RU2729224C2 - Information retrieval method and corporate information retrieval system - Google Patents

Abstract

FIELD: computer equipment.SUBSTANCE: invention relates to computer engineering. Disclosed is an information retrieval method implemented in a corporate information retrieval system configured to extract results from a plurality of different sources, comprising steps of receiving a request at an input; at the output, this request is issued in a set of different sources, wherein it is characterized by a hierarchical multi-tree list view of search results in form of one or more sets of related graphs, which nodes do not contain cycles, but can have duplicates and intersections with other sets or branches, wherein there can be a non-unique path, characterized by a sequence of nodes, from root to node, tree roots have a list representation and can be opened if necessary or interacting with the user interactively; includes execution of search requests; performing search query analysis by means of morphological analysis; accessing a search index database through a corporate information retrieval system and extracting search results therefrom; generating search results using a corporate information search system based on the right of the user making the request, and rights associated with the source information; filling a search index base; using information on document structure; aggregation of obtained information from files into documents, wherein files are considered as version of document; merging, via a corporate information search system, versions of files into a document for constructing a common index; finding, through a corporate information retrieval system, documents with similar structural data from meta-information for subsequent output of found information to a user in the presence of necessary rights; combining these documents through the corporate information retrieval system into higher-level structures, such as "project", "direction"; display when searching results in the form of "project", "direction".EFFECT: technical result is possibility of searching information with provision of information aggregation, information presentation form and compliance with information security requirements when providing access to documents.2 cl, 1 dwg

Description

The invention is intended for large and medium-sized enterprises, with a large volume of documents in the information system (IS) of the enterprise and a large number of employees.

The corporate information retrieval system (KSPI) acquires particular relevance in companies with an extensive organizational structure in different regions of the country or the world.

Enterprise search is a feature that should quickly and easily lead a user to a small number of results from millions of documents. This task is very different from searching the Internet over a huge amount of information.

Known Information Management System (patent GB2338324, IPC G06F17 / 30, op. 12/15/1999), used when searching for data from a plurality of data blocks, such as a collection of scientific articles, provides a display of a group of data units in virtual space and each data block indicates its correlation with search criteria. Links are provided between the various blocks of data to help navigate the data during searches. In addition, an indication of a previous or current search can be given by another search device, which will display the search path for users through a group of data units.

The disadvantage of the known technical solution is that only document links are built without determining their versioning.

Known Computerized method (patent US6738678, IPC G06F17 / 30, op. 18.05.2004), which determines the ranking of documents, including information content. This method uses content and connectivity analysis. The input set of documents is represented as a neighborhood graph in memory. In the graph, each node represents one document, and each directional front that connects a pair of nodes represents a link between the two documents. The input set of documents presented in the graph is evaluated according to the content of the documents. A subset of documents is selected from the input dataset if the content ranking of the selected documents exceeds the first specified threshold. Nodes representing any documents other than the selected documents are removed from the schedule. The selected subset of documents is evaluated according to the document anchor, and the output document set that exceeds a second predetermined threshold is selected for presentation to users.

The closest technical solution adopted as a prototype is the METHOD, SYSTEM AND COMPUTER SOFTWARE FOR SEARCHING, NAVIGATION AND RANKING OF DOCUMENTS IN THE PERSONAL NETWORK (RF Patent 2388050, IPC G06F17 / 30, op. 27.04.2010). The method determines an undirected weighted relationship between two of said plurality of documents based on similarity; determine directional weighted relationship between at least two of the above-mentioned plurality of documents; add said certain non-directional weighted links to said certain directional weighted links to create a hybrid network having links, and perform a link analysis algorithm receiving links of said hybrid network as its input, said algorithm including at least link analysis forward and backward link analysis. The output of this algorithm is a set of estimates in the analysis of links.

The disadvantage is that the known technical solution is universal and concerns only the construction of links between documents, but there are no links between versions of documents.

The objective of the invention is to develop a method for aggregating information, presentation forms and compliance with information security (IS) requirements when providing access to documents.

The technical result from the application of the invention is to provide the ability to search for information with the provision of information aggregation, the form of information presentation and compliance with information security (IS) requirements when providing access to documents.

This is achieved due to the fact that the information retrieval method implemented

in a corporate information retrieval system, capable of extracting results from a set of various sources, containing stages at which a request is received at the input; at the output, this query is issued in a set of different sources , according to the invention, it is characterized by a hierarchical multi-tree list representation of search results in the form of one or many sets of connected graphs, the nodes of which do not contain cycles, but may have duplicates and intersections with other sets or branches, when This can have a non-unique path, characterized by a sequence of nodes, from root to node, the roots of trees have a list view, and can be expanded when necessary or when interacting with the user interactively. The method includes executing search queries through a web interface, a mobile application, or in another system through an API providing services, as the main user interface, a search string can be used; further parsing the search query using morphological analysis, using a database or other storage of similar words or their components, to expand the search capabilities; using a corporate information retrieval system; accessing the search index base through the corporate information search system and extracting search results from it in the form of a link, document, project, file, object, contact information to interact with the information owner; formation of search results using the corporate information retrieval system, based on the authority of the user making the request and the authority associated with the initial information; filling the search index base through the interaction of the corporate information search system with the company's information system and receiving information from it to create a search index, on the basis of which the search is carried out; using information about the structure of the document to improve the quality of indexing, search and display of the document; aggregation of the information received from files into documents, while the files are considered as versions of the document; combining file versions into a document through the corporate search system to build a common index; finding documents with similar structural data from meta-information through the corporate information search system for the subsequent display of the found information to the user if he has the necessary powers; combining these documents according to these characteristics through the corporate information search system into higher-level structures, such as "project", "direction"; displaying the search results in the form of a "project", "direction", in which the name of the "project", creators, owners of the "project", "directions" and other information available to the user, according to his authority, can be indicated.

This is also achieved due to the fact that the corporate information retrieval system (KSPI) contains the main structures in the form of a data source agent (AID), a database management system (DBMS), an information aggregation system (AIS) , a user interface system (SIP) ; AID data source agent is intended for delivery of documents, files, extraction of keywords, ranking of keywords, indexing, reading properties, powers or other meta - information, as well as for using the technology of converting a graphic image of text into computer text using a graphic recognition algorithm ( OCR), the formation of index information that enters the database management system of the DBMS, the agent of the AID data source contains a set of filters that provide parsing of various file formats, and meta-information comes from both the agent and the filter from which it is extracted, the formation of a search an index to be stored in the DBMS database management system, and the AID data source agent can be implemented on several functional information systems (IS); database management system DBMS is designed to store: morphological word forms to expand search capabilities, staff or organizational structure of the enterprise obtained from the information systems of the enterprise's IS, authority, search index; AIS information aggregation system is designed to concentrate individual streams of information into a single structure, with the ability to generate index information to combine files into documents, documents into projects; the SIP user interface system is designed to implement interaction with the user through various software and hardware; wherein the output of the AID data source agent is connected to the input of the DBMS database management system, and the DBMS database management system is connected to the AIS information aggregation system and to the VIS user interface system; AID data source agent interacts with the information system, file servers, document management system, intranet and Internet sites, receives information and powers from there, AID data source agent completely reads the file itself, selects from the system all information about powers, file name, file name , tags, flags, that is, the external attributes of the file and sends along with all the meta - information, authority, date of creation, the creator, which he took from the IS information system, to the DBMS database management system; the AID data source agent is designed to process information, contains a recognition module (OCR), generates index information that can be placed in the database management system of the DBMS, contains a set of filters that provide parsing of various file formats, and meta-information comes from the agent and from the filter that extracts it; meta - information, as well as information for creating a search index, is sent to the DBMS database management system; database management system DBMS is a component of the corporate information retrieval system that stores indexes, when a request is received through the SIP user interface system, the DBMS database management system is accessed and the necessary information is obtained from the DBMS database management system, while the selection is based on information stored in the DBMS database management system using morphology, parsing the request; the AID data source agent is configured to compress and send the received information to the server where the DBMS database management system is located, where the information is deployed and embedded in the DBMS database management system, since the communication channels may be weak; the SIP user interface system generates requests to the DBMS database management system to retrieve information, taking into account the morphology, parsing words, performs the requests and clarifications themselves, the SIP user interface system carries out a two-way exchange of information with the DBMS database management system.

New essential features of this invention:

1. Hierarchical multi-tree list representation of search results, namely in the form of one or many sets of connected graphs, where the graph nodes do not contain cycles, but may have duplicates and intersections with other sets or branches, and also have a non-unique path (sequence of nodes) from a root to a node, while the roots of trees can have a list representation and expand when necessary or when interacting with a user as a search result, using a tree as a way to interactively refine a search query.

2. Aggregation of information - that is, combining versions of a document / files into a document, combining documents into a project, and so on along the available hierarchy; it is possible to use an aggregated index for search.

3. Sequential aggregation of document versions, due to the accumulation of differences between the initial version and the final one.

4. Compliance with the requirements of modern Information Security - namely, the management of permissions in the search.

Significant differences between the claimed invention and known technical solutions.

1. Corporate search - search can work with initially closed information. In known solutions, Internet search is a search for all initially open information.

2. Working with duplicates - the ability to consolidate all versions, operating on the "Document" entity, which may include many versions and many files. In known solutions, work with duplicates - the Internet excludes duplicate search, they are looking for the very first version or the most demanded by users and consider it the most relevant.

3. Aggregation of information can take place (aggregation of documents into projects, and further along the hierarchy). The possibility of using a hierarchy within which the search for documents can be considered. In the hierarchy, it is possible to use additional document properties (date of creation, author of creation) and document permissions, which can be provided by document storage systems, and in the KSPI it is also possible to select tags from a document (a word, a phrase that can be used in a search). In the known solutions, various documents are issued, there is no aggregation into projects, there can be a huge number of documents in which the user can get confused.

4. The system does not use relevance analysis based on information from users. In known solutions, the search takes into account the relevance of document analysis based on information from users. Statistics are accumulated on how many users have viewed and how much time was spent on viewing.

5. The search engine can take into account all information, all versions. In known solutions, based on graphs, document links are built to select the required document, discarding unnecessary versions.

6. The KSPI may lack links from document to document. There may also be no ranking for these attributes. (A project cannot be more popular if someone links to it). KSPI can work with classified information. In known solutions, when searching for documents, they are ranked based on document-to-document links. More links (more users viewed) is considered more relevant.

7. During the search, two independent options may appear:

and. hierarchical multi-tree list view of search results;

b. the ability to manage powers, in accordance with the powers, it is possible to display either a document, or indicate a means, a method for communicating with the owner of the document, so that it is possible to get the document itself (as a result of the search - the document or the department that developed the document - full name of the head of the department, phone, e-mail ). And it is also possible to request authorization for the wanted document through corporate authorization request systems (IDM system, etc.). In known solutions, the information found is presented in the form of links to documents.

8. KSPI can operate with the essence of the document, that is, information from all versions can be consolidated and used to search for this document. Documents may also not belong to one project, but be scattered across different data storage systems. In the known solutions, the information is not consolidated; different versions of documents are provided to the user.

9. In KSPI, the search can be conducted not for a separate document, but for the entire project. Known solutions search for a separate document.

10. The search index can be built on the basis of an object - a "document" consisting of many file versions of the document. The index possibly used in the proposed system is aggregated. It can contain information from all versions of the document. (information in different versions of the document may differ fundamentally). In known solutions, the search index is built one file at a time (aka document).

11. If there is an object - "project", the search can go through many files and many versions. It is possible to carry out additional aggregation at the top level - above the document level. (eg project, projects, department directory). In the known solutions, only a single document is searched, there is no aggregation.

12. A satisfactory search result is the provision of all versions of a document, or a project, or projects, or the name of the department that deals with this (is the owner of this information), depending on the authority of the requestor. In known solutions, the search result is an open document. This search and its results are independent of the authority of the searcher.

The essence of the technical solution.

KSPI provides the user with the ability to make search queries through a web interface, a mobile application, or in another system through an application programming interface (API) providing services. The search bar can be used as the main user interface.

The search engine is intended for large and medium-sized enterprises with a large volume of documents in the enterprise's IS and a large number of employees. The KSPI acquires particular relevance in Companies with an extensive organizational structure in different regions of the country or the world.

The novelty of the KSPI lies in the aggregation of information, the form of presentation and compliance with IS requirements when providing access to documents.

The search result can be a document, project, object, top-level structure of the IS and / or contact information for interacting with the owner of the information. KSPI generates a search result based on the authority of the user making the request and the authority associated with the initial information.

A search query can be parsed using morphological analysis, a database or other storage of similar words or their components can be used, which expands the search capabilities.

KSPI interacts with the Company's IS and receives information from it to create a search index on the basis of which the search is carried out.

With the IS of the Company, KSPI can receive information about the composition of security groups, read information about permissions from the attached files themselves, directories or other data storage structures.

KSPI can also receive information about the staff structure of the enterprise, contact information, and other additional information in the form of tags, flags of the document itself, document properties and other meta - information for the subsequent display of the found information to the user if he has the necessary powers.

Using information about the structure of the document can be used to improve the quality of indexing, search and display of the document.

The obtained information about files can be aggregated into documents. CSPI can combine versions and files into a document to build a common index. However, the output of individual documents is also possible.

Also KSPI can find documents with similar structural data (titles, tags, date of creation, other meta - information), according to these features, KSPI can combine these documents into a "project", when searching, the result can be displayed as a "project". It can indicate the name of the project, the creators, owners of the "project" and other information available to the user, according to his authority. Projects can be combined into higher-level structures using similar methods.

For clear understanding, this description relates to simplified embodiments of the present invention. Many embodiments of the present technical solution will be much more complex.

Providing information to the User.

After sending a search query (the query may contain one or several words), the user can receive an aggregate response in the form of collapsed hierarchical structures (trees). The roots are top-level structures, such as the States or Regions in which the company is represented. As the tree expands, groups such as:

a) region;

b) organizations;

c) divisions;

d) projects, structures, etc.

e) document / s;

f) document versions.

This list is not complete or exclusive, other groups may be represented in the tree, those skilled in the art may create other groups that remain within the scope of the present technology. The list can be formed on the basis of statistical information in automatic mode. This list of groups should not be interpreted as the only embodiment of this element of the present technology and / or what has been described above is the only embodiment of this element of the present technology.

The user when expanding the tree can reach:

1) before the project, document and document versions, if he has the authority:

2) to the directory of the department (unit), in the absence of authority, providing the user with information about:

a) subdivision owner;

b) the head of the department;

c) resource owner;

d) contact information;

This list is not complete or exclusive, other information may be provided.

This list of information should not be interpreted as the only embodiment of this element of the present technology and / or what has been described above is the only embodiment of this element of the present technology.

Based on the information received, the user can request the necessary powers, for example:

- Full name and position can tell who you need to contact for the necessary authority. A telephone or email can be provided for authorization requests.

With an Integrated Identity Management (IDM) system, the user can request permission to view documents automatically.

It is possible that other departments are working with the requested documents (versions, files, documents, projects, objects), this information can also be displayed in the tree if there is such information and the user's authority to obtain it. In the hierarchy of versions, it may appear that a third-party department (or third-party departments) worked with this project (version, file, document, object) and the results of the work of another department may be stored in another directory, resource or IS. Depending on whether the user has the necessary permissions, he will be able to see the full detailed information. If there is no authority, he will be able to request the necessary information from a third-party department using contact information or request authority through the IDM system, if available in the Company. The implementation of security management is optional and is customer specific.

The figure shows the scheme of work of the KSPI.

KSPI consists of the following basic structures: AID 1, DBMS 2, AIS 3, SIP 4. All examples and conditional constructions used in this description are intended to understand the principle of this technical solution, and not to establish the boundaries of its scope. Those of skill in the art can devise various schemes, not separately described or shown in this description, but which embody the principles of the present technology and are within its scope.

Data source agent - AID (1) - can carry out:

a) supply of documents (files);

b) parsing (analysis) of text information, including OCR conversion of documents;

c) extraction (extraction) of keywords;

d) ranking of keywords;

f) reading properties, powers or other meta - information;

e) formation of a search index for saving in the DBMS

This list of AID functionality should not be interpreted as the only embodiment of this element of the present technology and / or what has been described above is the only embodiment of this element of the present technology. AID can be implemented on several functional ICs with different functional divisions.

Database Management System DBMS (2) is a structure in which storage is carried out:

a) morphological word forms to expand search capabilities;

b) staff or organizational structure of the enterprise obtained from the enterprise's IP;

c) powers;

d) search index.

This list of information stored in the DBMS should not be interpreted as the only embodiment of this element of the present technology and / or what has been described above is the only embodiment of this element of the present technology.

The AIS Information Aggregation System (3) is a structure in which the analysis of individual blocks of information is carried out in order to identify communities and generate search information for combining files into documents, documents into projects.

The SIP User Interface System (4) can implement interaction with the user through various software and hardware.

All principles, aspects and embodiments of the present invention as claimed herein, as well as specific examples, are intended to indicate their structural and functional bases, whether they are currently known or will be developed in the future.

Interaction of system components:

1. Data Source Agent (AID), which interacts with the information system (file server, document management system, intranet and Internet sites) (everything that contains files), receives information and powers from there. AID completely reads the file itself, takes from the information system all information about the authority, the name of the document / file, the name of the document / file, tags, flags, that is, the external attributes of the file. AID contains a set of filters that provide parsing of various file / document formats. Filters extract both text information and meta information. For example - a Microsoft Word document contains information: the text itself, markup, headings, and other meta - information (who is the creator, date of creation, etc.). This document / file is a container containing various information. The filter is a means of parsing unification of various data formats. The list of filters can be expanded as needed to process new data formats. All information received from the AID goes to the DBMS to form a search index. AID may contain a text-to-image recognition (OCR) module.

2. The Database Management System (DBMS) is a component of the system that stores the information necessary for searching. When a request is received through the SIP, the DBMS database management system is accessed and the necessary information is received from the DBMS database management system. The selection is carried out on the basis of information stored in the database management system of the DBMS using morphology, parsing the request. A situation is possible when the AID compresses the information received and sends it to the server where the DBMS is located, where the information is deployed and embedded in the DBMS database management system, since the communication channels may be weak. There may be other architecture as well. Compression - optional, maybe not.

3. AIS interacts with the database management system of the DBMS and analyzes search information in order to identify the commonality of the stored information, identifying commonality allows you to combine versions of documents into documents, documents into projects moving up. Consolidation into higher-level structures is optional.

4. SIP (web interface, API, etc.) - through which the user and / or other system can interact. The user interface generates queries to the database management system DBMS to retrieve information. Taking into account the morphology. parsing performs the queries and clarifications themselves. The interface carries out a two-way exchange of information with the database management system DBMS.

Examples of how search works

Example 1: searching the corporate network

1. The user enters into the search box any word or phrase related to the topic of interest. For example - he can enter the word - plan 2019. Next, the user launches a search engine.

2. The front-end sends a request to a search server or a group of servers where information is stored.

3. The morphological parsing of the search query begins, that is, a base or other storage of similar words is used, which expands the search capabilities.

4. The extended request enters the database, and from there the necessary documents or different versions of the sought document are obtained in the form of files.

5. The user is provided with information in an aggregated form, in the form of collapsed hierarchical structures (trees). Each structure begins with a global group that unites the top-level group of documents, for example, by regional, organizational or other common feature (Region or Head Organization), which is the root.

6. A global feature is common for most documents in this structure, then the tree opens further - tree branches - hierarchy branches formed by common features with less weight of objects under them, the final leaves of the tree are documents or document versions. As an example - the first level from the root is - directory - Region or Head Organization. In the future, the tree opens up to the next levels - it will be - Organization, then - departments, then - projects, structures, etc., at the end - documents or document versions.

7. A distinctive feature of the search system can be the presentation of search results in accordance with the authority of the person who requests the information. That is, the user may or may not have the authority to view the requested documents or organizational groups.

8. If the user has the necessary permissions (the search program receives information about the permissions of each specific user from the information system of the enterprise in which this user works), the search engine will present information to him in an aggregated form, and the user, having received information in a collapsed form, will be able to expand it (by getting additional information at each level) to the project, to the document and to the specific version of the document being sought.

9. If the user does not have sufficient authority to view this document, the search program provides him with information also in an aggregated form, but the user, when expanding the levels, will be able to get only general information about the document he is looking for. The system can provide contact information about the owner (creator, manager - holding the maximum position) of this department or department in which the document of interest was created or is located.

10. The user can use the received contact information (for example, the owner's name, position, phone number, e-mail) provided by the search system to gain access to the requested document. For example, call the contact numbers, write to the e-mail of the owner of the document. When the enterprise uses an automation system for managing authorizations, the user will be able to request authorizations in automatic mode. The user can also request permissions to search - those are the ability to search in a given directory / resource by agreeing with the owner of the resource. Search authority does not mean being able to read documents.

11. It is also possible that the document is being dealt with in other departments or enterprises (in a large holding united into a single information network). This information will also be displayed when searching for a document. The hierarchy of versions can display which department worked or is working with which version of the document and where and who currently stores the version of the document. If you have permission to view, then all the information will be displayed, if not, then only information from whom you can request the necessary permission.

12. The option check credentials may or may not be in the information retrieval system. Depending on the wishes of the customer. If this option is not available, the search result can be opened to the final information - to document versions.

Example 2 (Internet search).

The user requests information about the software, such as a game. Since the user is lazy, he writes in the request just the name of the game and nothing more (hereinafter “GAME”). The system returns the result to it in the form of several trees, the roots are typical query refinements with this game:

- reviews "GAME"

- passing "GAME"

- purchase of "IGRA"

- cheat codes "GAME"

- download "GAME"

etc. ranking by popularity of requests.

The user immediately sees that he needs to clarify what exactly he wants to get. He can at the beginning open the tree of “overviews” and see there:

- reviews in Russian;

- reviews in English;

- reviews in other languages;

Those. in principle, obvious options for clarification.

For example, he chooses - reviews in Russian and he receives a typical search engine response with such a clarification.

Then the user selects "download" and gets various sources where this game is offered to download, grouped by type:

- torrent trackers;

- sites of official sale;

- FTP server;

etc. etc.

Example 3 (internet search).

The user is interested in health and makes a request for the name of the disease (hereinafter “Disease”).

As a result of the request, it receives a typical set of refiners:

- Diagnostics "Disease";

- "Disease" treatment;

- "Disease" symptoms;

- photo “Disease”;

etc.

Suppose the user already knows what kind of "Disease" he is interested in and wants to receive information about modern methods of treatment - for this he begins to open the "treatment" tree.

Then he gets:

- traditional methods of treatment "Disease";

- drugs for the treatment of "Disease";

- medical treatment centers “Disease”;

- foreign clinics of treatment "Disease";

- folk methods of treatment "Disease";

Accordingly, he immediately sees the picture and can navigate and decide where to go, and not open every site on the first page in the hope of finding the information he needs.

For example, he chose drugs - in response, he receives the entire list of drugs.

And already opening the right one, he can get:

- description;

- application;

- acquisition;

Those. all that is needed - read about the application - opened the purchase branch and ordered the drug in the store, having previously selected a store in your city (and not going through all the stores, or not doing a separate search for each branch).

Example 4 (internet search).

The user does not know what he is looking for and wants to know what it is - to show the complexity of the problem, let's imagine that a foreigner wants to understand the meaning of the word "Spit" in Russian. He makes a request for "Spit" and, accordingly, the search system should return him all options, including corrections:

- Braid (hairstyle) - hair, woven together in length;

- Scythe - agricultural hand-held wearable tool for mowing grass;

- Spit - a sandbank on the river;

- Scythe (wires) - a bundle of wires in the car connecting the engine, gearbox and computer;

- The goat is an animal;

etc.

Those. the user will immediately understand what he is dealing with and where the request can be specified

The examples show that the interactive clarification of questions within a hierarchical multi-tree list view of search results is very user-friendly and will be relevant especially for mobile users where it is difficult to enter additional text information.

In Internet search, there is no initial imposed hierarchy, but it can be formed on the basis of already available typical information based on statistical information about the search.

This method can allow the user to refine the search query in an interactive mode (in real time) or be a static structure that is disclosed in real time.

The user can reach the final document on the site - that is, at the end of the interactive search, a link is displayed to the site page with the information sought.

Claims (2)

1. A method for extracting information, implemented in a corporate information retrieval system, capable of extracting results from a set of different sources, comprising stages at which a request is received at the input; at the output, this query is issued in a set of different sources, characterized in that it is characterized by a hierarchical multi-tree list representation of search results in the form of one or many sets of connected graphs, the nodes of which do not contain cycles, but may have duplicates and intersections with other sets or branches , while they can have a non-unique path, characterized by a sequence of nodes, from root to node, the roots of trees have a list representation and can be opened when necessary or when interacting with the user interactively; involves executing search queries through a web interface, a mobile application, or in another system through the API providing services, as the main user interface, a search bar can be used; further parsing a search query using morphological analysis, using a database or another repository of similar words or their components, to expand the search capabilities, using a corporate information retrieval system; accessing the search index base through the corporate information search system and extracting search results from it in the form of a link, document, project, file, object, contact information to interact with the information owner; formation of search results using the corporate information retrieval system, based on the authority of the user making the request and the authority associated with the initial information; filling the search index base through the interaction of the corporate information search system with the company's information system and receiving information from it to create a search index, on the basis of which the search is carried out; using information about the structure of the document to improve the quality of indexing, search and display of the document; aggregation of the information received from files into documents, while the files are considered as versions of the document; combining file versions into a document through the corporate search system to build a common index; finding documents with similar structural data from meta-information through the corporate information search system for the subsequent display of the found information to the user if he has the necessary powers; combining these documents according to these characteristics through the corporate information search system into higher-level structures, such as "project", "direction"; displaying the search results in the form of a "project", "direction", in which the name of the "project", creators, owners of the "project", "directions" and other information available to the user according to his authority can be indicated. 2. Corporate information retrieval system (KSPI) for implementing the method according to claim 1, containing basic structures in the form of a data source agent, a database management system, an information aggregation system, a user interface system; the data source agent is intended for the delivery of documents, files, extraction of keywords, ranking of keywords, indexing, reading properties, powers or other meta-information, as well as for using the technology of converting a graphic image of text into computer text using an algorithm for recognizing graphic images, generating index information that enters the database management system; the data source agent contains a set of filters that provide parsing of various file formats, and the meta-information comes from both the agent and the filter from which it is extracted, forming a search index for saving in the database management system, and the data source agent can be implemented on several functional information systems; the database management system is designed to store: morphological word forms to expand search capabilities, staff or organizational structure of an enterprise obtained from enterprise information systems, powers, search index; the information aggregation system is designed to concentrate individual information flows into a single structure; with the ability to generate index information for combining files into documents, documents into projects; the user interface system is designed to implement interaction with the user through various software and hardware; wherein the output of the data source agent is connected to the input of the database management system, and the database management system is connected to the information aggregation system and to the user interface system; the data source agent interacts with the information system, file servers, document management system, intranet and Internet sites, receives information and powers from there, the data source agent completely reads the file itself, selects from the system all information about authority, file name, file name, tags , flags, that is, the external attributes of the file and sends, along with all the meta-information, authority, creation date, creator, which he took from the information system, to the database management system; the data source agent is designed to process information, contains a recognition module, generates index information that can be placed in the database management system, contains a set of filters that provide parsing of various file formats, and meta-information comes from both the agent and the filter that extracts; the database management system receives meta-information, as well as information for creating a search index; the database management system is a component of the corporate information retrieval system that stores indexes; when a request is received through the user interface system, the database management system is accessed and the necessary information is obtained from the database management system; the selection is carried out on the basis of information that is stored in the database management system using morphology, parsing the request; the data source agent is configured to compress and send the received information to the server where the database management system is located, where the information is deployed and embedded in the database management system, since the communication channels may be weak; the user interface system generates requests to the database management system to retrieve information, taking into account morphology, parsing words, performs the requests and clarifications themselves, the user interface system carries out a two-way exchange of information with the database management system.

Images