What this category's data looks like
This page addresses credit management for primary and secondary school teacher training, centering on three types of materials: training notices, credit recognition forms, and participant lists. A training notice provides the training name, class hours, credit standards, and the date of issuance; a credit recognition form records the teacher's name, recognized credits, recognition date, and review signature; a participant list lists the participants, attendance days, and sign-in receipts. Materials are often archived in batches by academic year or semester, and the update cadence is synchronized with training batches; after revision, older versions may still remain in the knowledge base. When these three types of materials enter knowledge base retrieval and recall, each serves as a source of candidate chunks: training notices provide the basis for credit standards, credit recognition forms provide the basis for recognition results, and participant lists provide the basis for participation facts. Field characteristics concentrate on four categories: names, quantities, dates, and signatures; the same record needs to be aligned item by item across different materials.
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 mode, merge, rerank, filter, and limit trimming). The basis entries that can be traced back to the original text are completed by subsequent nodes or manually.
| Object | Check | Verification action |
|---|---|---|
| Training notice | Check whether the training name, class hours and credits, date of issuance, and signature or receipt are complete, and whether they are consistent with the current batch | In the search test, switch searchMode to hybrid and check whether the candidate chunks cover the notice's credit clauses and date line |
| Credit recognition form | Check whether the teacher's name, recognized credits, recognition date, and review signature match the original notice | Compare the number of candidate chunks and scores to confirm that the recognition form chunks contain the name, credits, and signature fields |
| Participant list | Check whether the participants, attendance days, and sign-in receipts are consistent with the recognition form's headcount and dates | Compare the top few rankings before and after enabling usingReRank, and check whether the list chunks are ranked in relevant positions |
| Retrieval output | Check whether candidate chunks can be traced back to the original text, whether similarity passes the similarity filter, and whether the number of citations is trimmed by limit | Record the number of candidate chunks and scores, record the cited text tokens separately, and do not compare the number of chunks with the budget as equal |
Three common errors
- Phenomenon: the same teacher's name is written inconsistently in the recognition form and the participant list. Cause: the two materials were entered separately, and names were not aligned to the same record.
- Phenomenon: the training notice has been revised, but the old credit standard is still recalled from the knowledge base. Cause: versions were not synchronized, and old materials were not marked as invalid.
- Phenomenon: the recognition form lacks a review signature or the participant list lacks a receipt, yet it is still recalled as a candidate. Cause: signature and receipt completeness were not checked when the materials were archived.
How to verify the setup
This feature's acceptance: test once each with synonym-rewritten questions, compare candidates and scores, then toggle rerank to see the top few rankings; at the same time switch searchMode, adjust similarity and limit, and record the number of candidate chunks and scores.
Business test: ask a cross-material basis question, see whether the answer gives a traceable source, and check whether that source can be traced back to the original text chunks of the training notice, credit recognition form, or participant list.
- In the search test, switch among semantic, full-text, and hybrid modes for the same basis question, record the number of candidate chunks and scores, and check whether both types of recall results are included.
- Trace each returned candidate back to the original text one by one, check whether the four field categories of names, quantities, dates, and signatures match, and separately mark entries missing signatures or receipts.