What the Data for This Category Looks Like
This category's data comes from the Chinese Center for Disease Control and Prevention Information System, medical institution HIS systems, infectious disease reporting systems, and laboratory testing systems. The update frequency is daily. Daily documents include infectious disease report cards, epidemic analysis reports, case correction records, clustered epidemic reports, and epidemiological investigation reports. Fields cover case identification, diagnostic information, reporting time limits, correction status, clustered event association information, and more. All data is structured and stored strictly in accordance with legal infectious disease reporting regulations, with no additional statistical processing applied.
What Constraints These Characteristics Impose on the Knowledge Base Retrieval and Recall Link
Daily updated data requires the knowledge base retrieval and recall link to support incremental synchronization mechanisms. This avoids repeated full data pulls that occupy system resources. Structured documents include multi-dimensional legal report fields. Retrieval processes must ensure that field association information is not split or broken. Otherwise, matching accuracy for business scenarios such as case correction and time limit verification will be affected. Clustered epidemic reports and epidemiological investigation reports have strong content relevance. Retrieval must prioritize recalling associated documents under the same event. It must also filter invalid old report cards that have been corrected, to ensure recall results are the latest valid data. Legal reporting regulations require recall results to strictly correspond to report fields. Irrelevant content must not interfere with business judgment, as this may cause business process errors.
Configuration Settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
similarity | 0.75–0.85 | Meets the accuracy requirements for matching legal report fields, prevents irrelevant documents with low similarity from being recalled |
limit | Top 8 | Covers associated information from multiple report cards and analysis reports within a single day, while controlling context length |
searchMode | keyword + vector | Balances keyword matching for report numbers and diagnostic information with vector matching for semantic content, adapting to structured document retrieval |
usingReRank | Enabled | Filters low-relevance content from preliminary recall results, improving matching accuracy for correction records and epidemiological investigation reports |
datasetSearchUsingExtensionQuery | Enabled | Automatically expands disease control professional terminology for queries related to party building learning, adapting to semantic understanding in professional scenarios |
collectionIds | Specify the exclusive knowledge base collection for this category | Limits the retrieval scope to avoid mixing irrelevant data from other categories |
Three Common Configuration Mistakes
- Phenomenon: Online conversations continue to search the knowledge base without response, with no valid return results. Cause: No reasonable
limitandsimilarityvalues are configured. The number of recalled documents exceeds the context carrying limit, causing request timeouts. - Phenomenon: Knowledge base IDs and chunk IDs in the knowledge base details page cannot be copied. Cause: The page interaction logic is not bound to a copy trigger event, or permission configuration does not allow front-end copy operations.
- Phenomenon: The professional enhanced retrieval function cannot be enabled normally, and semantic expansion matching cannot be triggered. Cause: The
datasetSearchUsingExtensionQueryconfiguration is not enabled, or the corresponding function permission is not open in the current environment. For personal self-built scenarios, confirm the function switch status.
How to Verify Successful Configuration
- Manually trigger knowledge base incremental synchronization, check that the update time in the synchronization log matches the daily update frequency of the data source.
- Enter a query containing a specific case number, verify that the recall results include the complete field information of the corresponding report card.
- After enabling
usingReRank, compare the number of recall results before and after enabling to confirm that redundant low-relevance content has been effectively filtered. - View the
collectionIdsconfiguration of the knowledge base, confirm that only the document collection for this category is included, with no cross-category data mixed in.
The values provided are common starting points and should be evaluated 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.