What this category’s data includes
Data sources for this category include China Government Procurement Network, local government procurement websites, fiscal integration systems, procurer budget systems, and supplier credit records. The update frequency is daily. Daily processed documents cover procurement requirements, procurement plans, bidding documents, tender documents, bid evaluation reports, winning bid announcements, procurement contracts, responses to objections, and other materials. This category involves entities including procurer units, the government procurement supervision department of the finance department, procurement agencies, and evaluation experts. Key fields include procurement numbers, supplier credit ratings, bid evaluation results, and additional relevant fields.
Constraints on the document parsing and chunking link
The high-frequency daily update requirement demands stable and timely parsing processes, to avoid delaying procurement workflows due to parsing delays. There are significant structural differences across document types: some are long texts with hierarchical headings, while others are business forms with embedded multiple tables. Parsing tools must adapt to all these formats, to prevent damage to the integrity of key business logic such as procurement requirements and bid evaluation results during chunking. Document formats vary across source systems. Parsing tools must support structured files exported from fiscal integration systems and converted files publicly released on government procurement websites. It is also necessary to avoid parsing errors that could lead to subsequent bid invalidation, complaints, and other negative outcomes.
Configuration settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Ensures semantic integrity for long texts such as procurement requirements and bid evaluation reports in government procurement documents, avoiding split damage to business field associations |
chunkSplitter | Split by paragraph + heading | Government procurement documents have hierarchical heading structures such as procurement plans and winning bid announcements. Splitting by headings preserves business logic associations |
customPdfParse | Enabled | Adapts to multi-table PDF documents such as bidding documents and bid evaluation reports, preventing table data loss from default parsing |
PARSE_FILE_TIMEOUT_SECONDS | 120–180 seconds | Government procurement documents are mostly multi-page long files, reserving sufficient parsing time to avoid timeout failures |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Covers the upload requirements for complete tender documents and bid evaluation reports of large-scale procurement projects |
chunkSettingMode | Manual configuration | Adapts to chunking needs for different document types, allowing parameter adjustments for short texts such as procurement contracts and response to objections |
Three common configuration errors
- Symptom: PDF parsing returns an error, and logs show parsing timeout. Cause: The
PARSE_FILE_TIMEOUT_SECONDSparameter was not adjusted, and parsing time for long government procurement documents exceeded the default threshold. - Symptom: Uploaded bidding documents do not retain table structures, and the large model cannot identify quotation details in procurement requirements after embedding. Cause: The
customPdfParseconfiguration was not enabled, and default parsing rules caused table data loss. - Symptom: When calling the API interface for creating a file collection in the knowledge base, PDF parsing parameters cannot be specified. Cause: The
customPdfParsefield was not correctly configured in the API request, and the system default parsing logic was used.
How to confirm correct configuration
- Upload a standard bidding document PDF, and check if the parsed text retains table and hierarchical heading structures.
- Submit a bid evaluation report with more than 100 pages, and verify that the parsing task completes within 180 seconds with no timeout errors.
- Call the API interface for creating a file collection in the knowledge base, and confirm that the request parameters include the
customPdfParseconfiguration item. - View the chunked text fragments, and confirm that each chunk’s length falls within the preset range, without damaging key business fields such as procurement numbers and supplier names.
The values provided are common starting points and should be measured against the reader's own 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.