What this category's data looks like
This category's data includes enrollment plans, filing rules, and admission rosters, with fields covering institution code, major group, planned number, filing ratio, minimum admission score, date, and signature. Enrollment plans come from provincial admissions agencies and institution submissions, and are updated centrally by batch; filing rules are released with the policy documents of the year and are often appended as attachments; admission rosters are returned by institutions after each batch ends, containing candidate identifiers and major placements. When these three types of materials enter the multi-turn dialogue and prompt feature, their roles differ: enrollment plans and filing rules serve as prior context constraining the answer scope, while admission rosters serve as a record source that can be queried further. Names, quantities, dates, and signatures are most prone to version differences when batches switch, and must be aligned by batch and date before being placed into the dialogue.
Configuration and implementation checks
This part is executed by the memory turns and prompt configuration in the application settings, outputting the referenced historical message range and answer constraints. The externally available entry points and authorization scope are determined by the publishing service and file permissions; the model only processes filtered materials, while the publishing service and file permissions handle identity and file authentication.
| Object | Check | Verification action |
|---|---|---|
| Enrollment plan | Institution code, major group, planned number, date, and signature are consistent across the three materials | Check the memory turns value in the application settings, run a trial with two consecutive questions, and compare the plan fields |
| Filing rules | Filing ratio, batch, and date align with the policy documents of the year, with no missing signature or receipt | Run a trial with two consecutive questions and check whether the second round carries forward the reference from the first round |
| Admission roster | Candidate identifiers, major placements, and planned numbers match across the two materials, with synchronized versions | Compare the usage records of two runs to confirm that consumption changes accordingly after adjusting the turns |
| External entry point | Visitor identity matches the visible scope, and unauthorized access is not visible | Access once with a restricted identity and see whether the visible scope matches the setting |
Three common errors
- Symptom: the major name for the same candidate does not match between the enrollment plan and the admission roster. Cause: the two materials' versions are not synchronized, their update rhythms differ, and receipts are missing.
- Symptom: the attachment date of the filing rules is inconsistent with the body date, and the signature is missing. Cause: policy documents are appended in batches, and the old version was not replaced.
- Symptom: when the second question uses a reference, the answer does not address the question, and responses become slower with increased consumption. Cause: the memory turns are too large, bringing in irrelevant prior context and opening messages.
How to verify the setup
This feature's acceptance: ask two consecutive follow-up questions and see whether the second question carries forward the reference from the first; check the memory turns in the application settings, and inspect the values of chatId, messages, and variables within the historical message range.
Business test: access once with a restricted identity and see whether the visible scope matches the setting
- Check the current value of memory turns in the application settings and match it with the prior context range required by the question
- Run a trial with two consecutive questions and confirm that the second round can correctly carry forward the reference from the first round