Document retrieval system

Abstract

To efficiently retrieve a document according to an input keyword.SOLUTION: A document retrieval system stores a database for document retrieval indicating a table of contents hierarchical structure of one or more documents. A document retrieval system receives a retrieval keyword input from a user, retrieves an item matching the retrieval keyword in a database for document retrieval, and determines a result item to be presented to the user from a plurality of items by selecting the items in order from a high-order hierarchy to a low-order hierarchy according to an index hierarchical structure including the plurality of items matching the retrieval keyword. The selection in the at least one hierarchy is in accordance with a selection by the user from multiple choice items presented to the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to document retrieval by input keywords. [Background technology]

[0002] The number of product-related inquiries from users is increasing. Support staff or automated chatbots respond to inquiries by searching product manuals. However, the same term may appear in various manuals covering multiple devices and functions. When a keyword search returns a large number of results, it takes a long time to find the desired information.

[0003] Japanese Patent Application Publication No. 2018-18279 is a related technology of the present disclosure. Japanese Patent Application Publication No. 2018-18279 discloses a document search device that can output search results faster than when a full-text search is performed using classification of attribute information. More specifically, the document search device 10 includes a reception unit 12 that receives multiple hierarchical search terms, and a setting unit 16 that sets a first content among multiple hierarchical contents included in structured document data 26 as the search range for a first search term received by the reception unit 12, and sets a second content that is at the same level as the first content or at a lower level than the first content as the search range for a second search term that is at a lower level than the first search term (e.g., abstract). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2018-18279 Summary of the Invention [Problem to be solved by the invention]

[0005] Patent Document 1 discloses a document search method based on a hierarchical table of contents structure. With this method, when a search keyword is specified, the name of a document containing the first keyword from multiple documents is extracted, and then a chapter containing the second keyword is searched for in the table of contents of that document. This method repeats the same procedure to identify sections, paragraphs, and so on. Therefore, unless the user enters a keyword that matches the terminology of a chapter or section, they will not be able to find the desired chapter or section. [Means for solving the problem]

[0006] A document search system according to one embodiment of the present invention includes a processor and a storage device, wherein the storage device stores a document search database that indicates a hierarchical table of contents structure of one or more documents, and the processor receives search keywords entered by a user, searches the document search database for items that match the search keywords, and determines result items to be presented to the user from the multiple items by selecting items from a higher hierarchy to a lower hierarchy according to the hierarchical table of contents structure that includes multiple items that match the search keywords, wherein the selection in at least one hierarchy follows the selection made by the user from multiple-choice items presented to the user. [Effects of the Invention]

[0007] According to one aspect of the present invention, documents can be efficiently searched for according to input keywords. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an overview of a document search system according to a first embodiment. [Figure 2] 1 shows an example of the configuration of a document search system. [Figure 3] The GAD manual is shown. [Figure 4] An example of a thesaurus is shown below. [Figure 5] 1 shows an example of a document search database. [Figure 6]10 is a flowchart of an example of pre-processing for enabling a user to search for documents. [Figure 7] 10 shows a flowchart of an example of a search process. [Figure 8] 1 shows an example of a dialogue between a document retrieval system and a user in document retrieval by input keywords. [Figure 9] 1 shows an example of the configuration of a document search system 10 according to a second embodiment. [Figure 10] 1 shows an example of the configuration of a storage device. [Figure 11] 10 shows an example of the configuration of various information included in the configuration information. [Figure 12] 10 shows an example of a dialogue between a document search system and a user in a document search using an input keyword in the second embodiment. [Figure 13] 10 shows a flowchart of an example of a search process in the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of the present specification will be described with reference to the accompanying drawings. In the accompanying drawings, functionally identical elements may be designated by the same numerals. Note that the accompanying drawings show specific embodiments and implementation examples in accordance with the principles of the present invention, but these are for understanding the present invention and are not to be used to interpret the present invention in a limiting manner.

[0010] Although the embodiments of the present specification are described in sufficient detail to enable those skilled in the art to practice the present invention, it should be understood that other implementations and forms are possible, and that changes in configuration and structure and substitutions of various elements are possible without departing from the scope and spirit of the technical concept of the present invention. Therefore, the following description should not be interpreted as being limited thereto.

[0011] Additionally, embodiments herein may be implemented in software running on a general purpose computer, in specialized hardware, or a combination of software and hardware, as described below.

[0012] When describing each process in the examples of this specification using "each processing unit as a program" as the subject (acting entity), the program is executed by a processor (CPU, etc.) to perform the specified process using memory and a communication port (communication control device), so the description may also use the processor as the subject. [Example]

[0013] FIG. 1 is a diagram illustrating an overview of a document search system 10 according to an embodiment of the present specification. FIG. 1 shows an example of the document search system 10 that searches for various manuals 17 for storage systems. The manuals are examples of documents. FIG. 1 shows, as examples of manuals, a GAD (Global Active Device) manual and a TI (Thin Image) manual, which are manuals related to storage systems.

[0014] The document search system 10 includes a thesaurus 11 and a document search database 13. The thesaurus 11 indicates the correspondence between technical terms specific to a manual and general terms in a manual 17. The document search database 13 includes information on the hierarchical table of contents (chapter divisions) extracted from each of the various manuals 17, and the text (including titles and main text) at each level. Each document can have a different hierarchical table of contents. For example, the hierarchical table of contents of a document can include, from the top down, all or part of chapters, sections, and paragraphs. Note that each document can have any number of levels, and each level can include the title and / or main text of the corresponding item. Furthermore, the names of the levels are arbitrary.

[0015] The document search system 10 can efficiently search for the manual and the portion of the manual that the user is looking for through dialogue with the user 50. Figure 1 illustrates a process for extracting a portion corresponding to an inquiry by the user 50 from various manuals 17 related to storage systems.

[0016] A user 50 inputs a search keyword to the document search system 10. In the example shown in FIG. 1, "pair creation method" is input as the search keyword. The document search system 10 searches the document search database 13 for items (e.g., clauses or terms) at the lowest level of each document that matches the search keyword. The lowest level is a level that has no levels below it.

[0017] In the example shown in Figure 1, two items in the GAD manual and one item in the TI manual match the search keywords. Specifically, they are the items "1.1 Pair Management" and "2.2 Pair Operation" in the GAD manual and the item "1.1 Pair Management" in the TI manual.

[0018] The document search system 10 extracts the hierarchical table of contents structure above the items that match the search keywords from the document search database 13. For example, the titles of each level are extracted. For example, for the item "1.1 Pair Management" in the GAD manual, the title of the chapter immediately above it, "Chapter 1 CLI," and the title of the manual above that, "GAD Manual," are extracted. CLI stands for Command Line Interface.

[0019] The document retrieval system 10 presents the information, in this case the titles, from the top to the bottom of the extracted hierarchical structure of the table of contents to the user 50, and asks questions to have the user select items sequentially. When presenting information to the user 50, the document retrieval system 10 refers to the thesaurus 11 and converts technical terms specific to the manual into general terms. Note that the thesaurus 11 or the processing referring to it may be omitted in part or in whole.

[0020] For example, GAD, TI, and RM (RAID Manager) are technical terms specific to the manual, while CLI, GUI (Graphical User Interface), API (Application Programming Interface), and HA (High Availability) are general technical terms (general terms).

[0021] As shown in FIG. 1, the document search system 10 may present titles using both technical terms and general terms, or may present only general terms. General terms can make user selections easier. The thesaurus 11 may also be used for searches using search keywords. The document search system 10 may also search for synonyms of the search keywords in the search database 13.

[0022] 1, the document retrieval system 10 first presents candidates for documents at the highest level and allows the user to select one. In this example, "GAD Manual" and "TI Manual" are presented, and "GAD Manual" is selected.

[0023] Next, the document retrieval system 10 presents candidate items in the chapter, which is the next level of the selected "GAD Manual," and allows the user to select one. In this example, "CLI" and "API" are presented, and "CLI" is selected.

[0024] By repeating the user's selection along the above hierarchy, the items to be presented to the user as search results are finally determined. In the example of Figure 1, the item "1.1 Pair Management" from the GAD manual is selected as the information to be presented.

[0025] The document search system 10 presents information about the selected item to the user. In the example shown in FIG. 1, the document search system 10 provides the text of the identified item to a generation AI (Artificial Intelligence) 15, which generates a summary. The document search system 10 presents the user with a link to the original manual along with the summary output by the generation AI 15. This allows the user to easily learn about the information about the target item and to easily refer to the original manual. Note that at least one of the summary and the link may be omitted, and the title of the item may be presented instead.

[0026] In the above example, the lowest level items that match the search keyword are searched for. This allows the most narrowed down items to be presented to the user. In another example, the document search system 10 searches not only the lowest level but also other levels. For example, all levels may be searched.

[0027] The document search system 10 stops the user's selection when the user selects an item that matches the search keyword. For example, in the above example, assume that the text (explanation) of the item "Chapter 1 CLI" includes "pair creation." When the user selects the item "Chapter 1 CLI," the document search system 10 determines that item as the final selected item. The document search system 10 generates a summary of the chapter using the generation AI 15 and presents it to the user along with a link to the original manual.

[0028] As described above, by having the user select items at each level through dialogue, the user can efficiently search the manual and find the information they want. Furthermore, by displaying general terms together with or instead of technical terms, even beginners who are unfamiliar with technical terms can select items appropriately.

[0029] 2 shows an example of the configuration of a document retrieval system 10 according to an embodiment of the present specification. The document retrieval system 10 includes a document retrieval computer 200 and a user terminal 51, which communicate with each other via a network.

[0030] An example of the hardware configuration of the document retrieval computer 2 is described below, but the user terminal 3 may have a similar configuration. The document retrieval computer 200 includes a CPU (processor) 201 that executes various programs, a memory (main storage device) 202 that stores the various programs, and an auxiliary storage device 203 that stores various data. The CPU 201 can include one or more cores, and the memory 202 is, for example, a DRAM that includes a volatile storage area. The auxiliary storage device 203 is, for example, an HDD (hard disk drive) or flash memory, and can provide a non-volatile storage area.

[0031] The document retrieval computer 200 further includes an output device 204 for presenting information to the user of the device, an input device 205 for inputting instructions, images, etc. from the user, and a network interface 206 for communicating with other devices. These are interconnected by a bus 207. Instead of the input / output devices of the document retrieval computer 200, the user may use a user terminal 51 connected to the document retrieval computer 200 via a network.

[0032] The functional units of the document retrieval computer 200 can be implemented, for example, by the CPU 201 operating in accordance with a program. The CPU 201 reads and executes various programs from the memory 202 as needed. The memory 202 can store programs and data used by them. Each program and reference data is loaded into the memory 202 from the auxiliary storage device 203, for example, and executed and processed by the CPU 201. At least a portion of the functional units may be configured with logic circuits.

[0033] The output device 204 is composed of devices such as a display, printer, and speaker. The input device 205 is composed of devices such as a keyboard, mouse, and microphone. The output device 204 presents input results from the user and also presents processing results by the document retrieval computer 200. Instructions from the user are input to the document retrieval computer 200 by the input device 205. When a user terminal 51 is used, its input / output devices function in the same way, and the output device 204 and input device 255 can be omitted.

[0034] The network interface 206 receives data transmitted from other devices connected via a network, including the user terminal 51, and transmits the processing results of the document retrieval computer 200 to other devices. Note that some devices may be omitted.

[0035] The document retrieval computer 200 stores a document analysis program 210 and a thesaurus generation program 220. The document analysis program 210 and thesaurus generation program 220 execute pre-processing for a user search. Note that the processing results of the document analysis program 210 and thesaurus generation program 220 may be prepared in advance by a system designer.

[0036] The document analysis program 210 includes a hierarchical structure extraction program 211, a main text extraction program 212, and a glossary extraction program 213. The hierarchical structure extraction program 211 references the table of contents of the document and extracts the table of contents hierarchical structure. The main text extraction program 212 extracts the main text at each level in the document. The glossary extraction program 213 extracts the glossary contained in the document. The glossary is an item that explains the terms in the document. The thesaurus generation program 220 generates the thesaurus 11.

[0037] The document retrieval computer 200 further includes a dialogue program 230, a document summary generation program 240, and a document retrieval program 250. The dialogue program 230 includes a hierarchical structure-based questioning program 231 and a terminology conversion program 232.

[0038] The dialogue program 230 is a program for executing dialogue with the user 50 for document search. The hierarchical structure-based question-reply program 231 asks questions in response to search keyword inputs from the user 50 and answers to the questions. The technical term conversion program 232 uses the thesaurus 11 to convert technical terms in document information presented to the user into general terms.

[0039] The document summary generation program 240 includes a generation AI 15, which generates a summary to be presented to the user 50 from one or more items in a document. The generation AI 15 may run on another computer and be called by the document summary generation program 240. The document search program 250 extracts items that match the search keywords entered by the user 50 and items above them from the document search database 13.

[0040] The document retrieval computer 200 stores data to be referenced or processed by a program. Specifically, the document retrieval computer 200 stores documents 17, a thesaurus 11, a document retrieval database 13, and authentication information 260. The authentication information 260 is used to access configuration information of a storage device described in the second embodiment.

[0041] FIG. 3 shows a GAD manual 310 as an example of one document included in document 17. GAD manual 310 includes three chapters, each titled "CLI," "API," and "Error Codes." Chapter 1, CLI, includes one section, "1.1 Pair Management." Chapter 2, API, includes two sections, "2.2 Overview" and "2.2 Pair Operations." Chapter 3, Error Codes, consists of a table that associates error codes with meanings.

[0042] The following explains the hierarchical structure of the table of contents of the GAD manual 310. For example, the hierarchy (item) of Chapter 2 includes the title "API" 311 but does not include the main text. The hierarchy of Section 2.1 includes the title "Overview" 312 and the main text 313. The hierarchy of Section 2.2 includes the title "Pair Operation" 314 and the main text 315.

[0043] The hierarchy of Chapter 3 includes the title "Error Code" 316, but does not include any text. The items below it are the entries in the table. One entry constitutes one item directly under Chapter 3. In each item, the error code 317 is the title, and the meaning of the error code 318 is the text.

[0044] FIG. 4 shows an example of the thesaurus 11. The thesaurus 11 shows the correspondence between technical terms specific to a manual and general terms. In the structural example shown in FIG. 4, the thesaurus 11 has a technical term column 111 and a general term column 112. The general term column 112 shows general terms corresponding to each technical term specific to a manual shown in the technical term column 111. As in the example of "LU" and "Logical Unit" shown in FIG. 4, the abbreviation and the official name may have a relationship of technical term and general term.

[0045] FIG. 5 shows an example of the document search database 13. The document search database 13 manages the hierarchical table of contents structure of each document 17. FIG. 5 shows, as examples, a document with the title "GAD Manual" and a document with the title "TI Manual" among the documents 17. The document 17 includes many more documents. The number of documents registered in the document search database 13 can be any number greater than or equal to one.

[0046] The document search database 13 contains information on the hierarchical table of contents (chapter structure) extracted from various manuals 17, and the text of each level (including titles and main text). Each document can have a different hierarchical table of contents. The number of levels and the number of items in each level of each document are arbitrary, and the items in each level can include titles and / or main text.

[0047] In Fig. 5, the top level 510 is a document level, which is made up of document items. In Fig. 5, the "GAD Manual" and the "TI Manual" are shown as examples, each of which does not include the main text and is made up of only a title. An example of the GAD Manual is shown in Fig. 3.

[0048] The next lower level 520 of the GAD manual is the chapter level, which is made up of three items: Chapter 1, Chapter 2, and Chapter 3. Each item includes a title but does not include a main text. The next lower level 540 of the "Chapter 1 CLI" item is the section level, which is made up of multiple items. Figure 5 shows one item, "1.1 Pair Management," as an example. This item includes the title "1.1 Pair Management" and the main text "Create a pair...".

[0049] The next lower level 550 below the "Chapter 2 API" item is a section level and is made up of multiple items. Figure 5 shows one item, "2.2 Pair Operation," as an example. This item includes the title "2.2 Pair Operation" and the text "Pair creation is...". The next lower level 560 below the "Chapter 3 Error Code" item is a level corresponding to a table and is made up of multiple items. Figure 5 shows three items as examples. These items include a title and a text. As shown in Figure 3, the title indicates the error code, and the text indicates the meaning of the error code.

[0050] The next lower level 530 of the TI manual is the chapter level, which is made up of multiple items. Figure 5 shows the item "Chapter 1 CLI" as an example. This item includes a title but not a body text. The next lower level 570 of the "Chapter 1 CLI" item is the section level, which is made up of multiple items. Figure 5 shows one item "1.1 Pair Management" as an example. This item includes the title "1.1 Pair Management" and the body text "Create a pair...".

[0051] As explained with reference to FIG. 1, assume that user 50 inputs "pair creation method" as a search keyword. Here, document search system 10 searches for items matching the search keyword in the titles and text of each item in document search database 13. In the example of FIG. 5, words matching the search keyword are found in the text of items 541, 551, and 571. These are candidates for the item that user 50 is looking for.

[0052] The document search system 10 extracts a hierarchical structure that includes an item that matches the search keyword, that is, extracts that item and items in a higher hierarchy. In the example shown in Fig. 5, the items in the hierarchy above item 541 are item 521 and item 511. The items in the hierarchy above item 551 are item 522 and item 511. The items in the hierarchy above item 571 are item 531 and item 512.

[0053] 1, the document search system 10 narrows down the items from the highest hierarchy in dialogue with the user 50, and ultimately determines the items for which information is to be presented to the user 50. Note that the search targets in the document search database 13 may be only the items in the lowest hierarchy (a hierarchy with no lower hierarchy), or only the main text of each item.

[0054] Next, the processing performed by the document retrieval system 10 will be described. Figure 6 is a flowchart of an example of pre-processing to enable document retrieval by the user 50. The pre-processing involves creating a document retrieval database 13 and a thesaurus 11. Note that the document retrieval database 13 and thesaurus 11 may be prepared in advance by a system designer instead of the document retrieval system 10.

[0055] The pre-processing for one document will be described with reference to Fig. 6. The hierarchical structure extraction program 211 analyzes the structure of the document and stores it in the document search database 13 (S11). Various document structure analysis techniques are known, and detailed explanations will be omitted. The hierarchical structure extraction program 211 may use any document structure analysis technique.

[0056] Next, the text extraction program 212 extracts text from the document in the document search database 13 (S12). The thesaurus generation program 220 then performs statistical processing on the extracted text of the document together with general text, such as text published on an open network, to create a thesaurus (S13). A word embedding technique that generates word vectors can be used to create a thesaurus. Various techniques are known for creating a thesaurus, and Word2Vec, for example, can be used. The thesaurus generation program 220 may use any of these techniques.

[0057] Next, the glossary extraction program 213 extracts the glossary from the document and adds the glossary terms as technical terms to the thesaurus 11 (S14). For example, pairs of abbreviations and full names, or terms generated from glossary explanations can be registered as general terms.

[0058] Next, a description will be given of the search processing by the document search system 10. Fig. 7 shows a flowchart of an example of the search processing. First, the interactive program 230 displays a keyword search screen on the display device of the user terminal 51 (S21). The user 50 inputs a search keyword on the keyword search screen, and the interactive program 230 receives the search keyword (S22).

[0059] Next, the document search program 250 searches for titles and texts that match the search keywords and extracts a hierarchical structure containing those items (S23). A search for sentences (including titles) that match the search keywords can use known techniques. For example, one method is to search for sentences that contain the input keywords or their synonyms. At this time, the thesaurus 11 may be referenced to absorb the variations in spelling between technical terms and general terms in keyword searches. Then, the input keywords are decomposed into N-grams to determine whether they are included in the sentence, or the input keywords are converted into vector embeddings to search for sentences that are converted into similar vectors.

[0060] Next, the hierarchical structure-based question-reply program 231 presents the user 50 with options for items, starting from the top of the extracted hierarchical structure, and allows the user 50 to select an item, ultimately narrowing down the items to be presented (S24). The options may be presented, for example, by displaying the title of the item, or by displaying the title and a summary of the text. The summary may be generated by the document summary generation program 240. This makes it possible to present information useful for the user's selection. The hierarchical information structure-based question-reply program 231 converts technical terms into general terms using the technical terminology conversion program 232, and displays these in place of or together with the technical terms. This allows the user to easily understand the question-reply even without knowledge of technical terms.

[0061] Next, the document summary generation program 240 uses the generation AI 15 to generate a summary for the narrowed-down item. Here, the document summary generation program 240 generates a summary of the narrowed-down item or the narrowed-down item and the main text at a lower level (S25). The summary enables the user to efficiently understand the content. Note that a summary may be generated without using the generation AI 15. Next, the dialogue program 230 displays the document summary and a link to the original text (S26). This ends this flow.

[0062] Here, the interactive narrowing down of items and generation of summaries will be described with reference to Fig. 5. In the example shown in Fig. 5, the items that match the input keyword are items 541 and 531. The items in the hierarchy above item 541 are items 521 and 511. The item in the hierarchy above item 531 is item 512.

[0063] The query program 231 presents items 511 and 512 in the top hierarchy 510 to the user 50 as options. If the user 50 selects item 511, the item whose summary is ultimately provided to the user is item 541. Item 541 is the lowest hierarchy, and there are no lower hierarchy levels. The document summary generation program 240 generates a summary of item 541.

[0064] On the other hand, if user 50 selects item 512 in hierarchy 510, the item whose summary is ultimately provided to the user is item 531. The hierarchical structure of item 531 includes a lower hierarchy 570. Document summary generation program 240 generates a summary from the text of item 531 and the items in lower hierarchy 570.

[0065] Another example will be described. Suppose one item matches the search keyword, and furthermore, multiple items in one lower hierarchy match the search keyword. For example, suppose item 521, item 541 in lower hierarchy 540, and other items match the search keyword. The query program 231 may perform the necessary selections up to the lowest hierarchy where items matching the search keyword exist, and then allow the user to select one of the three items. Alternatively, the query program 231 may stop the query when item 521 is selected, and display a summary of item 521 and its lower hierarchy. Here, the user may be prompted to select one item from item 541 and other items in lower hierarchy 540. This allows the user 50 to further narrow down the parts to be referenced.

[0066] If both items in the upper and lower hierarchy match the search keyword, for example, if both item 541 and its upper hierarchy item 521 match the search keyword, the document summary generation program 240 may generate a summary of the lower hierarchy item without referring to the upper hierarchy item.

[0067] Fig. 8 shows an example of a dialogue between the document retrieval system 10 and the user 50 in a document search using an input keyword. Fig. 8 corresponds to the dialogue described with reference to Fig. 1. The user 50 inputs "pair creation method" as a search keyword into the user terminal 51 (S31).

[0068] The document search system 10 finds items 541, 551, and 571 that match the input keyword in the document search database 13 shown in Figure 5. From the hierarchical structure of items 541, 551, and 571, the document search system 10 presents items 511 and 512 in hierarchy 510 as options (S32). That is, their titles, GAD Manual and TI Manual, are presented as options. At this time, the technical terms GAD and TI are converted into the general terms HA and Snapshot, and the technical terms and general terms are listed side by side.

[0069] Next, the user 50 selects the GAD manual from the GAD manual and the TI manual (S33). Next, the document retrieval system 10 presents items 521 and 522 in the hierarchy 520 as options (S34). That is, their titles, CLI and API, are presented as options.

[0070] Next, the user 50 selects a CLI from the CLI and API (S35). The document retrieval system 10 generates a summary of the text of the item 541 and presents it to the user 50 together with a link to the GAD manual (S36).

[0071] As described above, in the hierarchical table of contents structure of the keyword search results (found candidate items), the titles of the items in each hierarchy are displayed from the highest hierarchy to the lowest hierarchy, and the user is allowed to select an item, thereby enabling the user to efficiently find the item for which information should be presented. Note that instead of the dialogue using text information (visual information) as described above, the dialogue may be carried out using audio information. [Example]

[0072] An embodiment of the present specification will be described below. Differences from the first embodiment will be mainly described. Unless otherwise specified, the description of the first embodiment may be applied to this embodiment. In this embodiment, the document search system 10 automatically selects an item in the hierarchical table of contents structure by referring to the configuration information of the device used by the user, thereby reducing the number of times the user is asked questions.

[0073] 9 shows an example of the configuration of a document search system 10 according to an embodiment of the present specification. The document search system 10 acquires configuration information from a device used by a user using authentication information 260 for accessing the device. In the example configuration shown in FIG. 9, the device used by a user 50 is a storage device 40.

[0074] The authentication information 260 has a user ID column, a device ID column, a device IP address column, and a password column. The user ID column indicates the user ID. The device ID column indicates the ID of the device used by the user. The device IP address column indicates the IP address of the device. The password column indicates the password for accessing the device and obtaining configuration information. The authentication information 260 is registered in advance in the document search system 10.

[0075] The document search system 10, for example, includes a field for inputting a user ID on a keyword search screen and acquires the user ID from the user 50. The document search system 10 acquires information for accessing the storage device 40 used by the user from the authentication information 260 using the acquired user ID as a key. The document search system 10 accesses the IP address of the storage device 40 acquired from the authentication information 260, and acquires the configuration information 45 from the storage device 40 after authentication using a password.

[0076] In this embodiment, the document retrieval system 10 references the configuration information 45 of the storage device 40 used by the user 50 and generates a query in response to a search request specifying a keyword from the user 50. In the example shown in Fig. 9, the first query by the document retrieval system 10 is to select between "CLI" and "API" in the GAD manual. In the configuration example shown in Fig. 1, before presenting this option, the user 50 is presented with the options of the GAD manual and the TI manual. In this embodiment, this first query is omitted based on the configuration information of the storage device 40 used by the user, and the GAD manual is automatically selected.

[0077] 10 shows an example configuration of a storage device 40. The storage device 40 includes a CPU 401 that executes various programs, a memory 402 that stores the various programs, and a storage medium 408 that stores data from the host (host data). The CPU 401 can include one or more cores, and the memory 402 is, for example, a DRAM that includes a volatile storage area. The storage medium 408 is made up of multiple storage drives, and the storage drives are, for example, HDDs or SSDs (Solid State Drives).

[0078] The storage device 40 includes a network interface 406 for communicating with the host and the document search system 10, and an external storage interface 407 for communicating with other storage devices. These are interconnected by an internal network.

[0079] The memory 402 stores programs executed by the CPU 401 and configuration information 45 of the storage device 40. A storage system program 421 and a configuration information management program 422 are shown as examples of the programs. The storage system program 421 executes processing related to host data, including processing IO requests to the storage device 40. The configuration information management program 422 manages the configuration information 45 and updates the configuration information 45 in accordance with the operation of the storage device 40.

[0080] The configuration information 45 includes a current setting value 454, an operation history 452, and an event log 453. Fig. 11 shows an example of the configuration of various information included in the configuration information 45. The current setting value 451 includes device information 471, a function usage status 472, and fault information 473. The device information 471 includes the model name and version. The function usage status 472 indicates whether each of the functions implemented in the storage device 40 is being used. In the example of Fig. 11, the GAD function is being used, and the TI function is not being used. The fault information 473 is information about a fault within the storage device 40, and indicates the component in which the fault has occurred.

[0081] The operation history 452 shows the history of operations performed on the storage device 40. The operation history 452 includes a date and time column 481, a user column 482, an operation column 483, and a success / failure column 484. The date and time column 481 shows the date and time of the operation. The user column 482 shows the user who performed the operation. The operation column 483 shows the content of the operation. The success / failure column 484 shows the success or failure of the operation.

[0082] The event log 453 is a log of events that have occurred in the storage device 40. The event log indicates the date and time when the event occurred and its contents.

[0083] 9, the document search system 10 refers to the configuration information 45 of the storage device 40 used by the user 50 and generates a query in response to a search request specifying a keyword from the user 50. In the example shown in Fig. 9, the first query by the document search system 10 is a selection of "CLI" and "API" in the GAD manual.

[0084] The hierarchical structure query program 231 refers to the configuration information 45 before presenting the options of the GAD manual and the TI manual in the first query. The function usage status of the current setting value 451 indicates that the GAD function is in use and the TI function is not in use. Therefore, instead of a query requesting the selection of the manuals for these functions, the hierarchical structure query program 231 automatically selects the GAD manual and generates a query for selection in the next hierarchy. In FIG. 9, the first query presents the option of the CLI or API chapter in the GAD manual.

[0085] The hierarchical structure-based questioning program 231 may present recommended options instead of automatically selecting options. For example, the GAD manual and the TI manual may be presented as options, and a message indicating that GAD is recommended may be displayed. In addition, the reason for the recommendation, for example, that GAD is in use and TI is not in use, may be displayed.

[0086] In this way, the hierarchical structure-based question-reply program 231 refers to the configuration information 45 and determines whether to automatically select an option or present recommended options before each question for narrowing down items through dialogue as described in Example 1. The combination of the term and the portion of the configuration information to be referred to for the determination, and the rules for determining recommended options may be set in advance in the hierarchical structure-based question-reply program 231.

[0087] For example, when device functions are included in the information presented for selection, whether or not the functions are in use in the current setting values ​​451 is referenced, and functions in use are given priority over unused functions. When an entry within a predetermined period in the operation history 452 or the event log 453 includes a term of an option (such as a function or component) or an item previously associated with the option, that term may be given priority. When multiple terms are included in entries within a predetermined period, the term in the most recent entry may be given priority. For example, a function or component with the most recent reported failure may be recommended.

[0088] Different priorities may be assigned to different types of information in the configuration information 45. If different terms are assigned higher priorities for different types of information, the recommended terms may be selected according to the priority of the information type. Alternatively, if different terms are prioritized for different types of information, all options may be presented to the user without automatic selection or recommendation.

[0089] 12 shows an example of a dialogue between the document retrieval system 10 and the user 50 in a document retrieval using an input keyword. Compared to the example shown in FIG. 8, the first query S52 by the document retrieval system 10 is different from the first query S32 in FIG. 8. The query S52 recommends the selection of a GAD manual and indicates the reason for the recommendation. As described above, the GAD manual may be automatically selected by the document retrieval system 10 without query S52.

[0090] FIG. 13 shows a flowchart of an example of search processing in this embodiment. Compared to the processing example shown in FIG. 7, step S34 is replaced with step S64. The other steps are the same. In step S64, the hierarchical structure-based questioning program 231 automatically selects an item or presents options to the user to select, starting from the top level of the hierarchical structure, and ultimately narrows down the items for which information is presented. As with the processing example shown in FIG. 7, the options are presented as general terms instead of or in addition to technical terms. As described above, recommended options may be presented instead of the automatic selection indicated by selection instructions. [Explanation of symbols]

[0091] 10 Document Search System 15 Summary generation AI 201 CPU 202 memory 11 Thesaurus 13 Document search database 17 documents 50 users

Claims

1. A document retrieval system, comprising: a processor; a storage device, The storage device includes: storing a document search database showing a hierarchical table of contents of one or more documents; The processor: Receive search keywords entered by the user, searching the document search database for items matching the search keywords; determining a result item to be presented to the user from among the plurality of items by selecting items in order from a higher hierarchy to a lower hierarchy according to a hierarchical table of contents structure including the plurality of items that match the search keyword; A document retrieval system, wherein the selection in at least one tier follows a selection by the user from multiple choice items presented to the user.

2. 2. The document retrieval system according to claim 1, The processor presents multiple-choice items to the user at each level until the number of item options is one, and selects one from the multiple-choice items in accordance with a selection made by the user.

3. 2. The document retrieval system according to claim 1, The processor includes, in the presentation of the multiple choice items to the user, general terms converted from technical terms in the multiple choice items using a thesaurus.

4. 4. A document retrieval system according to claim 3, The processor includes both the technical terms and the general terms in presenting the multiple choice items to the user.

5. 2. The document retrieval system according to claim 1, A document search system in which the processor uses a thesaurus to absorb spelling variations between technical terms and general terms when searching for items that match the search keywords.

6. 2. The document retrieval system according to claim 1, The processor presents to the user a summary based on the text of the result items and items at a lower level than the result items.

7. 2. The document retrieval system according to claim 1, the storage device stores configuration information of a device used by the user; the processor searches the configuration information for information related to each of the multiple choice items in a selection from the multiple choice items in the hierarchy; A document search system that selects one option item on behalf of the user based on information related to each of the multiple option items in the configuration information, or presents recommended option items to the user.

8. A document retrieval method, comprising: The system stores a document search database that indicates a hierarchical table of contents of one or more documents; The document search method includes the steps of: Receive search keywords entered by the user, searching the document search database for items matching the search keywords; determining a result item to be presented to the user from among the plurality of items by selecting items in order from a higher hierarchy to a lower hierarchy according to a hierarchical table of contents structure including the plurality of items that match the search keyword; A document retrieval method, wherein the selection in at least one hierarchy is in accordance with a selection by the user from a selection of candidate items presented to the user.

Images