What this category of data looks like
Data for this category comes from the China Disease Prevention and Control Information System, hospital HIS systems, infectious disease reporting systems, and laboratory testing systems. The update frequency is daily. Daily generated documents include infectious disease reporting cards, epidemic analysis reports, case correction records, clustered epidemic reports, and epidemiological investigation reports. Document structures include structured fields and free-text analysis content. Core fields include case unique ID, diagnosis time, reporting unit, test results, correction time, and others. Common units are case count, positive test count, and reporting time limit.
What constraints do these characteristics place on the citation source and traceability process
Decentralized storage of multi-source data requires the traceability process to link unique data identifiers across different systems. This prevents mixing of reporting content from different medical institutions. The daily update requirement means traceability data must use the latest corrected version. Outdated or unupdated old data cannot be cited. Statutory document formatting rules require citations to correspond to specific fields, not just general references to report content. The large volume of case correction records requires the traceability process to distinguish between different versions of report data. This ensures that only the latest valid content is cited.
How to set configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
quoteMaxToken | 800–1200 token | Matches the core text length of a single infectious disease reporting card or epidemic analysis report, ensuring complete citation of key information |
responseData | Return original source metadata | Links unique data identifiers from the China Disease Prevention and Control Information System and hospital HIS systems, meeting traceability verification requirements |
detail | Enable full traceability display | Shows specific fields of the cited document, such as case ID, reporting time limit, correction time, and complies with compliance requirements for disease prevention and control reports |
| Single search limit | Top 6 recall results | Covers decentralized data from multiple sources, avoiding missed key case correction records or monitoring signals |
| Maximum context | 15000 characters | Accommodates content from multiple associated reports, supporting cross-document traceability analysis for clustered epidemics |
Three common misconfigurations
- Setting
quoteMaxTokentoo small causes cited content to be truncated. Only fragments appear in the interface, preventing full verification of case reporting time limits. This occurs when the single-text length requirements of disease prevention and control documents are not matched. - Not enabling the
responseDataconfiguration results in no citation ID field in the conversation request API response. This makes tracing the original source of the data impossible. - Not configuring the citation disable rule for no matching results leads to forced citation of existing documents when querying an abnormal monitoring scenario not present in the knowledge base. This violates the rigor requirements of disease prevention and control reports.
How to confirm the configuration is correct
- Initiate a query that includes a specific case ID. Verify whether the returned result contains the metadata identifier of the corresponding document.
- Check the conversation API return fields for FastGPT v4.17.0. Confirm that citation-related metadata content is present.
- Test a query with no matching data. Confirm that no unfounded citations are generated in the result.
- Adjust the
quoteMaxTokenparameter. Verify whether the text length of the returned citation meets expected requirements.
Note: The configuration values provided are common starting points and should be evaluated against specific samples and use cases.
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.