What this category's data looks like
This page addresses grade entry and credit recognition scenarios, with materials including three types: transcripts, deferred examination applications, and make-up examination arrangements. Transcripts carry course name, credits, grade, and grade point fields; deferred examination applications record the application reason, original examination date, and signature field; make-up examination arrangements list the make-up examination time, location, and receipt requirements. Material sources are academic affairs system exports and departmental submissions, with an update rhythm that changes with the semester examination cycle, and field characteristics are the coexistence of multilingual course names and terminology, inconsistent date formats, and scattered signature and receipt statuses. When materials enter the model access and configuration function, each has a different role: transcripts serve as anchor text for terminology comparison, while deferred examination applications and make-up examination arrangements serve as parallel materials that need to refer back to the original text. The unified terminology table and comparison draft are produced by the model bound to dialogue and retrieval use, and the configuration step is only responsible for the model binding relationship and channel response.
Configuration and implementation checks
This step is executed by the server-side model channel configuration, outputting the model binding relationship and channel response. The "unified terminology table and comparison draft" is subsequently completed by AI nodes, code nodes, or manual work. Models are divided into five types: language, index, rerank, speech synthesis, and speech recognition. Different steps bind different model types, and during configuration it is necessary to check the type annotation and the model name returned by the channel.
| Object | Check | Verification action |
|---|---|---|
| Transcript | Course name, credits, grade, and grade point fields need to be aligned item by item with deferred examination applications and make-up examination arrangements | In the model configuration, check the type annotation, confirm that the language model and index model are bound to dialogue and retrieval respectively, and compare against the model name returned by the channel |
| Deferred examination application | Application reason, original examination date, and signature field need to be consistent with transcript records, and date formats unified | Check the channel configuration and key source, confirm the index model is available, and avoid retrieval errors related to the index model |
| Make-up examination arrangement | Make-up examination time, location, and receipt requirements need to be consistent with the course names, quantities, and dates in deferred examination applications and transcripts | In dialogue and retrieval, check whether the values of maxContext, maxTokens, and quoteMaxToken match the material length, and separately record the number of items in the citation array |
| Model identifier | Consistent with the names listed in the documentation, and the answer style should not change abruptly after switching models | Compare against the model name returned by the channel, and confirm that toolChoice is configured according to the documentation in the steps where needed |
Three common errors
- Symptom: The same course has inconsistent names in the transcript and make-up examination arrangement. Cause: The two material versions were not synchronized, and the terminology table did not cover the translated name of that course.
- Symptom: The dates and quantities in the deferred examination application and transcript do not match. Cause: The materials were entered at different times, and item-by-item verification referring back to the original text was not performed.
- Symptom: The make-up examination arrangement lacks a signature or receipt. Cause: The signature page was not collected during material circulation, and the receipt status was not updated with the version.
How to verify the setup
Acceptance for this function: run one dialogue and one retrieval, check the model identifier in the response, confirm that the language model and index model are bound correctly, and check whether maxContext, maxTokens, and quoteMaxToken match the materials.
Business test: sample three terms and check whether their spelling is consistent throughout the entire text, and confirm that translated names can refer back to the original text.
- Check the channel configuration and key source, confirm authentication succeeds and the model type corresponds to the step.
- Compare against the model name returned by the channel, and confirm that the model identifier is consistent with the names listed in the documentation.