What this category's data looks like
Materials for the business basis query for undergraduate program setup and dynamic adjustment come from program setup application forms, talent cultivation plans, expert demonstration opinions, and documents such as the university's program setup and adjustment management measures. These materials are updated by academic year or batch, and their fields include program name, program code, length of study, degree-granting category, enrollment number, adjustment date, demonstration conclusion, and signature field. The program setup application form records the proposed program and the reasons for adjustment, the talent cultivation plan records the course and credit structure, and the expert demonstration opinions record the demonstration conclusion and signatures. After the materials enter dialogue and retrieval, they respectively take on the roles of source text basis, version verification, and signature verification, and when replying to parents or writing them into a plan, they must refer back to the source text.
Configuration and implementation checks
This link is executed by the server-side model channel configuration, which outputs the model binding relationships and channel responses. The basis entries that can refer back to the source text are then completed by subsequent AI nodes, code nodes, or manual work. Models are divided into five categories: language, indexing, reranking, speech synthesis, and speech recognition, and different links bind different model types.
| Object | Check | Verification action |
|---|---|---|
| Program setup application form | Whether the fields, dates, and signatures are complete, and whether the program name and program code match the source text | Check the program name, program code, length of study, adjustment date, and signature field in the application form, and confirm they are consistent with the corresponding records in the talent cultivation plan; in the model configuration, check that the language model identifier is consistent with the model name returned by the channel, and check whether maxContext and maxTokens cover the long text of the application form |
| Talent cultivation plan | Whether the version is the current one, and whether the course and credit counts can refer back to the source text | Check the version number and effective date of the talent cultivation plan, and confirm they are consistent with the adjustment date in the program setup application form; check that the indexing model type label is consistent with the connected service, and during retrieval check the reference text tokens corresponding to quoteMaxToken, recording the number of array entries separately |
| Expert demonstration opinions | Whether the signatures and receipts are complete, and whether the demonstration conclusion can refer back to the source text | Check the expert signatures, dates, and receipts in the expert demonstration opinions, and confirm they are consistent with the same record names, quantities, and dates in the program setup application form and talent cultivation plan; check the binding relationship between the reranking model and the language model, and confirm that toolChoice is consistent with the name listed in the document |
| Model binding relationships | Whether the model type corresponds to the link in which it is to be used, and whether the model identifier is consistent with the name listed in the document | In the model configuration, check the type labels of the language model, indexing model, and reranking model; check the model identifier against the model name returned by the channel; confirm that the channel and key are consistent with the connected service, and when authentication fails, first check the channel configuration and key source |
Three common errors
- Symptom: the name, quantity, and date of the same program do not match between the program setup application form and the talent cultivation plan. Cause: the two materials are maintained by different links, and cross-material field verification was not performed.
- Symptom: the version of the talent cultivation plan is not synchronized, and retrieval still returns the old courses and credits. Cause: after the materials were updated, they were not re-indexed, and the indexing model does not correspond to the current version.
- Symptom: the expert demonstration opinions lack signatures or receipts, and the basis entries cannot refer back to the source text. Cause: the signature field and receipts were not archived together with the materials, and verification only checked fields but not signatures.
How to verify the setup
Acceptance of this feature: run one dialogue and one retrieval, check whether the model identifier in the response is consistent with the name listed in the document, and check the model name returned by the channel; check whether maxContext, maxTokens, quoteMaxToken, and toolChoice are configured according to the document, and record the number of citations and the citation content budget separately.
Business test: ask a basis question that spans materials, and see whether the answer gives a citable source, and whether the source matches the source text of the program setup application form, talent cultivation plan, and expert demonstration opinions.
- Check the fields, dates, and signatures of the program setup application form, talent cultivation plan, and expert demonstration opinions, and confirm that the name, quantity, and date of the same record are consistent.
- In the model configuration, check the type labels and model identifiers of the language model, indexing model, and reranking model, and confirm that the channel and key are consistent with the connected service.