Character and character extraction method in cad file AI translation

Abstract

The invention discloses a character and character extraction method in cad file AI translation, and relates to the technical field of AI translation. The method comprises the following steps: S1, character data receiving preparation: starting a server monitoring service, establishing a Socket communication link at a specified port, and waiting for client connection; s2, establishing client connection: receiving a connection request of a client, and establishing an independent processing thread for each client connection; through a multi-thread server architecture and an encryption transmission mechanism, a high-concurrency, safe and reliable data receiving basis is constructed, text character classification, hierarchical analysis and structured storage are adopted, efficient organization and rapid access of text data are achieved, and through deep linkage of an extraction process and a user instruction, the extraction efficiency of the text data is improved. In the initial stage of extraction, the source language content needing to be reserved is accurately filtered, the defects that in the prior art, extraction efficiency is low, data management is disordered, and intelligent screening cannot be achieved are overcome, and the accuracy of translation content is guaranteed from the source.

Description

Technical Field

[0001] This invention belongs to the field of AI translation technology, and in particular relates to a method for extracting text and characters in AI translation of CAD files. Background Technology

[0002] With the continuous improvement and widespread application of AI large-scale model technology, direct translation of AutoCAD-related applications through AI large-scale models has become possible. CAD technology is widely used in various industries, covering multiple fields such as architectural design, transportation engineering, municipal engineering, chemical engineering design, mechanical design, and electronic information. AI automatic translation of CAD drawings has gradually replaced manual translation, and directly outputting translation results has become a core application technology.

[0003] Existing methods typically perform simple segmentation of the received raw text data stream, failing to establish a hierarchical parsing process. This results in the system's inability to effectively verify the compliance of data formats, accurately separate file identification information from the main text content, and further break down the main text content into independent, manageable text entry units. Simultaneously, the lack of a unified and efficient storage structure to organize these parsed text data and their index relationships makes subsequent querying, filtering, and access operations complex and slow.

[0004] Most importantly, regarding the accuracy and intelligence of the extraction process, while existing technologies can extract text from documents, they typically perform indiscriminate, one-time extraction of all text content. This crude extraction method fails to intelligently filter and select based on user intent, such as the need to retain specific blocks of the source language, from the initial extraction stage. The root cause lies in the fact that the text extraction process of existing technologies is an isolated, "blind" operation, failing to effectively link with the structured parsing information of the document, such as block relationships, and user-input control commands. This results in core content that needs to retain the source language, such as drawing frames, company logos, and project information, being incorrectly included in the translation set, severely compromising the standardization and integrity of the translated drawings.

[0005] To address this issue, we provide a method for extracting text and characters from AI translation of CAD files. Summary of the Invention

[0006] The purpose of this invention is to provide a method for extracting text and characters in AI translation of CAD files, which solves the problem that existing technologies indiscriminately extract all text in the AI ​​translation of CAD files and cannot perform intelligent extraction according to user instructions.

[0007] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution.

[0008] This invention provides a method for extracting text and characters in AI translation of CAD files, comprising the following steps: S1. Text data reception preparation: Start the server listening service, establish a Socket communication link on the specified port, and wait for the client to connect; S2. Establish client connection: Receive client connection requests, create an independent processing thread for each client connection, establish a two-way communication channel, perform two-way authentication when establishing a connection, and establish an encrypted transmission channel after successful authentication. S3. Receive the CAD file content sent by the client through the established communication channel, receive and read the CAD source file, review the file data, perform structured parsing and content extraction, and extract all text entities and all their associated graphic attributes. S4. Separate text data packets: Separate independent text data packets, separating text content from non-text content such as graphics, primitives, and blocks; S5. Parse text data structure: Parse the received text data packet. First, verify whether the data format meets the preset specification requirements. Then, separate the file identifier part and the text content part from the data packet. Split the text content part into independent text entries according to the preset delimiter. Organize all the split text entries into an ordered list. S6. Text character classification: The extracted text content from the CAD file is classified into Chinese characters, English characters, numbers, general symbols, and special characters, which serve as the basis for data parsing. S7. Categorization and analysis: English words, English symbols, numbers, and general symbols are treated as non-translation work content and stored directly as ordinary markers. S8. The language to be translated, such as Chinese to English, is extracted, preprocessed and stored separately to prepare for translation parsing; S9. Specific content marking: A certain part of the content in the language to be translated needs to be excluded, while the source language is retained. For example, company map frame, specific block, etc. After the user specifies the block name, the text in the block is extracted, marked in a specific way, and pre-stored. The text content in the specified block is then translated and excluded. S10. Special characters, such as ㎡, m 3 It parses and encodes special characters such as Φ and ¥, encodes and stores special characters from different industries, and performs preprocessing before storage. S11. Establish a text data storage structure: Organize the text data, after classification, parsing, preprocessing and encoding, into structured data objects and store them with the file identifier as the key; S12. Provide a text data access interface: Establish an external access interface to support querying the corresponding structured text data object by file identifier.

[0009] The present invention is further configured to include a text filtering processing step, which involves receiving the name of the image block to be filtered input by the user, excluding text entries associated with the image block to be filtered from the global text dictionary based on the user-input image block name, and extracting only the remaining text entries to form a set of text to be translated.

[0010] The present invention is further configured such that text character classification and parsing processing is performed on the client side, and a structured extraction data packet is formed. The extraction data packet internally partitions and stores the following contents: original text entries, character category label sequence, separated character sequence to be translated, and encoding and position mapping table of special characters. This data packet is the basic unit of the text data packet sent by the client to the server.

[0011] The present invention is further configured such that the text character classification and parsing processing supports refined extraction of mixed text entries. For text entries containing multiple types of characters, the system performs context association analysis to identify and separate character combinations that logically belong to the same semantic unit. Among them, combinations of numbers with adjacent English characters or special symbols are identified as units of measurement or parameters, and combinations of Chinese characters with adjacent numbers or symbols are identified as modified terms. Sub-entry associations are established for these combinations and encapsulated together in the extraction data packet.

[0012] The invention is further configured to include a text translation processing step, in which the filtered set of texts to be translated is input into a preset large model translation engine, the translation engine performs translation processing on the set of texts, and generates a translated set of texts, supporting the selection and configuration of multiple translation engines.

[0013] The present invention is further configured such that the large model translation engine includes multiple preset translation engines, and the user selects any engine for translation. The translation engine supports text translation in industries and professions such as architecture, transportation engineering, municipal engineering, chemical design, mechanical design, and electronic information.

[0014] The present invention is further configured to include a text extraction quality control step, wherein after extraction, the text data is subjected to quality checks to detect duplicate text entries, empty text entries, or text entries with abnormal formatting, and a text data preview function is provided to support viewing and editing text content, support merging, splitting, and reordering of text entries, and provide text extraction statistics.

[0015] The present invention is further configured to include a text extraction caching mechanism, which caches the parsing results and storage structure of processed text data. When the same file identifier appears again, the text data is retrieved from the cache first. The cached data is set with an expiration period, and is automatically re-parsed and stored after expiration. Users can manually clear the cache or configure a cache management strategy.

[0016] The invention is further configured to include a text extraction template function, which allows users to create and save commonly used text extraction configurations as templates. The templates include preset delimiter settings, storage structure configurations, and caching strategy information. When processing similar text data in the future, the template configurations can be directly applied. The invention also supports template import, export, and version management functions.

[0017] The present invention is further configured to include exception handling and log recording. During the text extraction process, if a data format error, connection error, or processing failure is detected, an exception log is automatically generated, a prompt message is sent to the user explaining the exception and processing suggestions, and the unfinished processing results are retained for the user to choose to continue processing or restart.

[0018] The present invention has the following beneficial effects.

[0019] This invention constructs a high-concurrency, secure, and reliable data receiving foundation through a multi-threaded server architecture and encrypted transmission mechanism. It adopts text character classification, hierarchical parsing, and structured storage to achieve efficient organization and rapid access to text data. Furthermore, through deep linkage between the extraction process and user commands, it accurately filters out the source language content that needs to be retained in the early stages of extraction, solving the shortcomings of existing technologies such as low extraction efficiency, chaotic data management, and inability to intelligently filter, thus ensuring the accuracy of the translated content from the source. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0021] Figure 1 This is a flowchart of a method for extracting text and characters in AI translation of CAD files.

[0022] Figure 2 This is an example image before translation in Implementation 1 of a method for extracting text and characters in AI translation of CAD files.

[0023] Figure 3 This is an example image of the translated text and character extraction method in AI translation of CAD files, as shown in Implementation Example 1.

[0024] Figure 4 This is an example image before translation in Embodiment 2 of a method for extracting text and characters in AI translation of CAD files.

[0025] Figure 5 This is an example image of the translated text and character extraction method in CAD file AI translation, as shown in Example 2.

[0026] Figure 6 This is an example image before translation in a method for extracting text and characters in AI translation of CAD files.

[0027] Figure 7 This is an example image of the translated text and character extraction method used in AI translation of CAD files.

[0028] Figure 8 This is an example image before translation in Embodiment 3 of a method for extracting text and characters in AI translation of CAD files.

[0029] Figure 9 This is an example image of the translated text and character extraction method in CAD file AI translation, as shown in Example 3.

[0030] Figure 10 This is an example image before translation in Example 4 of a method for extracting text and characters in AI translation of CAD files.

[0031] Figure 11 This is an example image of the translated text and character extraction method in AI translation of CAD files, shown in Example 4. Detailed Implementation

[0032] The technical solutions of the present invention will be described below with reference to the accompanying drawings. The described embodiments are only some embodiments of the present invention, and not all embodiments.

[0033] Please see Figure 1-11 This invention provides a method for extracting text and characters in AI translation of CAD files, comprising the following steps: S1. Text data reception preparation: Start the server listening service, establish a Socket communication link on the specified port, and wait for the client to connect; S2. Establish client connection: Receive client connection requests, create an independent processing thread for each client connection, establish a two-way communication channel, perform two-way authentication when establishing a connection, and establish an encrypted transmission channel after successful authentication. S3. Receive the CAD file content sent by the client through the established communication channel, receive and read the CAD source file, review the file data, perform structured parsing and content extraction, and extract all text entities and all their associated graphic attributes. S4. Separate text data packets: Separate independent text data packets, separating text content from non-text content such as graphics, primitives, and blocks; S5. Parse text data structure: Parse the received text data packet. First, verify whether the data format meets the preset specification requirements. Then, separate the file identifier part and the text content part from the data packet. Split the text content part into independent text entries according to the preset delimiter. Organize all the split text entries into an ordered list. S6. Text character classification: The extracted text content from the CAD file is classified into Chinese characters, English characters, numbers, general symbols, and special characters, which serve as the basis for data parsing. S7. Categorization and analysis: English words, English symbols, numbers, and general symbols are treated as non-translation work content and stored directly as ordinary markers. S8. The language to be translated, such as Chinese to English, is extracted, preprocessed and stored separately to prepare for translation parsing; S9. Specific content marking: A certain part of the content in the language to be translated needs to be excluded, while the source language is retained. For example, company map frame, specific block, etc. After the user specifies the block name, the text in the block is extracted, marked in a specific way, and pre-stored. The text content in the specified block is then translated and excluded. S10. Special characters, such as ㎡, m 3 It parses and encodes special characters such as Φ and ¥, encodes and stores special characters from different industries, and performs preprocessing before storage. S11. Establish a text data storage structure: Organize the text data, after classification, parsing, preprocessing and encoding, into structured data objects and store them with the file identifier as the key; S12. Provide a text data access interface: Establish an external access interface to support querying the corresponding structured text data object by file identifier.

[0034] It also includes a text filtering process, which receives the name of the map to be filtered from the user input, excludes text entries associated with the map to be filtered from the global text dictionary based on the name of the map to be filtered from the user input, and extracts only the remaining text entries to form a set of text to be translated.

[0035] The text character classification and parsing process is executed on the client side and forms a structured extraction data packet. The extraction data packet internally partitions and stores the following contents: original text entries, character category label sequence, separated character sequence to be translated, and encoding and position mapping table of special characters. This data packet is the basic unit of the text data packet sent by the client to the server.

[0036] The text character classification and parsing process supports refined extraction of mixed text entries. For text entries containing multiple types of characters, the system performs contextual association analysis to identify and separate character combinations that logically belong to the same semantic unit. Among them, combinations of numbers with adjacent English characters or special symbols are identified as units of measurement or parameters, and combinations of Chinese characters with adjacent numbers or symbols are identified as modified terms. Sub-entry associations are established for these combinations and encapsulated together in the extraction data packet.

[0037] It also includes a text translation processing step, in which the filtered set of texts to be translated is input into a preset large model translation engine. The translation engine translates the text set and generates a translated text set. It supports the selection and configuration of multiple translation engines.

[0038] The large model translation engine includes multiple preset translation engines. Users can select any engine for translation. The translation engines support text translation in industries and professions such as architecture, transportation engineering, municipal engineering, chemical design, mechanical design, and electronic information.

[0039] It also includes a text extraction quality control step, which performs a quality check on the text data after extraction, detects duplicate text entries, empty text entries, or text entries with abnormal formatting, provides a text data preview function, supports viewing and editing text content, supports merging, splitting, and reordering text entries, and provides text extraction statistics.

[0040] It also includes a text extraction caching mechanism, which caches the parsing results and storage structure of processed text data. When the same file identifier appears again, the text data is retrieved from the cache first. The cached data has an expiration period, and it is automatically re-parsed and stored after expiration. Users can manually clear the cache or configure cache management strategies.

[0041] It also includes a text extraction template function, which allows users to create and save commonly used text extraction configurations as templates. The templates include preset delimiter settings, storage structure configurations, and caching strategy information. When processing similar text data later, the template configurations can be directly applied. It also supports template import, export, and version management functions.

[0042] It also includes exception handling and logging. If data format errors, connection errors, or processing failures are detected during text extraction, an exception log is automatically generated, and a prompt message is sent to the user explaining the exception and handling suggestions. The unfinished processing results are retained for the user to choose to continue processing or start over.

[0043] Example 1 Please see Figure 2-3This embodiment is applied to the Purple Triangle CAD drawing AI artificial intelligence translation network to extract text from CAD files of a basement air conditioning water renovation project.

[0044] The Purple Triangle platform's server first starts a multi-threaded listening service on a preset communication port, establishing a stable Socket communication link and preparing to receive connection requests from various clients. When a user client initiates a connection request, the server immediately allocates an independent processing thread for that connection and establishes an encrypted transmission channel with strict two-way authentication to ensure the security of data transmission.

[0045] Internally, the system performs in-depth character classification and preprocessing on the raw text read from CAD drawings. Specifically, the system accurately identifies and categorizes the characters in each text entry into Chinese characters, English characters, numeric characters, general symbols, and industry-specific characters. For example, the entry "fire compartment diagram" is categorized as a pure Chinese character sequence, while the entry "339.01m" is identified as a mixture of numeric and English characters. Based on this, the system parses and reorganizes the categorized characters: Chinese character sequences are separated and stored independently as the language content to be translated; mixed entries such as "339.01m" are identified as a single semantic unit representing length through contextual analysis; and special characters such as "㎡" and "Φ" that may appear are identified and uniformly encoded.

[0046] After the above preprocessing is completed, the client sends the data packet containing the file identifier and the structured extraction data generated according to the above rules to the server through a secure channel. The server supports resuming interrupted downloads to ensure complete data reception, then verifies the format, separates the content, and finally organizes the data into a structured data object containing multi-level information, indexed by the file identifier, for system storage. The standardized data access interface provided by the system enables subsequent modules to accurately query all Chinese content to be translated or technical parameters containing encoded special characters, laying a solid data foundation for high-fidelity translation.

[0047] Example 2 Please see Figure 4-5 This embodiment is applied to the translation of CAD design drawings for a building pressurized air supply room in the construction industry.

[0048] The system's background service first starts a Socket server with multi-threaded processing capabilities on a designated network port and enters a listening state. The dedicated client responsible for parsing the design files, after completing the initial analysis of the drawings, initiates a connection request to the server, and establishes an encrypted two-way communication channel after successful authentication.

[0049] Within this dedicated client, the core text extraction and preprocessing logic is executed. It analyzes the text in the drawings line by line, such as "name of pressurized air supply room" and "duct specification parameter DN500". The system first performs fine character classification, distinguishing between Chinese, English, numbers, and symbols. For the parameter "DN500", the system not only classifies the characters but also identifies "DN" as an industry-specific symbol representing nominal diameter and encodes it independently, while associating the number "500" with it as a whole semantic. All Chinese character sequences to be translated are extracted and cached separately.

[0050] The client packages the processed structured data with file identifiers and sends them in batches to the server via an encrypted channel. The server establishes an ordered queue to process these data packets sequentially. After verification and parsing, it finally stores the semantically rich structured data objects into the system using the file identifiers as keys. This allows the intelligent translation engine to quickly obtain pre-processed sets of Chinese text to be translated via standard interfaces, ensuring the structural integrity of professional parameters such as "DN500," thereby guaranteeing the professionalism and accuracy of the drawing translation.

[0051] Example 3 Please see Figure 8-9 This embodiment is implemented in the translation of CAD assembly drawings for emulsion pumps in the field of mechanical design.

[0052] The system server program keeps listening on a preset port using multiple threads, ready to respond to connection requests from dedicated preprocessing tool clients at any time.

[0053] This preprocessing tool performs in-depth local analysis of the emulsion pump assembly drawing, executing the key extraction steps of this method. Faced with technical parameters such as "nominal pressure 31.5 MPa" and "plunger diameter 50 mm," the tool first performs rigorous character classification, identifying Chinese characters, numbers, and industry-specific unit symbols such as "MPa" and "mm." Subsequently, the system performs intelligent analysis and encoding: "MPa" and "mm" are identified as special unit symbols and uniformly encoded and stored; through contextual analysis, the number "31.5" is closely associated with the encoded unit "MPa," forming a complete semantic unit for the pressure parameter; all Chinese descriptive text is extracted as a sequence to be translated.

[0054] After local preprocessing is complete, the client transmits the generated structured extraction data package along with the file identifier to the server. Once the server receives and stores this structured data, the system's text extraction quality control module can retrieve the data via an interface. This allows for not only basic checks such as duplicates and null values, but also in-depth validation based on character category tags. For example, it verifies whether all pressure-related entries contain the correct unit symbol codes, thus ensuring the high accuracy and completeness of the extracted content from the mechanical drawings from the source.

[0055] Example 4 Please see Figure 10-11 This embodiment demonstrates the specific implementation process of the text extraction method in the field of municipal environmental protection engineering, specifically for complex CAD drawings of odor collection and emission systems in sewage treatment plants.

[0056] The system's background service establishes a multi-threaded listening service on a designated port. A dedicated client, having completed the initial drawing analysis, establishes a secure data transmission channel with the server, encrypted and authenticated twice. On the client side, the system performs detailed extraction preprocessing. For example, for the entry "Design air volume 10000m³ / h..." 3 The system performs multi-level character classification, recognizing Chinese characters, numbers, and "m". 3 Special symbols such as " and " and "m 3 "This is a special character used in the industry and is specially encoded. The system further uses intelligent analysis to distinguish the number "10000" from the encoded "m". 3 The symbol, division sign, and unit "h" are associated as a unified semantic unit for air volume design parameters. Users input block names that need to retain the source language, such as "design institute drawing frame," into the front-end interface. During the preprocessing process, the client annotates the block information to which each text entry belongs and integrates it into the data package.

[0057] Subsequently, the client sends a data packet containing the file identifier and the aforementioned structured data to the server. The server supports resuming interrupted downloads to ensure complete reception. After verification, parsing, and storage, the user instruction processing module can obtain the data through a unified interface and quickly filter and exclude all text entries belonging to specific tiles based on the tile names provided by the user. This achieves the integration of user intent during the data extraction stage and enables precise isolation of the content that needs to be retained.

[0058] Example 5 This embodiment implements the method in the field of chemical engineering design, which emphasizes safety and compliance, with a particular focus on data verification and anomaly handling. For reactor flow diagrams, an audit server deployed in a secure environment initiates monitoring and employs advanced security protocols. A specially certified CAD security plugin acts as a client, requiring complex two-way authentication to establish a connection with the server.

[0059] After successful verification, the plugin performs rigorous extraction preprocessing locally. For key parameters such as "reaction temperature 80℃" and "working pressure 0.3MPa", the plugin performs precise character classification, identifies and encodes unit symbols such as "℃" and "MPa" as industry-specific characters, extracts all Chinese text as sequences to be translated, and records each step of the operation. Subsequently, the plugin transmits encrypted structured data packets containing file identifiers to the server via a dedicated line.

[0060] The server implements full monitoring and extremely rigorous data format and content compliance verification during the receiving process. Only after full verification is the structured data object parsed and stored, and all operations are automatically recorded in encrypted exception and audit logs. When any verification failure, parsing error, or encoding anomaly occurs, the system not only generates detailed logs and sends alerts, but also retains a snapshot of the problematic data for subsequent auditing, ensuring the entire process from source extraction is secure, reliable, and traceable.

[0061] The preferred embodiments of the present invention disclosed above are only for the purpose of illustrating the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation described herein. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can better understand and utilize the present invention.

Claims

1. A method for extracting text and characters in AI translation of a cad file, characterized in that: Includes the following steps: S1. Text data reception preparation: Start the server listening service, establish a Socket communication link on the specified port, and wait for the client to connect; S2. Establish client connection: Receive client connection requests, create an independent processing thread for each client connection, establish a two-way communication channel, perform two-way authentication when establishing a connection, and establish an encrypted transmission channel after successful authentication. S3. Receive the CAD file content sent by the client through the established communication channel, receive and read the CAD source file, review the file data, perform structured parsing and content extraction, and extract all text entities and all their associated graphic attributes. S4. Separate text data packets: Separate independent text data packets, separating text content from non-text content such as graphics, primitives, and blocks; S5. Parse text data structure: Parse the received text data packet. First, verify whether the data format meets the preset specification requirements. Then, separate the file identifier part and the text content part from the data packet. Split the text content part into independent text entries according to the preset delimiter. Organize all the split text entries into an ordered list. S6. Text character classification: The extracted text content from the CAD file is classified into Chinese characters, English characters, numbers, general symbols, and special characters, which serve as the basis for data parsing. S7. Categorization and analysis: English words, English symbols, numbers, and general symbols are treated as non-translation work content and stored directly as ordinary markers. S8. The language to be translated, such as Chinese to English, is extracted, preprocessed and stored separately to prepare for translation parsing; S9. Specific content marking: A certain part of the content in the language to be translated needs to be excluded, while the source language is retained. For example, company map frame, specific block, etc. After the user specifies the block name, the text in the block is extracted, marked in a specific way, and pre-stored. The text content in the specified block is then translated and excluded. S10, special characters such as m2, m 3 , Φ, ¥, etc. Analyze special characters and encode them. Encode and store special characters for different industries and perform preprocessing storage. S11. Establish a text data storage structure: Organize the text data, after classification, parsing, preprocessing and encoding, into structured data objects and store them with the file identifier as the key; S12. Provide a text data access interface: Establish an external access interface to support querying the corresponding structured text data object by file identifier.

2. The method of claim 1, wherein the method further comprises: determining whether the text and characters are in a text box or a character box; and extracting the text and characters from the text box or the character box. It also includes a text filtering process, which receives the name of the map to be filtered from the user input, excludes text entries associated with the map to be filtered from the global text dictionary based on the name of the map to be filtered from the user input, and extracts only the remaining text entries to form a set of text to be translated.

3. The method of claim 1, wherein the method further comprises: determining whether the text and characters are in a font of the CAD file; and if the text and characters are in the font of the CAD file, extracting the text and characters from the CAD file. The text character classification and parsing process is executed on the client side and forms a structured extraction data packet. The extraction data packet internally partitions and stores the following contents: original text entries, character category label sequence, separated character sequence to be translated, and encoding and position mapping table of special characters. This data packet is the basic unit of the text data packet sent by the client to the server.

4. The method of claim 2, wherein the method further comprises: determining whether the text and characters are in the same language; and if the text and characters are in the same language, extracting the text and characters in the same language. The text character classification and parsing process supports refined extraction of mixed text entries. For text entries containing multiple types of characters, the system performs contextual association analysis to identify and separate character combinations that logically belong to the same semantic unit. Among them, combinations of numbers with adjacent English characters or special symbols are identified as units of measurement or parameters, and combinations of Chinese characters with adjacent numbers or symbols are identified as modified terms. Sub-entry associations are established for these combinations and encapsulated together in the extraction data packet.

5. The method for extracting text and characters in AI translation of CAD files according to claim 1, characterized in that: It also includes a text translation processing step, in which the filtered set of texts to be translated is input into a preset large model translation engine. The translation engine translates the text set and generates a translated text set. It supports the selection and configuration of multiple translation engines.

6. The method for extracting text and characters in AI translation of CAD files according to claim 5, characterized in that: The large model translation engine includes multiple preset translation engines. Users can select any engine for translation. The translation engine supports text translation in industries and professions such as architecture, transportation engineering, municipal engineering, chemical design, mechanical design, and electronic information.

7. The method for extracting text and characters in AI translation of CAD files according to claim 1, characterized in that: It also includes a text extraction quality control step, which performs a quality check on the text data after extraction, detects duplicate text entries, empty text entries, or text entries with abnormal formatting, provides a text data preview function, supports viewing and editing text content, supports merging, splitting, and reordering text entries, and provides text extraction statistics.

8. The method for extracting text and characters in AI translation of CAD files according to claim 1, characterized in that: It also includes a text extraction caching mechanism, which caches the parsing results and storage structure of processed text data. When the same file identifier appears again, the text data is retrieved from the cache first. The cached data has an expiration period, and it is automatically re-parsed and stored after expiration. Users can manually clear the cache or configure cache management strategies.

9. The method for extracting text and characters in AI translation of CAD files according to claim 1, characterized in that: It also includes a text extraction template function, which allows users to create and save commonly used text extraction configurations as templates. The templates include preset delimiter settings, storage structure configurations, and caching strategy information. When processing similar text data later, the template configurations can be directly applied. It also supports template import, export, and version management functions.

10. The method for extracting text and characters in AI translation of CAD files according to claim 1, characterized in that: It also includes exception handling and logging. If data format errors, connection errors, or processing failures are detected during text extraction, an exception log is automatically generated, and a prompt message is sent to the user explaining the exception and handling suggestions. The unfinished processing results are retained for the user to choose to continue processing or start over.

Images