What Data Looks Like for This Category
This category uses data from the Chinese Center for Disease Control and Prevention Information System, hospital HIS systems, infectious disease reporting systems, and laboratory testing systems. Data updates occur daily. Daily documents include infectious disease report cards, epidemic analysis reports, case correction records, clustered epidemic reports, and epidemiological investigation reports. Core fields cover basic case information, clinical diagnosis results, reporting time limits, correction status, and laboratory test data. The data structure must comply with standardized format requirements for statutory infectious disease reporting.
What Constraints These Characteristics Impose on the Knowledge Base Retrieval and Recall Link
Daily data updates require the knowledge base to support incremental synchronization, which avoids performance losses from full reconstruction. Mixed retrieval of multiple document types requires adaptation to each document’s structural features, distinguishing structured fields in report cards from unstructured text in analysis reports. Strict statutory reporting rules require retrieved content to match field definitions and format standards, preventing non-compliant information. Underreporting or delayed reporting carries severe consequences, so the accuracy of retrieved results must be improved to reduce interference from irrelevant content.
How to Configure Settings
| Configuration Item | Recommended Value Range | Rationale |
|---|---|---|
similarity | 0.75–0.85 | Matches the field accuracy requirements of statutory reports, filters low-correlation results, and avoids mixing irrelevant case information |
limit | Top 10 entries | Balances the amount of retrieved information and context length, adapting to the concise retrieval needs of disease control documents |
searchMode | vector_search | Supports mixed retrieval of structured and unstructured documents, prioritizing semantically relevant report content |
usingReRank | Enabled | Improves the sorting priority of highly relevant results, reducing mismatches for correction records and report cards |
datasetSearchUsingExtensionQuery | Enabled | Automatically expands retrieval keywords, covering synonyms of statutory fields such as case diagnosis and reporting time limits |
collectionIds | Bind the corresponding disease control knowledge base collection | Limits the retrieval scope to avoid introducing irrelevant documents outside this category |
Three Common Misconfigurations
- The
insufficient_quota Current group upstream load is saturated, please try again latererror occurs. The cause is that reasonable retrieval concurrency limits are not configured, and too many computing resources are occupied during knowledge base incremental updates, leading to upstream service overload. - Generated documents do not match knowledge base content. The cause is that
datasetSearchUsingExtensionQueryis not enabled, or thesimilaritythreshold is set too low, resulting in retrieval of content from non-target documents, causing generated results to deviate from statutory reporting rules. - Slow retrieval response, high token consumption, or inability to upload files larger than 100MB. The cause is that the
limitvalue is not restricted, resulting in retrieval of too many redundant documents, and large-volume documents are not split compliantly, causing the context length to exceed a reasonable range, increasing token consumption and upload failure risk.
How to Verify Successful Configuration
- In the FastGPT v4.17.0 retrieval debugging tool, enter "Clustered epidemic report from a medical institution in February 2025" as the retrieval keyword, and verify that retrieved results include the core fields of the corresponding document.
- View the retrieval log to confirm that the number of results filtered by the
similaritythreshold meets expectations, with no large number of low-correlation non-target documents returned. - Simulate the daily incremental update process to verify that the knowledge base synchronization has no errors, and documents larger than 100MB can be uploaded after compliant splitting.
- Test retrieval for different document types to confirm that retrieved results for both structured infectious disease report cards and unstructured epidemic analysis reports match business requirements.
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.