency of analysis and statistics.

[0193] The above document analysis method can also be applied to the analysis of legal documents. For example, in an application scenario of a case analysis system, users can construct label items in advance according to dimensions such as the subject of behavior S, the object of behavior O, and the case facts E. For example, S.1 is "natural person", S.2 is "legal person", O.1 is "property", E.1 is "peaceful means", E.2 is "violent means", E.3 is "transfer of possession", etc. When the user analyzes a case document, if the label is S.1, O.1, E.1 and E.3, it can be analyzed that the suspect in the case is suspected of theft. If the label is S.1, O.1, E.2 and E.3, it can be analyzed that the suspect in the case is suspected of robbery.

[0194] It can be seen from this that the above-mentioned document analysis system can be used to analyze legal documents, and can also be used to complete the analysis of a single document. By statistically analyzing the label identifiers on a single document, the specific content involved in the document (such as the specific crime) can be analyzed. The above-mentioned document analysis system is easy to operate and uses label identifiers that can express parent-child relationships, allowing users to build a more systematic label set. As long as the user pre-segmented and set label items according to a certain classification logic, the document can be accurately indexed and analyzed through indexing, thereby improving the accuracy of document analysis.

[0195] It should be noted that the above application scenario is an application scenario of a document analysis system with a B / S architecture, and the document analysis method is not limited to programs with a B / S architecture, but can also be based on programs with a C / S architecture or independent desktop programs and mobile apps. In the application scenario where the C / S architecture program implements the document analysis method, the functions implemented by the client program are similar to those implemented by the browser in the B / S architecture. The client program is responsible for interface presentation, user interaction and partial data processing, while the storage of tag sets, indexing records, and data query management and other services can be carried by the server. In the application scenario where the independent desktop program implements the document analysis method, the independent desktop program executes the above document analysis method in its entirety. In the application scenario where the mobile terminal APP implements the above document analysis method, the functions implemented by the mobile terminal APP are similar to those implemented by the browser in the B / S architecture. The mobile terminal APP is responsible for interface presentation, user interaction and partial data processing, while the storage of tag sets, indexing records, and data query management and other services can be carried by the server of the APP. In other words, the implementation of the above document analysis method can be implemented based on a variety of program languages ​​and a variety of program architectures, and the computer program itself is not limited in the embodiments of the present invention.

[0196] To solve the above problems, the present invention also proposes a document analysis device, such as Fig.13As shown, it includes: a document display module 102, a tag set display module 104, a tag indexing module 106 and an indexing record storage module 108, wherein:

[0197] The document display module 102 is used to display target document information in the first interface area.

[0198] The label set display module 104 is used to display the target label set in the second interface area through the first tree control. The target label set includes label items with parent-child relationships. The label items have label identifiers, and the label identifiers express the parent-child relationships of the corresponding label items. The label items are displayed by the nodes of the first tree control, and the parent-child relationships of the label items are consistent with the parent-child relationships of the corresponding nodes on the first tree control.

[0199] The label indexing module 106 is used to receive an indexing instruction, obtain a first target node pointed to by the indexing instruction, and obtain a first target label identifier of a label item corresponding to the first target node.

[0200] The indexing record storage module 108 is used to store the first target tag identifier in an indexing record corresponding to the target document information and the target tag set, wherein the indexing record includes one or more tag identifiers.

[0201] In one embodiment, the indexing record includes one or more sub-partitions, and the indexing instruction includes a sub-partition identifier.

[0202] The indexing record storage module 108 is also used to obtain the target sub-partition identifier corresponding to the indexing instruction; obtain the indexing record corresponding to the target document information and the target tag set, and store the first target tag identifier in the sub-partition corresponding to the target sub-partition identifier.

[0203] In one embodiment, Fig.13 As shown, the device also includes a label set setting module 110, which is used to display the target label set through the second tree control in the third interface area, receive the label setting instruction for the second tree control; obtain the second target node pointed to by the label setting instruction, and modify the label item corresponding to the second target node according to the label setting instruction.

[0204] In one embodiment, Fig.13 As shown, the device also includes an indexing record change module 112, which is used to record the modification record of the label item in the modification order when the label identifier of the label item of the second target node is modified; obtain a target indexing record set corresponding to the target label set, the target indexing record set including one or more indexing records; and modify the label identifier in the indexing record in the target indexing record set according to the modification record.

[0205] In one embodiment, the second target node may be one or more, the modification record may include one or more, and the modification record may include a pre-modification identifier and a post-modification identifier. The indexing record changing module 112 is also used to traverse the indexing records in the target indexing record set; for the traversed indexing records, sequentially obtain modification records one by one, obtain the pre-modification identifier and the post-modification identifier contained in the modification record, and modify the pre-modification identifier contained in the traversed indexing record to the post-modification identifier.

[0206] In one embodiment, Fig.13 As shown, the device further includes a modification record merging module 114, which is used to merge two or more modification records for the same second target node.

[0207] In one embodiment, Fig.13 As shown, the device further includes an indexing backup module 116 for backing up the target tag set and the indexing records corresponding to the target tag set.

[0208] In one embodiment, Fig.13 As shown, the device also includes a label item verification module 118, which is used to prompt an error when the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node does not conform to the parent-child relationship of the second target node in the second tree control, or automatically adjust the label identifier so that the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node conforms to the parent-child relationship of the second target node in the second tree control.

[0209] In one embodiment, Fig.13 As shown, the device also includes an indexing statistics module 120, which is used to obtain an input filtering expression, the filtering expression includes one or more tag identifiers in the target tag set; obtain an input document information set, obtain indexing records corresponding to the documents in the document information set and the target tag set; and filter out documents corresponding to the indexing records whose contained tag identifiers meet the filtering expression.

[0210] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0211] In one embodiment, Fig.14 As shown, Fig.14A terminal of a computer system for running the document analysis method is shown. The computer system may be a terminal device such as a smart phone, a tablet computer, a PDA, a laptop computer or a personal computer. Specifically, it may include an external input interface 1001, a processor 1002, a memory 1003 and an output interface 1004 connected via a system bus. Among them, the external input interface 1001 may optionally include at least a network interface 10012. The memory 1003 may include an external memory 10032 (such as a hard disk, a CD or a floppy disk, etc.) and an internal memory 10034. The output interface 1004 may include at least a display screen 10042 and other devices.

[0212] In this embodiment, the operation of this method is based on a computer program, the program file of which is stored in the external memory 10032 of the aforementioned computer system, loaded into the internal memory 10034 during operation, and then compiled into machine code and transferred to the processor 1002 for execution, so that a logical document display module 102, a label set display module 104, a label indexing module 106, and an indexing record storage module 108 are formed in the computer system based on the von Neumann system. In addition, during the execution of the shape setting method of the image display area, the input parameters are received through the external input interface 1001, transferred to the memory 1003 for buffering, and then input to the processor 1002 for processing, and the result data of the processing is either buffered in the memory 1003 for subsequent processing, or transferred to the output interface 1004 for output.

[0213] The document analysis method, device, computer storage medium and equipment provide a convenient operation mode for document tag indexing analysis. When tagging a document, a tag set is displayed through a tree control. The nodes on the tree control correspond to tag items. The tag items have a parent-child relationship (subordinate relationship) when designed, and the parent-child relationship is expressed through tag identifiers. When the tag items are mapped to the tree control, they are also consistent with the parent-child relationship of the corresponding nodes on the tree control. The user can complete the indexing of the tag item corresponding to the node by simply operating the node on the tree control, and the operation traversal is fast.

[0214] At the same time, since the label identifier of the label item added to a document contains the parent-child relationship information of the label item (the subordinate relationship of the analysis feature classification), the user can understand the hierarchy and subordinate relationship of the label item in the entire label classification through the label identifier of the label item. Therefore, the overall classification information is expressed through only one label identifier, and the amount of information expressed is larger. The user only needs to reasonably design the label classification to obtain the content summary expressed by the target document through the feature information such as the name, description and keywords of the label item indexed for the target document, thereby realizing the analysis of the document. Compared with the method of actively reading the document manually and then writing an analysis summary, the embodiment of the present invention is a reverse analysis method, which is to pre-standardize the analysis features, and then index the features in the pre-defined specifications; therefore, the document analysis process is simpler and the analysis accuracy is more objective and accurate.

[0215] At the same time, since a tag identifier containing a parent-child relationship is set for the tag item, the statistics and screening work is simpler and more efficient. Only a simple string match is needed to filter and count the indexed records accordingly. Compared with the Excel table statistics method in traditional technology, the statistical efficiency is significantly higher and more accurate. After indexing, only a filter expression needs to be entered, and the computer can match and search according to the specific string of the tag identifier in the filter expression, so as to quickly filter out the document list or count the number of documents, making the operation simpler and the execution efficiency higher. In addition, due to the design of the tag identifier, the screening and statistics of the indexed records can be based on a flexible screening expression. The screening expression only needs to contain a specific tag identifier and use a corresponding logical expression to constrain it, so as to complete a variety of document screening and statistical work, making the document analysis work more flexible.

[0216] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A document analysis method, characterized in that: The method comprises: Display target document information in the first interface area; The target label set is displayed in the second interface area through the first tree control, the target label set includes label items with a parent-child relationship, the label items have label identifiers in the form of strings, the label identifiers express the parent-child relationship of the corresponding label items, the label items are displayed by the nodes of the first tree control, and the parent-child relationship of the label items is consistent with the parent-child relationship of the corresponding nodes on the first tree control; the label identifiers in the form of strings of the label items express the path from the root node of the first tree control to the node corresponding to the label item; Receive an indexing instruction, obtain a first target node pointed to by the indexing instruction, and obtain a first target label identifier of a label item corresponding to the first target node; storing the first target tag identifier in an indexing record corresponding to the target document information and the target tag set, wherein the indexing record includes one or more tag identifiers; The method further comprises: Displaying the target tag set in the third interface area through a second tree control, and receiving a tag setting instruction for the second tree control; Acquire a second target node pointed to by the label setting instruction, and modify a label item corresponding to the second target node according to the label setting instruction; When the label identifier of the label item of the second target node is modified, recording the modification record of the label item of the second target node in the modification order; Merging two or more modification records for the same second target node; Acquire a target indexing record set corresponding to the target tag set, wherein the target indexing record set includes one or more indexing records; Traversing the indexing records in the target indexing record set; For the traversed indexing records, the modification records are obtained one by one in sequence, the pre-modification identifier and the post-modification identifier contained in the modification record are obtained, and the pre-modification identifier contained in the traversed indexing record is modified to the post-modification identifier.

2. The document analysis method according to claim 1, characterized in that: The indexing record includes one or more sub-partitions, and the indexing instruction includes a sub-partition identifier; The storing the first target tag identifier into the indexing record corresponding to the target document information and the target tag set includes: Get the target sub-partition identifier corresponding to the indexing instruction; An indexing record corresponding to the target document information and the target tag set is obtained, and the first target tag identifier is stored in a sub-partition corresponding to the target sub-partition identifier.

3. The document analysis method according to claim 1, characterized in that: Before traversing the indexing records in the target indexing record set, the method further includes: The target tag set and the indexing record corresponding to the target tag set are backed up.

4. The document analysis method according to claim 1, characterized in that: Modifying the label item corresponding to the second target node according to the label setting instruction includes: When the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node does not conform to the parent-child relationship of the second target node in the second tree control, an error prompt is given, or the label identifier is automatically adjusted so that the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node conforms to the parent-child relationship of the second target node in the second tree control.

5. The document analysis method according to any one of claims 1 to 4, characterized in that: The method further comprises: Obtaining an input screening expression, wherein the screening expression includes one or more tag identifiers in the target tag set; Obtain an input document information set, and obtain indexing records corresponding to documents in the document information set and a target tag set; The documents corresponding to the indexing records whose tag identifiers meet the filter expression are filtered out.

6. A document analysis device, characterized in that: The device comprises: A document display module, used to display target document information in the first interface area; A label set display module, used for displaying a target label set in a second interface area through a first tree control, wherein the target label set includes label items with a parent-child relationship, the label items have label identifiers in the form of strings, the label identifiers express the parent-child relationship of the corresponding label items, the label items are displayed by nodes of the first tree control, and the parent-child relationship of the label items is consistent with the parent-child relationship of the corresponding nodes on the first tree control; the label identifiers in the form of strings of the label items express the path from the root node of the first tree control to the node corresponding to the label item; A label indexing module, used to receive an indexing instruction, obtain a first target node pointed to by the indexing instruction, and obtain a first target label identifier of a label item corresponding to the first target node; An indexing record storage module, used to store the first target tag identifier in an indexing record corresponding to the target document information and the target tag set, wherein the indexing record includes one or more tag identifiers; The device also includes a label set setting module, which is used to display the target label set through a second tree control in the third interface area, receive a label setting instruction for the second tree control; obtain a second target node pointed to by the label setting instruction, and modify a label item corresponding to the second target node according to the label setting instruction; The device further comprises a modification record merging module, which is used to merge two or more modification records for the same second target node; The device also includes an indexing record changing module, which is used to record the modification record of the label item in the modification order when the label identifier of the label item of the second target node is modified; obtain a target indexing record set corresponding to the target label set, the target indexing record set including one or more indexing records; modify the label identifier in the indexing record in the target indexing record set according to the modification record; The second target node is one or more than one, the modification record includes one or more than one, and the modification record includes a pre-modification identifier and a post-modification identifier; The indexing record change module is also used to traverse the indexing records in the target indexing record set; for the traversed indexing records, the modification records are obtained one by one in sequence, the pre-modification identifier and the post-modification identifier contained in the modification record are obtained, and the pre-modification identifier contained in the traversed indexing record is modified to the post-modification identifier.

7. The document analysis device according to claim 6, characterized in that: The indexing record includes one or more sub-partitions, and the indexing instruction includes a sub-partition identifier; The indexing record storage module is used to obtain the target sub-partition identifier corresponding to the indexing instruction; obtain the indexing record corresponding to the target document information and the target tag set, and store the first target tag identifier in the sub-partition corresponding to the target sub-partition identifier.

8. The document analysis device according to claim 6, characterized in that: The device further comprises an indexing backup module, which is used to back up the target tag set and the indexing records corresponding to the target tag set.

9. The document analysis device according to claim 6, characterized in that: The device also includes a label item verification module, which is used to prompt an error when the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node does not conform to the parent-child relationship of the second target node in the second tree control, or automatically adjust the label identifier so that the parent-child relationship expressed by the label identifier of the label item corresponding to the second target node conforms to the parent-child relationship of the second target node in the second tree control.

10. The document analysis device according to any one of claims 6 to 9, characterized in that: The device also includes an indexing statistics module, which is used to obtain an input filtering expression, wherein the filtering expression includes one or more tag identifiers in the target tag set; obtain an input document information set, obtain indexing records corresponding to the documents in the document information set and the target tag set; and filter out documents corresponding to the indexing records whose contained tag identifiers meet the filtering expression.

11. A computer storage medium, characterized in that: The computer storage medium stores one or more computer instructions, and the one or more instructions are suitable for being loaded by a processor and executing the document analysis method according to any one of claims 1 to 5.

12. A document analysis device, characterized in that: The document analysis device includes a memory and a processor, wherein the memory stores a group of programs, and the processor is used to call the programs stored in the memory to execute the document analysis method according to any one of claims 1 to 5.

Images