What Data This Category Includes
Data for this category comes from software asset management systems, procurement systems, terminal scan results, and vendor authorization documents. Update frequency is monthly or quarterly. Daily used documents include license ledgers, software usage lists, procurement contracts, and compliance audit reports. Core fields are software name, authorized quantity, installed terminal count, expiration date, and vendor information. Units follow standard IT asset measurement standards, such as units, pieces, and days.
Constraints for Knowledge Base Retrieval and Recall
Low data update frequency eliminates need for high-frequency retrieval dataset synchronization, reducing retrieval resource usage. Multiple closely linked field dimensions require precise matching of core fields like authorized quantity and installation status. This prevents irrelevant results from disrupting operations and ticket decision-making. Document types include structured ledgers and unstructured contracts, requiring support for multi-format content recall. Queries for authorization verification have strict accuracy requirements. The correlation threshold for retrieval results must be tightly controlled to ensure returned content can directly support decision-making.
Configuration Settings
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
similarity | 0.75–0.85 | Matches the accuracy requirements for core fields of enterprise software license management |
limit | 10 | Balances context length and retrieval efficiency |
usingReRank | Enabled | Optimizes sorting logic for multi-field results |
datasetSearchUsingExtensionQuery | Enabled | Adapts to colloquial ticket queries from operations staff |
collectionIds | Corresponding exclusive dataset ID for license management | Isolates irrelevant data to improve retrieval accuracy |
Three Common Mistakes
- No change in search result ranking after configuring
usingReRank. The cause is that the re-ranking permission for the corresponding dataset is not enabled, or the re-ranking model call times out without triggering. getaddrinfo EAI_AGAIN mongoerror occurs after restarting the server. The cause is that MongoDB connection configuration was not persisted during knowledge base synchronization, or local network resolution has an exception.- Retrieval results cannot extract text content from PDF images. The cause is that the OCR extension processing module is not enabled, or the uploaded PDF only contains image layers without native text.
How to Verify Successful Configuration
- Submit a test query that includes authorized quantity and expiration date. Verify whether the field matching degree of returned results meets expectations.
- View retrieval logs in the FastGPT v4.17.0 backend. Confirm that the call status of
usingReRankis successful. - Upload a test PDF file with embedded images. Verify whether text content from images can be extracted.
- Adjust the value range of
similarity. Observe changes in the matching accuracy of retrieval results to confirm that the threshold configuration takes effect.
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 2026-10-19.