This configuration applies to FastGPT v4.17.0.
What Data for This Category Looks Like
Documents used for heat treatment-related bids and quotations come from material grade databases, customer technical specification documents, temperature control instrument export data, furnace temperature uniformity test reports, hardness tester and metallographic equipment test results, and workshop MES systems. Document types include heat treatment process cards, furnace loading layout drawings, furnace temperature curve records, hardness and metallographic analysis reports, equipment inspection forms, and exception handling sheets.
Data updates per furnace charge or batch. Fields include material grade, heating temperature, holding duration, cooling method, hardness value, and metallographic structure grade. Units involve degrees Celsius, hours, Vickers hardness grades, and similar units.
What Constraints Do These Characteristics Impose on the Document Parsing and Chunking Link
This category has diverse document types with large structural differences: plain text process cards, PDF furnace temperature curves with embedded charts, and structured furnace loading diagram tables. The parsing logic must adapt to mixed content across different formats.
The per-furnace charge update requirement means chunking must retain contextual association for individual furnace charges, avoiding incorrect splitting of cross-furnace charge data. Bid documents require precise matching with customer technical specifications, so parsed chunks must retain field units and original meanings without losing key process parameters. Additionally, frequently updated single documents require controlled parsing time to avoid impacting batch upload efficiency.
How to Set the Configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Adapts to single-paragraph process descriptions and single-furnace charge data block lengths in heat treatment documents, avoiding splitting of cross-furnace charge associated information |
chunkSplitter | Split by paragraph + table boundaries | Adapts to structured content in process cards and furnace loading diagrams, retaining inline position and data association for table rows |
customPdfParse | Enabled | Optimizes extraction accuracy of text associated with embedded charts for PDF documents like furnace temperature curves and metallographic reports, which mix charts and text |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Covers parsing time for large single furnace temperature curve PDF documents, preventing parsing processes from being interrupted by timeouts |
chunkSettingMode | Manual mode | Allows adjustment of chunking rules for different document types such as process cards, furnace temperature curves, and furnace loading diagrams to adapt to multi-document combinations for bid documents |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Accommodates large furnace temperature curve PDF documents summarizing multiple furnace charges, preventing parsing from being rejected due to exceeding file size limits |
Three Common Mistakes
- Phenomenon: Uploading heat treatment-related HTML process documents after changing their file extension to .txt results in empty parsing results. Reason: Document format recognition relies on the original file header and extension; changing the extension does not trigger the corresponding parsing rules, leading to failure to extract structured content.
- Phenomenon: After using chunk mode to call the pushdata API for upload, the interface shows "indexing" for an extended period. Reason: The
PARSE_FILE_TIMEOUT_SECONDSconfiguration value is too low, or the single document volume exceeds theUPLOAD_FILE_MAX_SIZElimit, so the parsing process does not complete. - Phenomenon: Uploading a PDF document containing furnace temperature curve charts, and the parsing result does not include data associated with the charts. Reason: The
customPdfParseconfiguration is not enabled; the default parsing logic cannot extract associated text and coordinate information from embedded vector charts.
How to Confirm the Configuration Is Correctly Set
- Upload a single heat treatment process card PDF, check if the parsing result includes complete heating temperature and holding time fields, with no content truncation or cross-paragraph splitting.
- Upload a large PDF document containing furnace temperature curves, verify that the parsing time does not exceed the
PARSE_FILE_TIMEOUT_SECONDSconfiguration value, and no timeout errors occur. - Upload a copy of the heat treatment HTML report with its file extension changed to .txt, confirm whether the parsing result triggers a format recognition exception prompt, or that normal data extraction is possible after uploading with the original file extension.
- Switch
chunkSettingModeto manual mode, adjustchunkSize, then re-upload the same furnace loading diagram document, and verify that the chunk length meets expectations.
The values provided here are common starting points and should be evaluated against the reader's own samples.