What data looks like for this category
Data for the condiment category comes primarily from internal enterprise inventory ledgers, compliance reports issued by third-party testing institutions, category monitoring documents released by industry associations, and filing materials required by regulatory authorities. Data update cadences include monthly inventory updates, quarterly cost accounting reports, and annual industry analysis documents. Document formats include structured tables with fixed fields, long-form compliance descriptions, encrypted PDF test reports, and Excel inventory ledgers. Fields include batch number, production date, raw material purchase unit price, finished product ex-factory price, physical and chemical indicator values, and more. Units include grams per 100 grams, yuan, cases, and other granular measurement formats.
What constraints these characteristics impose on document parsing and chunking
The data characteristics of the condiment category impose multiple constraints on the document parsing and chunking process. Mixed document formats including tables, long text, and encrypted PDFs require the parsing process to support multi-format compatibility and encrypted content unlocking. Structured tables with fixed fields require precise matching of headers to corresponding values to avoid field misalignment and confusion. Numeric fields with multiple units must retain original measurement identifiers to prevent data bias from unit conversion. Long-form compliance descriptions contain a large number of specialized terms, so chunking should follow semantic paragraphs instead of fixed lengths to avoid breaking professional expressions. Frequently updated monthly documents also require adaptation to batch parsing speed requirements to prevent parsing timeouts.
How to set configurations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
pdf_marker_version | v2.1.0 | Supports encrypted PDF parsing and structured table output, compatible with complex formats of condiment test reports |
chunk_size | 800–1200 characters | Balances semantic completeness and chunk granularity for both long-form compliance text and structured tables in condiment documents |
max_paragraph_depth | 3 | Matches the majority of chapter and paragraph hierarchy levels within condiment due diligence reports, which are within three levels |
api_parse_timeout | 120 seconds | Covers the typical processing time for batch parsing multiple condiment inventory ledgers and test reports |
preserve_original_units | Enabled | Retains original measurement identifiers for condiment fields to avoid data bias caused by unit conversion |
enable_table_structured_parse | Enabled | Accurately extracts fields and corresponding values from structured tables in condiment documents |
The parameter values provided on this page are general recommendations to serve as a starting point for configuration. Actual values are affected by material format, data volume, and business rules. Specific issues require targeted analysis. It is recommended to test on your own samples before finalizing settings.
Three common mistakes
- Issue: The
pdf_marker_v2version cannot be selected in the interface to enable enhanced parsing, and a 404 error is returned when calling the/v2/parse/fileAPI. Cause: The deployed parsing plugin has not pulled the v2 branch code, or the API path is not configured correctly. - Issue: After passing chunking parameters via the API documentation, the length of the chunk results does not match the
chunk_sizesetting, and the returned chunk list does not include theindex_sizefield. Cause: Thechunk_sizeandindex_sizeparameters are not passed correctly in the API request, or the parameter unit uses characters instead of tokens as required. - Issue: In the parsed condiment inventory ledger documents, unit price fields are misaligned with corresponding batch numbers, and some fields are empty. Cause: The
enable_table_structured_parseconfiguration is not enabled, so the parsing engine fails to recognize the table header structure and cannot match fields to their corresponding values.
How to confirm configurations are set correctly
- Upload a standard condiment test report PDF, and check if the table fields in the parsing result retain the original units and values completely.
- Call the
/api/v1/parse/fileAPI, pass the preset configuration parameters, and check if the returned chunk list includes theparagraph_depthfield and that its value matches the set value. - Upload multiple condiment documents of different formats in batch, and check if all parsing task statuses show as completed with no timeout errors.
- Check the parsing logs to confirm that the
pdf_marker_versionparameter is correctly identified as the set version, with no version compatibility-related errors.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-14.