Drawing description generation program and drawing description generation device

Abstract

From the drawing, a description about the drawing is generated. [Solution] The description generation device 1 includes a Markdown notation conversion unit 11 that inputs a first prompt including a drawing 21 and instructions to convert to Mermaid notation to a generation AI system 3, and obtains a chat completion object 52 including Mermaid notation text 23 from the generation AI system 3, a Markdown notation extraction unit 12 that extracts the Mermaid notation text 23 from the chat completion object 52, and a description conversion unit 13 that inputs a second prompt including instructions to generate a description of the chat completion object 52 and the drawing 21 to the generation AI system 3, and obtains a chat completion object 54 including the description 25 of the drawing 21 from the generation AI system 3.

Description

[Technical field]

[0001] The present invention relates to a program for generating a description of a drawing and a device for generating a description of a drawing. [Background technology]

[0002] With the recent development of AI technology, it has been proposed to support the generation of various contents with the assistance of generative AI. For example, Patent Document 1 describes an invention that suppresses the variation in quality when creating claims, or reduces the time required to create a patent application manuscript including claims. In this way, many technologies are being implemented in which humans create text with the assistance of generative AI.

[0003] Patent Document 2 describes a method for supporting the creation of an explanation of an invention, which reduces the effort required for the creation of the explanation. Patent Document 2 defines a diagram representing the invention, and generates the explanation from a text-format definition using a predetermined syntax. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-095653 A [Patent Document 2] Patent Application No. 2024-017234 Summary of the Invention [Problem to be solved by the invention]

[0005] In the invention described in Patent Document 2, the description of a drawing is generated not from the drawing itself but from a text-format definition of the drawing. However, when actually creating the description, the drawing itself is often provided. In that case, a person must manually write a text-format definition of the drawing from the drawing, which is not much labor-saving compared to when a person writes the description of the drawing.

[0006] In addition, generative AI such as ChatGPT (registered trademark) by OpenAI and Gemini by GOOGLE can accept an image and provide prompts related to the image, and in the case of a simple flowchart, it is possible to generate natural language explanations of the image to a certain extent. However, in flowcharts that include complex branches and loops, the explanations could not be generated well, and explanations of the branches and loops were often omitted. This is thought to be due to the ambiguity inherent in natural language. Therefore, an object of the present invention is to generate an explanatory text relating to a drawing from the drawing. [Means for solving the problem]

[0007] That is, the above-mentioned object of the present invention is achieved by the following configuration. The drawing description generation program of the present invention causes a computer to perform the steps of inputting a first prompt to a generation AI, the first prompt including instructions to convert a drawing into text in a specified Markdown notation, and obtaining a first answer from the generation AI including text written in the specified Markdown notation, extracting the text written in the specified Markdown notation from the first answer, and inputting a second prompt to the generation AI, the second prompt including instructions to generate a description of the drawing, and obtaining a second answer from the generation AI including the description of the drawing.

[0008] The drawing description generation program of the present invention causes a computer to execute the steps of inputting a prompt to a generation AI, the prompt including instructions to convert a drawing and text in a specified Markdown notation, and instructions to generate the text and a description of the drawing, to obtain an answer including text written in the specified Markdown notation and a description of the text, extracting the text written in the specified Markdown notation from the answer, and extracting the description of the drawing from the answer. The other configurations will be explained in each embodiment. Effect of the Invention

[0009] According to the present invention, it is possible to generate an explanatory text relating to a drawing from the drawing. [Brief description of the drawings]

[0010] [Figure 1] 1 is a logical configuration diagram showing an explanation generation device according to a first embodiment. [Diagram 2] FIG. 2 is a hardware configuration diagram illustrating the explanatory text generation device. [Diagram 3] 1 is a screen displayed by the description generation device. [Figure 4] 13 is a flowchart of an explanation generation process. [Diagram 5] A diagram showing a request body to be sent to the generation AI system. [Figure 6] FIG. 11 is a diagram illustrating the content of a first prompt. [Figure 7] A diagram showing a chat completion object received from the generative AI system. [Figure 8] FIG. 13 is a diagram illustrating the content of a reply included in a chat completion object. [Figure 9] A diagram showing JSON data of a request body to be sent to the generation AI system. [Figure 10] FIG. 13 is a diagram for explaining the content of a second prompt. [Figure 11] A diagram showing JSON data of a chat completion object received from the generation AI system. [Figure 12] FIG. 13 is a diagram illustrating an example of an explanatory text. [Figure 13] FIG. 11 is a logical configuration diagram showing an explanation generation device according to a second embodiment. [Figure 14] 13 is a flowchart of an explanation generation process. [Figure 15] A diagram showing a request body to be sent to the generation AI system. [Figure 16] FIG. 13 is a diagram illustrating the contents of a prompt. [Figure 17] A diagram showing a chat completion object received from the generative AI system. [Figure 18] FIG. 13 is a diagram illustrating the content of a reply included in a chat completion object. [Figure 19] 1 is an example of a mode transition diagram. [Figure 20] This is Mermaid text generated from a mode transition diagram. [Figure 21] 1 is an example of a sequence diagram. [Figure 22] This is Mermaid notation text generated from a sequence diagram. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. FIG. 1 is a logical configuration diagram showing an explanation generation device 1 according to the first embodiment. The explanatory text generation device 1 includes a Markdown notation conversion unit 11, a Markdown notation extraction unit 12, an explanatory text conversion unit 13, and an explanatory text extraction unit 14. The explanatory text generation device 1 cooperates with a generation AI system 3 to generate an explanatory text 25 from a drawing 21.

[0012] When the drawing 21 and drawing type 22 are input, the Markdown notation conversion unit 11 instructs the generative AI system 3 to convert the drawing 21 into Mermaid notation text 23 by a request body 51 including a first prompt. Mermaid notation is a type of Markdown notation. Note that the notation of the text converted by the Markdown notation conversion unit 11 is not limited to Mermaid notation, and may be any Markdown notation that expresses various drawings in text.

[0013] Natural text describing a drawing is often ambiguous, with small nodes and topology missing. Therefore, when a generation AI learns a drawing and natural text describing the drawing, the natural text generated based on the drawing may also be ambiguous and may have missing information in the description. In contrast, text in Mermaid notation accurately indicates the nodes and topology of the drawing. Furthermore, the method of converting Mermaid notation text into natural text is a simple conversion. Therefore, the description generation device 1 can generate a more accurate description of the drawing by first converting the drawing into Mermaid notation text and then converting the Mermaid notation text into natural text.

[0014] In addition, the explanatory text generation device 1 divides the process into one for generating a text in Mermaid notation from a drawing and another for generating an explanatory text from the Mermaid notation text. This reduces the number of tokens required by the AI ​​system 3 compared to executing two processes with a single prompt, increasing the likelihood that the processes will be executed properly.

[0015] The drawing 21 is, for example, a flow chart, a sequence diagram, a mode transition diagram, etc. The drawing type 22 is information explaining the type of the drawing 21, and is, for example, a character string such as "flow chart", "sequence diagram", or "mode transition diagram".

[0016] This request body 51 is written in JSON format. The request body 51 is executed by the generation AI system 3, for example, by calling an API (Application Programming Interface) of the generation AI system 3, and a chat completion object 52 is generated. Then, the generation AI system 3 transmits the chat completion object 52 in JSON format to the Markdown notation conversion unit 11. This chat completion object 52 corresponds to the first answer including the Mermaid notation text 23.

[0017] The Markdown notation extraction unit 12 extracts the Mermaid notation text 23 from the chat completion object 52. Specifically, when it finds a line in which three backquotes indicating the start of a code block and mermaid are written in Markdown notation, it extracts as the Mermaid notation text 23 the text thereafter up to the line in which three backquotes indicating the end of the code block are written.

[0018] When the Mermaid notation text 23, the subject of action 24, and the writing style 26 are input, the explanatory text conversion unit 13 instructs the generation AI system 3 to convert the Mermaid notation text 23 into explanatory text 25 by a request body 53 including a second prompt. For example, if the drawing 21 is a flowchart, the subject of action is the subject of the flowchart. If the drawing 21 is a mode transition diagram, the subject of action is the subject of the mode transition diagram. Note that if the drawing 21 is a sequence diagram, the subject of action is shown, so there is no need to specify the subject of action.

[0019] The writing style 26 is a plain style in which the end of a sentence is written in the "dearu-da" style, or a polite style in which the end of a sentence is written in the "desumasu" style. The explanatory text conversion unit 13 generates explanatory text in the specified writing style 26, and is therefore able to select either a plain style suitable for official documents, or a polite style suitable for user instructions, etc.

[0020] The request body 53 is written in JSON format. This request body 53 is executed by the generation AI system 3, for example by calling the API of the generation AI system 3, and a chat completion object 54 is generated. Then, the generation AI system 3 returns the chat completion object 54 in JSON format to the explanation conversion unit 13. This chat completion object 54 is a second answer including the explanation 25 of FIG. 21.

[0021] The explanatory text extraction unit 14 extracts the explanatory text 25 from the chat completion object 54. This allows the explanatory text generation device 1 to generate the explanatory text 25 that explains the drawing 21 from the drawing 21 and the drawing type 22.

[0022] FIG. 2 is a hardware configuration diagram showing the explanation generation device 1. As shown in FIG. The explanation generation device 1 includes a CPU (Central Processing Unit) 101, a RAM (Random Access Memory) 102, and a ROM (Read Only Memory) 103. The CPU 101 is a processing device that executes a computer program and controls the explanation generation device 1. The RAM 102 is a volatile readable and writable memory, and is used as a work area for the computer program by the CPU 101. The ROM 103 is a non-volatile readable-only memory, and stores, for example, a BIOS (Basic I / O System) and the like.

[0023] The explanatory text generation device 1 further includes a display unit 104, an input unit 105, a communication unit 106, and a storage unit 107. The display unit 104 is, for example, a liquid crystal display, and displays characters, figures, images, etc. The input unit 105 is, for example, a mouse, keyboard, or touch panel, and allows a user to input information. The communication unit 106 is, for example, a network interface card, and communicates information with, for example, the generation AI system 3.

[0024] The storage unit 107 is a large-capacity storage device such as a hard disk or a solid state drive (SSD). The storage unit 107 stores an explanation generation program 1071. When the CPU 101 executes the explanation generation program 1071, each of the functional units in FIG. 1 is realized.

[0025] FIG. 3 shows an explanation generation screen 41 that the explanation generation device 1 displays on the display unit 104. The description generation screen 41 includes a type combo box 411, an action subject text box 412, a drawing area 413, a Mermaid notation text area 414, a description area 415, a Mermaid notation text generation button 416, and a description generation button 417. This description generation screen 41 is a screen for generating a description of a drawing by inputting a drawing, the type of drawing, the action subject, and the style of the description.

[0026] The type combo box 411 is a combo box for selecting the type of drawing. The writing style combo box 418 is a combo box for selecting whether the writing style of the explanation is to be plain or polite.

[0027] The drawing area 413 is an area for displaying an input drawing. For example, a drawing is registered in the drawing area 413 by the user dragging and dropping a drawing file into the drawing area 413. The Mermaid notation text generation button 416 is a button for generating Mermaid notation text from the registered drawing.

[0028] The Mermaid notation text area 414 is an area for displaying Mermaid notation text generated from the drawing displayed in the drawing area 413. The action subject text box 412 is a text box for inputting the action subject of a flowchart or the like explained in the drawing. The Mermaid notation text area 414 displays text in an editable manner. This allows the Mermaid notation text generated by the generation AI system 3 to be manually corrected to obtain a correct explanation, even if there is an error in the Mermaid notation text generated by the generation AI system 3.

[0029] The description generation button 417 is a button for generating a description from the Mermaid notation text displayed in the Mermaid notation text area 414. The description area 415 is an area for displaying a description generated from the Mermaid notation text displayed in the Mermaid notation text area 414.

[0030] FIG. 4 is a flowchart of the explanation generation process. First, the Markdown notation conversion unit 11 accepts input of a drawing, its type, and the subject of the action (step S10). Then, the Markdown notation conversion unit 11 creates a prompt that specifies the drawing and the type of the drawing and converts it into Mermaid notation text (step S11). The Markdown notation conversion unit 11 sends a request body including the generated prompt to the generation AI system 3 via the API (step S12), and receives a chat completion object from the generation AI system 3 (step S13).

[0031] Next, the Markdown notation extraction unit 12 extracts Mermaid notation text from the chat completion object (step S14). Next, the explanatory text conversion unit 13 creates a prompt to convert the Mermaid notation text, the subject of the action, and the writing style into an explanatory text (step S15).The explanatory text conversion unit 13 then transmits a request body including the generated prompt to the generation AI system 3 via the API (step S16), and receives a chat completion object from the generation AI system 3 via the API (step S17).

[0032] Next, the caption extracting unit 14 extracts a caption from the chat completion object (step S18), and displays the extracted caption of the drawing on the caption generating screen 41 (step S19), and ends the processing of FIG.

[0033] FIG. 5 shows a request body 51 including a first prompt 31 to be sent to the generation AI system 3. As shown in FIG. This request body 51 includes a model item and a messages item. In the Model item, the ID of the generation AI model is specified after a semicolon. In the Messages item, a list of messages that make up the conversation so far is stored, and various message types such as text, image, and audio can be set. In the Messages item, the role item and content item are written in curly brackets after a semicolon.

[0034] In the role item, enter system or user. In the contents item, enter text in the type item and enter the first prompt 31 in Fig. 6 in the text item. In the contents item, enter image_url in the type item and enter image data in BASE64 format or an image URL (Uniform Resource Locator) in the image_url item.

[0035] FIG. 6 is a diagram for explaining the contents of the first prompt 31. As shown in FIG. The first prompt 31 contains the instruction text "The attached drawing is {drawing type}. Please convert this drawing to text in Mermaid notation" and "{drawing type} = flowchart." This causes the generative AI system 3 to interpret the attached drawing as a flowchart and convert it into text.

[0036] FIG. 7 is a diagram showing the chat completion object 52 received from the generation AI system 3. As shown in FIG. The chat completion object 52 includes an id item, an object item, a created item, a model item, a choices item, and a usage item.

[0037] The id field is an identifier that uniquely identifies the chat completion object 52. The object field indicates the object type, which is always chat.completion. The created field is the Unix timestamp in seconds when the chat completion was created. The model field is the model used for the chat completion. The choices field is a list of options for the chat completion, which may be multiple. The usage field is usage statistics for the completion request.

[0038] The choices item is composed of an index item, a role item and a content item contained in a message item, and a finish_reason item. The index item is the number of the option for completing the chat. The role item included in the message item stores assistant. The content item included in the message item stores the answer 32 generated by the AI ​​generation system 3.

[0039] FIG. 8 is a diagram for explaining the content of the answer 32 contained in the chat completion object 52. As shown in FIG. Answer 32 contains Mermaid text within a Markdown code block. This Mermaid text is a description of the drawing.

[0040] Below is the Mermaid notation text of Figure 8. flowchart TD Start(["Start"]) S101{"Control decision?"} S102 ["Getting population density"] S103{"Overall control?"} S104 ["Determination of supply space"] S105 ["Control according to decision results"] S106 ["Independent Control"] Start --> S101 S101 -- "Y" --> S102 S101 -- "N" --> S106 S102 --> S103 S103 -- "Y" --> S104 S103 -- "N" --> S106 S104 --> S105 S105 --> S106

[0041] FIG. 9 shows a request body 53 including a second prompt 33 to be sent to the generation AI system 3. As shown in FIG. This request body 53 includes a model item and a messages item. In the Model item, the ID of the generation AI model is specified after a semicolon. In the Messages item, a list of messages that make up the conversation so far is stored, and various message types such as text, image, and audio can be set. In the Messages item, the role item and content item are written in curly brackets after a semicolon.

[0042] In the role item, enter system or user. In the contents item, enter text in the type item and enter the second prompt 33 in Fig. 10 in the text item.

[0043] FIG. 10 is a diagram for explaining the contents of the second prompt 33. As shown in FIG. In the second prompt 33, the instruction text on the first line reads, "The Markdown notation text is in Mermaid notation. Please generate an explanation of the [drawing type] indicated in this Markdown notation text with {actor} as the subject. Use {style} as the writing style." Then, on the second line, "{actor} = terminal" is written. On the third line, "{style} = normal" is written. On the fourth line, "#Markdown notation text" is written, indicating the Markdown notation text. From the fifth line onwards, the Mermaid notation text is written. As a result, the generation AI system 3 converts the Mermaid notation text from the fourth line onwards into an explanation with the action subject as the subject and the writing style specified.

[0044] FIG. 11 is a diagram showing the chat completion object 54 received from the generation AI system 3. As shown in FIG. The chat completion object 54 includes an id item, an object item, a created item, a model item, a choices item, and a usage item.

[0045] The id field is an identifier that uniquely identifies the chat completion object 54. The object field indicates the object type, which is always chat.completion. The created field is the Unix timestamp in seconds when the chat completion was created. The model field is the model used for the chat completion. The choices field is a list of options for the chat completion, which may be multiple. The usage field is usage statistics for the completion request.

[0046] The choices item is composed of an index item, a role item and a content item contained in a message item, and a finish_reason item. The index item is the number of the option for completing the chat. The role item included in the message item stores assistant. The content item included in the message item stores the answer 34 generated by the AI ​​generation system.

[0047] FIG. 12 is a diagram for explaining the content of the reply included in the chat completion object 54. As shown in FIG. The answer 34 includes an explanatory text. The explanatory text is a text that explains the drawing. The explanatory text actually generated by the AI ​​generation system 3 is transcribed below.

[0048] "The flowchart according to the present invention shows the procedure for determining whether or not control by the terminal is necessary and for executing appropriate control according to the situation. Each step is explained in detail below. First, the terminal starts processing (step "START"). Next, the terminal judges whether control is necessary (step S101). In this judgment, if it is determined that control is necessary, the terminal proceeds to a people density acquisition process (step S102). On the other hand, if it is determined that control is not necessary, the terminal directly executes independent control (step S106). In step S102, the terminal acquires information about the density of people in the vicinity. Next, the terminal judges whether or not overall control is necessary (step S103). If it is judged that overall control is necessary as a result of this judgment, the terminal proceeds to determining a supply space (step S104) and selects an appropriate supply space. On the other hand, if it is judged that overall control is not necessary, the terminal proceeds to independent control (step S106). After the supply space is determined, the terminal executes appropriate control based on the determination result (step S105). Finally, the terminal executes independent control (step S106) to complete the control flow. The flowchart of the present invention enables the terminal to appropriately determine the necessity of control according to the situation and the necessity of overall control, and to determine the supply space as necessary. This allows efficient control to be realized, and the terminal can perform the optimal operation suited to the surrounding situation."

[0049] FIG. 13 is a logical configuration diagram showing an explanation generation device 1A according to the second embodiment. The explanatory text generation device 1A includes a conversion unit 15, a Markdown notation extraction unit 12, and an explanatory text extraction unit 14. The explanatory text generation device 1A cooperates with a generation AI system 3 to generate an explanatory text 25 from a drawing 21.

[0050] When the conversion unit 15 receives the drawing 21, drawing type 22, action subject 24, and writing style 26, it instructs the generation AI system 3 to convert the drawing 21 into Mermaid notation text 23 and explanatory text 25 by a request body 55 including a prompt. The drawing 21 is, for example, a flow chart, a sequence diagram, a mode transition diagram, etc.

[0051] The drawing type 22 is information that explains the type of drawing 21, and is, for example, a character string such as "flowchart," "sequence diagram," or "mode transition diagram." The action subject 24 is, for example, the action subject of the flowchart if the drawing 21 is a flowchart. If the drawing 21 is a mode transition diagram, it is the action subject of the mode transition diagram. If the drawing 21 is a sequence diagram, the action subject is illustrated, so there is no need to specify the action subject. The writing style 26 is a plain style in which the sentence ends in the "desu / da" style, or a polite style in which the sentence ends in the "desu / masu" style. The conversion unit 15 can select either a plain style suitable for official documents or a polite style suitable for user instructions, etc., by generating an explanatory text in the specified writing style 26.

[0052] This request body 55 is written in JSON (JavaScript Object Notation) format. The request body 55 is executed by the generation AI system 3, for example, by calling the API of the generation AI system 3, and a chat completion object 56 is generated. Then, the generation AI system 3 returns the chat completion object 56 in JSON format to the conversion unit 15. This chat completion object 56 includes the Mermaid notation text 23.

[0053] The Markdown notation extractor 12 extracts the Mermaid notation text 23 from the chat completion object 56 . The explanatory text extraction unit 14 extracts the explanatory text 25 from the chat completion object 56. This enables the explanatory text generation device 1 to generate the explanatory text 25 that explains the drawing 21 from the drawing 21 and the drawing type 22.

[0054] FIG. 14 is a flowchart of the explanation generation process. First, the conversion unit 15 accepts input of a drawing, its type, and the subject of the action (step S30). The conversion unit 15 then specifies the drawing, the type of the drawing, and the subject of the action, converts them into Mermaid notation text, and creates a prompt to convert the prompt into an explanatory text for the drawing (step S31). The conversion unit 15 sends a request body including the generated prompt to the generation AI system 3 via the API (step S32), and receives a chat completion object from the generation AI system 3 (step S33).

[0055] Next, the Markdown notation extraction unit 12 extracts Mermaid notation text from the chat completion object (step S34).

[0056] Next, the caption extracting unit 14 extracts a caption from the chat completion object (step S35), and displays the extracted caption of the drawing on the caption generating screen 41 (step S36), and then ends the processing of FIG.

[0057] FIG. 15 is a diagram showing a request body 55 to be sent to the generation AI system 3. As shown in FIG. This request body 55 includes a model item and a messages item. In the Model item, the ID of the generation AI model is specified after a semicolon. In the Messages item, a list of messages that make up the conversation so far is stored, and various message types such as text, image, and audio can be set. In the Messages item, the role item and content item are written in curly brackets after a semicolon.

[0058] In the role item, enter system or user. In the contents item, enter text in the type item, and enter prompt 35 in Fig. 16 in the text item. In the contents item, enter image_url in the type item, and enter image data in BASE64 format or an image URL (Uniform Resource Locator) in the image_url item.

[0059] FIG. 16 is a diagram for explaining the contents of the prompt 35. As shown in FIG. Prompt 35 includes the instruction text "The attached drawing is {drawing type}. Please convert this drawing to Mermaid text. Then, generate a description for the patent specification of {drawing type} shown in this Mermaid text, with {actor} as the subject. Please use {style}. Please include the Mermaid text and description in your answer.", along with "{drawing type} = flowchart", "{actor] = terminal", and "{style} = plain". This causes the generation AI system 3 to interpret the attached drawing as a flowchart and convert it to Mermaid text. Then, generate a description for the patent specification of the flowchart shown in Mermaid text, with the subject of the action being the terminal.

[0060] FIG. 17 is a diagram showing the chat completion object 56 received from the generation AI system 3. The chat completion object 56 includes an id item, an object item, a created item, a model item, a choices item, and a usage item.

[0061] The id field is an identifier that uniquely identifies the chat completion object 56. The object field indicates the object type, which is always chat.completion. The created field is the Unix timestamp in seconds when the chat completion was created. The model field is the model used for the chat completion. The choices field is a list of options for the chat completion, which may be multiple. The usage field is usage statistics for the completion request.

[0062] The choices item is composed of an index item, a role item and a content item contained in a message item, and a finish_reason item. The index field is the number of the option for completing the chat. The role field in the message field stores assistant. The content field in the message field stores the answer 36 generated by the AI ​​generation system.

[0063] FIG. 18 is a diagram for explaining the content of the answer 36 contained in the chat completion object 56. As shown in FIG. Answer 36 contains Mermaid text in a Markdown code block. This Mermaid text is a description of the drawing. After that, a description of the drawing is provided.

[0064] FIG. 19 is an example of a mode transition diagram. The device starts in battery operation mode. When the power is turned ON, the device transitions to the first main power operation mode. If an attendance error occurs while in the first main power operation mode, the device returns to battery operation mode. After a certain period of time has passed, the device transitions from the first main power operation mode to the second main power operation mode. If the power is turned OFF while in the second main power operation mode, the device returns to battery operation mode.

[0065] FIG. 20 shows Mermaid notation text generated from a mode transition diagram. The Mermaid text is transcribed below. The explanation in Figure 19 is an explanation that was automatically generated from the Mermaid text. stateDiagram [*] --> Battery operation Battery operation --> 1st main power operation: Power ON 1st main power source operation --> Battery operation: Attendance error 1st main power supply operation ---> 2nd main power supply operation: for a certain period Second main power supply operation --> Battery operation: Power OFF

[0066] FIG. 21 is an example of a sequence diagram. The sequence diagram according to the present invention shows a series of steps in which a server responds to a processing request from a user terminal and stores the processing results in a cache database. Each step is described in detail below.

[0067] First, the user terminal transmits a processing request to the server. The server, upon receiving the processing request, transmits the processing request to the cache database to check whether the corresponding processing result exists as a cache in the processing result cache database.

[0068] Next, if the processing result cache database does not have the corresponding cache, it returns a "cache not found" response to the server. The server that receives this response executes the processing according to the request and generates the processing result.

[0069] After the processing is completed, the server stores the generated processing results in a processing result cache database so that future similar requests can be responded to quickly using the cache.The server then sends the processing results back to the user terminal so that the user can check the results.

[0070] Finally, the results cache database stores records of results received from the server, allowing for faster response times to future requests and improved system efficiency. According to the present invention, by using a cache function, the processing time for similar requests can be reduced, and the performance of the entire system can be improved.

[0071] Figure 22 shows Mermaid notation text generated from a sequence diagram. The Mermaid text is transcribed below. The explanation in Figure 21 is an explanation that was automatically generated from the Mermaid text. sequenceDiagram Participant User terminal Participant Server Participant processing result cache DB User terminal ->> Server: Processing request Server ->> Processing result cache DB: Processing request Processing result cache DB -->> Server: Cache not found Server ->> Server: Process execution Server ->> Processing result cache DB: Processing result Server -->> User's terminal: Processing result Processing result cache DB ->> Processing result cache DB: Record

[0072] (Modification) The present invention is not limited to the above-described embodiment, and modifications can be made without departing from the spirit of the present invention, for example, the following (a) to (d).

[0073] (a) The text into which the generating AI converts the drawing is not limited to Mermaid text, but can be any Markdown notation that uniquely indicates the nodes and topology of the drawing. (b) The generative AI system is not limited to one that runs on a cloud environment and may also run in an on-premise environment. (c) The description of the drawings is not limited to a description for a patent specification, but may be a description for a software specification or a description for a user instruction manual. In that case, it is recommended to specify in the prompt, "Generate a description for a software specification" or "Generate a description for a user instruction manual." (d) The generation AI system that sends the first prompt and the generation AI system that sends the second prompt may be the same or different, and are not limited thereto. [Explanation of symbols]

[0074] 1. Description generator 1A Description Generator 101 CPU 102 RAM 103 ROM 104 Display section 105 Input section 106 Communications Department 107 Storage section 1071 Description Generation Program 11 Markdown conversion section 12 Markdown syntax extractor 13 Description conversion section 14 Description Extraction Unit 15 Conversion section 21 Drawings 22 Drawing Types 23 Mermaid text 24 Actor 25 Description 3. Generative AI system 31 First prompt 32 answers 33 Second prompt 34 answers 35 Prompt 36 answers 51 Request body 52 Chat Complete Object (1st Answer) 53 Request body 54 Chat Complete Object (2nd Answer) 55 Request body 56 Chat Completed Object 41 Description generation screen 411 Type Combo Box 412 Action subject text box 413 Drawing Area 414 Mermaid text area 415 Description Area 416 Mermaid text generation button 417 Description generation button

Claims

1. On the computer, inputting a first prompt including a drawing and an instruction to convert the drawing into text in a predetermined Markdown notation to a generating AI, and obtaining a first answer including the text written in the predetermined Markdown notation from the generating AI; extracting text written in the predetermined Markdown notation from the first answer; inputting a second prompt to a generating AI, the second prompt including instructions to generate the text and a description of the drawing, and obtaining a second response from the generating AI including the description of the drawing; A drawing description generating program for implementing the above.

2. The type of the diagram is one of a flowchart, a sequence diagram, and a mode transition diagram; The second prompt includes a subject of the action indicated by the type of drawing and an instruction to make the subject the subject of the description of the drawing.

2. A drawing description generating program according to claim 1.

3. the second prompt includes instructions for a style of the legend to the drawing; 2. A drawing description generating program according to claim 1.

4. the first prompt includes information indicating a type of the drawing; 2. A drawing description generating program according to claim 1.

5. On the computer, A step of inputting a prompt including an instruction to convert a drawing and text in a predetermined Markdown notation and an instruction to generate the text and a description of the drawing to a generating AI, and obtaining an answer including the text written in the predetermined Markdown notation and the description of the text; extracting text written in the predetermined Markdown notation from the response; extracting the drawing legend from the response; A drawing description generating program for implementing the above.

6. The type of the diagram is one of a flowchart, a sequence diagram, and a mode transition diagram; The prompt includes a subject of the action indicated by the type of drawing and an instruction to make the subject the subject of the description of the drawing.

6. A drawing caption generating program according to claim 5.

7. the prompt includes information indicative of the type of drawing; 6. A drawing caption generating program according to claim 5.

8. a Markdown notation conversion unit that inputs a first prompt including a drawing and an instruction to convert the drawing into a predetermined Markdown notation to a generating AI and obtains a first answer including text written in the predetermined Markdown notation from the generating AI; a Markdown notation extraction unit that extracts text written in the predetermined Markdown notation from the first answer; an explanatory text conversion unit that inputs a second prompt including an instruction to generate an explanatory text for the text and the drawing to a generation AI and obtains a second answer including the explanatory text for the drawing from the generation AI; A drawing description generating device comprising:

9. A conversion unit that inputs a prompt including an instruction to convert a drawing and text in a predetermined Markdown notation and an instruction to generate an explanation of the text and the drawing to a generation AI, and obtains an answer including the text written in the predetermined Markdown notation and the explanation of the text; a Markdown notation extraction unit that extracts text written in the predetermined Markdown notation from the answer; an explanatory text extraction unit that extracts explanatory text for the drawing from the response; A drawing description generating device for implementing the above.

Images