What data for this category looks like
Typical subjects of this category are battery pack, energy storage cabinet and BMS system manufacturers. Daily documents include grouping rules, BOM lists, welding and torque records, BMS flashing records, insulation and voltage withstand reports, EOL test reports, and system configuration lists. Data sources cover cell capacity grading data, MES systems, welding equipment, torque tools, BMS software version records, and EOL test benches. Update frequency is once per produced pack or energy storage cabinet, with updates synchronized to BMS software version changes. Document fields include cell capacity, welding torque value, insulation resistance value, and BMS version number. Corresponding units are Ah, N·m, MΩ, and version number respectively. Some documents contain complex tables and hierarchical structures.
Constraints on document parsing and chunking
Multi-source heterogeneous document formats create parsing adaptation difficulties. CSV and PDF reports directly output by equipment have inconsistent formats, requiring parsing logic adapted to different structures. Frequently updated documents require the parsing process to support batch processing, to avoid parsing delays caused by fast update cycles. Specialized fields and units require retaining the association between parameters and units during chunking, to prevent semantic fragmentation after splitting. For documents with complex tables and hierarchical structures, chunking logic must avoid destroying valid cross-row and cross-column data and hierarchical relationships. This ensures subsequent Q&A can associate complete technical information.
How to configure settings
All configurations below are based on FastGPT v4.17.0.
| Configuration Item | Recommended Value Range | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Adapts to the mixed structure of this category’s documents, which include both short parameter fields and long technical paragraphs, to avoid splitting critical context |
chunkSplitter | By paragraph + semantic boundary | Adapts to documents with hierarchical logic such as BOMs and grouping rules, retaining natural associations between content |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Supports upload requirements for single large-capacity EOL test reports and batch cell capacity grading data files |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Addresses parsing time for complex PDF reports and multi-sheet tables, to avoid mid-process timeout failures |
customPdfParse | Enabled | Adapts to formatted technical parameter pages and complex tables common in this category’s documents, improving parsing accuracy |
Three common configuration mistakes
- A 413 error appears after file upload. The cause is failure to adjust the
UPLOAD_FILE_MAX_SIZEparameter to a value suitable for this category’s large-size documents. - Only the first two columns of CSV table data are parsed. The cause is failure to correctly configure the
chunkSplitter’s adaptation rules for table structures; the default chunking logic incorrectly truncates cross-column data. - The external pdf-maker parser cannot be called normally. The cause is failure to correctly set the associated key for the corresponding parser in system configurations, causing parsing requests to fail verification.
How to confirm correct configuration
- Upload a typical EOL test report PDF for this category. Confirm the parsed text fully retains table row and column structures and specialized unit information.
- Upload a cell capacity grading data CSV file with multiple columns. Verify all column data is fully parsed with no omissions or truncations.
- Adjust
chunkSizewithin the test range. Generate chunking results, and confirm long paragraphs of grouping rules are not forcibly split. - Upload a single document under 1000 MB. Confirm the parsing task does not trigger timeout or 413 errors.
The values provided are common starting points and should be measured against the reader's own samples.