What this category of data includes
This category of data primarily comes from China’s Disease Prevention and Control Information System, hospital HIS systems and infectious disease reporting systems, and laboratory testing systems. The update frequency is daily. Documents processed daily include infectious disease report cards, epidemic analysis reports, case correction records, clustered outbreak reports, and epidemiological investigation reports. Core fields cover case diagnosis information, reporting time limits, correction flags, clustered event association information, and similar items. Data must be processed strictly in accordance with legal infectious disease reporting regulations, with no additional custom statistical indicators.
What constraints these characteristics impose on deployment and upgrades
The daily update nature of this category of data requires a stable incremental synchronization mechanism to be configured during deployment, to avoid excessive system resource usage from full synchronization. The multi-type document structure, including structured report cards and unstructured epidemiological investigation reports, requires the knowledge base to support multi-format parsing and chunk adaptation, to ensure core information is not truncated. Legal reporting regulations require the data processing chain to be immutable. A full backup of business data must be completed before upgrade operations. For intranet deployment scenarios, dependency loading and version upgrade processes for environments without external network access must be adapted in advance, to avoid deployment failures caused by missing dependencies. Errors in core data directly affect public health response, so a complete rollback mechanism must be available during upgrade operations.
How to configure settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | Top 20 entries | Matches the daily volume of new case reports and monitoring data, to avoid exhausting the database connection pool |
REDIS_URL | Local intranet Redis service address, port 6379 | Supports retrieval caching and session storage, adapts to the network environment of intranet deployment |
FILE_TOKEN_KEY | Set based on actual testing | Adapts parsing rules for medical forms and report documents in this category, to ensure complete extraction of fields |
STORAGE_S3_ENDPOINT | Intranet object storage service address | Stores daily updated epidemic reports and historical case data, to ensure intranet data security |
| Chunk length | 4000–5000 tokens | Matches the average length of single documents in this category, to avoid truncation of core diagnosis and report information |
| Single search limit | Top 10 entries | Controls the context length of retrieval results, adapting to the accuracy requirements of business process guidance |
Three common configuration mistakes
- Phenomenon: The configuration reference limit is set to 1500, but retrieval results still exceed this length. Cause: The knowledge base chunk length and reference limit are not configured in linkage, causing chunked content to directly exceed the limit.
- Phenomenon: Version upgrade fails during intranet deployment, with a missing dependency prompt. Cause: Offline dependency packages are not prepared in advance, and the upgrade process relies on external network resource pulling.
- Phenomenon: After uploading an XLSX-format infectious disease report card, table fields cannot be parsed correctly. Cause:
FILE_TOKEN_KEYis not configured to adapt to parsing rules for medical structured forms, resulting in lost fields.
How to confirm the configuration is complete
- Log in to the FastGPT backend, view the database connection pool monitoring panel, confirm that the
DB_MAX_LINKconfiguration is in effect, and the connection count does not show a continuous peak. - Upload a local infectious disease report card XLSX file, verify that table fields can be parsed correctly and valid chunks are generated.
- Perform an incremental synchronization test, confirm that daily updated epidemic data can be normally imported into the knowledge base and retrieved.
- Trigger a version upgrade rollback operation, verify that the backed-up image and business data can normally restore the system configuration.
The values provided are common starting points and should be measured against the user'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.