What this category's data looks like
This page addresses credit management for primary and secondary school teacher training. The materials come from training notices, credit recognition forms, and participant lists. Training notices are issued by the project office and include the training name, start and end dates, conversion rules between class hours and credits, signature block, and seal; credit recognition forms are filled in by individual teachers or school administrators and include the training project, recognized credits, review comments, and signature; participant lists are compiled by the organizing unit and include name, unit, course name, attendance, and receipt. The three types of materials are updated by project batch, and the same project often has notices, recognition forms, and lists existing side by side. When entering this function, the notice serves as the basis source, the recognition form as the result source, and the list as the process source; each is split and written into data rows, and the index service generates vector identifiers to support subsequent retrieval and citation trimming for supporting materials. Field characteristics are concentrated in names, quantities, dates, and signatures, and cross-material verification is based on the original text entries.
Configuration and implementation checks
This step is performed by the index service and vector database, outputting vectors and index status (whether recall is possible and whether data in the same group is merged). "Basis entries that can be traced back to the original text" are then completed by subsequent AI nodes, code nodes, or manual work (retrieval and citation trimming only take supporting materials). The index model is consistent with the vector model actually used by the knowledge base, and data rows record the corresponding multiple vector identifiers in the index field.
| Object | Check | Verification action |
|---|---|---|
| Index model | Consistent with the vector model actually used by the knowledge base | Check the index model in the knowledge base settings and confirm it is the same as the model used when the knowledge base was created |
| Index field of the data row | Multiple vector identifiers for one piece of data are all present, and the status is indexed | Open a data row and check the index field to confirm that multiple vector identifiers are complete and the status is indexed |
| Rebuild task record | After changing the model, trigger only for data whose indexing has already been completed | Check the data range covered by the rebuild task record and confirm that unindexed data has not been included |
| Training notice and credit recognition form | Whether the name, quantity, and date of the same record match across the two materials | Sample the same project and compare the training name, number of class hours, start and end dates, and signature fields item by item |
Three common errors
- Symptom: duplicate entries appear in retrieval results. Inspect the following: the same group of original data corresponds to multiple vectors, and after recall they are not merged back into the original data by identifier, or the index field does not record all vector identifiers.
- Symptom: recall quality drops after changing the model. Inspect the following: the rebuild task record only covers data whose indexing has already been completed; data pending indexing does not participate in recall, so indexing must be completed before rebuilding.
- Symptom: a data row has content but cannot be retrieved. Inspect the following: the index field is empty or the status has not reached indexed, the vector has not been written to the vector database, or missing receipts or signatures cause the material not to be ingested.
How to verify the setup
Acceptance for this function: sample a data row and check the index field to confirm multiple vector identifiers and status indexed, then run a retrieval once to see whether the same group is merged and whether the score takes the highest score.
Business test: ask a cross-material basis question and see whether the answer provides a traceable source, and whether the source matches the original text entries in the training notice, credit recognition form, or participant list.
- Check the index model in the knowledge base settings, confirm it is consistent with the vector model actually used by the knowledge base, and record the values of
defaultTokenandmaxToken. - Open the data row and check the index field to confirm that multiple vector identifiers are complete and the status is indexed, and check the data range covered by the rebuild task record.