What this category's data looks like
When kindergarten operation behavior and quality self-assessment materials enter knowledge base retrieval and recall, the self-assessment report, issue list, and rectification plan each play different roles. The self-assessment report provides an overall description of dimensions such as operating conditions, safety and hygiene, care and education, faculty and staff, and internal management; the issue list records specific items, responsible persons, and discovery dates; the rectification plan carries measures, completion deadlines, and signed receipts. The materials mostly come from kindergarten submissions, feedback from higher-level inspections, and summaries of parent questionnaires, and the update rhythm is synchronized with semester inspections and special supervision. Field characteristics are concentrated in names, quantities, dates, signatures, and receipts. The same record in two materials can easily have mismatched names, quantities, and dates, with versions not synchronized or signatures and receipts missing. The retrieval service cuts the three types of materials into candidate chunks for subsequent nodes or manual generation of external entry points and authorization scopes.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), outputting candidate chunks and similarity scores (after pattern, merge, rerank, filter, and upper-limit trimming). The external entry points and authorization scopes are then completed by subsequent nodes or manually.
| Object | Check | Verification action |
|---|---|---|
| Self-assessment report | Whether the dimension fields, dates, and signatures are consistent with the issue list and rectification plan | In the search test, test once each with synonymous questions, record the number of candidate items and similarity scores, and check whether the names, quantities, and dates in the report can be recalled |
| Issue list | Whether the item names, quantities, and discovery dates match the rectification plan | Switch searchMode to semantic, full-text, or hybrid, compare whether the candidates simultaneously contain chunks from both types of materials, and check the names and quantities of the same record |
| Rectification plan | Whether the measures, completion deadlines, signatures, and receipts are complete and the versions are synchronized | Compare the top few rankings before and after enabling usingReRank, observe whether key paragraphs are retained, and check the signature and receipt fields |
| External entry points and authorization scope | Whether the visitor identity matches the visible scope, and unauthorized access is not visible | Access once with a restricted identity to see whether the visible scope matches the setting; adjust similarity and limit, record the number of candidate items and scores, and record quoted text tokens separately |
Three common errors
- Phenomenon: The names, quantities, and dates of the same record in the self-assessment report and the issue list do not match. Cause: The two materials are submitted separately and no field-level verification is performed; retrieval only returns candidate chunks and does not automatically compare them.
- Phenomenon: The rectification plan versions are not synchronized, and old measures and new deadlines are recalled at the same time. Cause: The material update rhythm is inconsistent, and version markers and signature/receipt verification are lacking.
- Phenomenon: There are many candidates, but the ones ranked first are all irrelevant, and the answer lacks key paragraphs. Cause: The retrieval mode does not match the material form, rerank is not enabled, or similarity filtering and quoted upper-limit trimming are too tight.
How to verify the setup
Acceptance for this feature: test once each with synonymously rewritten questions, compare candidates and similarity scores, then toggle rerank on and off to observe the top few rankings; switch searchMode, observe whether candidates simultaneously contain semantic and full-text results, record the number of candidate items and scores, record quoted text tokens separately, and adjust limit and similarity as needed.
Business test: access once with a restricted identity to see whether the visible scope matches the setting.
- In the knowledge base search test, test the same question once each with different phrasings, record the number of candidate items and similarity scores, and check whether the names, quantities, and dates in the self-assessment report, issue list, and rectification plan are consistent.
- View the top few rankings once each with rerank enabled and disabled, check whether the signature and receipt fields are retained with the candidates, and confirm that the external entry points and authorization scope are completed by subsequent nodes or manually.