What the data for this category looks like
Data for this category primarily comes from public sector and state-owned enterprise procurement centers, tender offices, business requirement departments, and audit and discipline inspection departments. It covers documents including procurement plans, requirement specifications, tender documents, bid evaluation records, contracts, acceptance slips, and supplier complaint materials. Data sources include enterprise procurement platforms, public resource trading platforms, supplier databases, contract management systems, and audit issue databases, with a daily update frequency. Documents include both structured fields and unstructured text. Fields cover procurement methods, budget basis, scoring methods, supplier qualification requirements, and similar items. Units include currency amounts, quantities, scoring points, and other relevant metrics.
Constraints on the document parsing and chunking link
Data for this category mostly consists of mixed-format documents updated in batches daily. The parsing link must support batch file processing. Documents include both structured tables and unstructured text. Chunking must retain table integrity and contextual association, and avoid splitting key compliance fields such as procurement methods and scoring methods. Some documents come from external public resource platforms, and their formats may be non-standard. The system must adapt to non-standard layouts. Audit and discipline inspection-related documents require precise retention of associated information. Chunk length must cover complete compliance review logic units, to avoid omitting key content that affects compliance judgments.
Configuration Settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Covers complete compliance review units such as procurement methods and scoring methods, avoiding splitting key fields |
chunkSplitter | Split by paragraph + table | Adapts to the mixed structured and unstructured document characteristics of this category, retaining table integrity |
customPdfParse | Enabled | Adapts to non-standard PDF formats common in tender and procurement documents, such as multi-column layouts and embedded tables |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Addresses time consumption requirements for parsing large batches of files, avoiding parsing failures due to timeout |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Supports uploading large-capacity documents such as batch tender documents and bid evaluation records in a single operation |
chunkSettingMode | Manual configuration | Adjusts chunking parameters for different document types (such as tender documents, contracts) to fit compliance review scenarios |
Three Common Configuration Mistakes
- Issue: Uploaded PDF and DOCX files have empty fields after parsing, or target files cannot be read. Cause: Correct file storage path is not configured, or file synchronization permissions pushed by external systems are not enabled, causing the parsing link to fail to obtain files to be processed.
- Issue: Parsing tasks time out and fail, returning a timeout status code. Cause:
PARSE_FILE_TIMEOUT_SECONDSis not set to a value suitable for large batch files, or single parsed files exceed theUPLOAD_FILE_MAX_SIZElimit. - Issue: Chunking results lose associated fields within tables, such as scoring methods and corresponding scores not appearing in the same chunk. Cause: Chunking configuration only splits by character length, without optimizing for table structures, destroying the contextual integrity of tables.
How to Verify Correct Configuration
- Upload a single tender document PDF, check the parsed text structure, confirm that tables are not split and key compliance fields are complete.
- Upload batch documents of the same type, verify that all parsing tasks are completed within the duration set by
PARSE_FILE_TIMEOUT_SECONDS. - Configure external systems to push test files, confirm that parsed data can be normally retrieved and accessed in the knowledge base.
- Adjust the
chunkSizeparameter, verify whether chunking results cover complete compliance review logic units.
The values provided here are common starting points, and should be evaluated against the relevant samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.