What this category's data looks like
Materials for regular higher education institution admissions enrollment filing include three categories: admissions plans, filing rules, and admission rosters. Sources are provincial admissions examination institutions and university admissions offices, updated by batch and year. Fields include institution code, major group, planned number, filing ratio, minimum admission score, candidate number, name, signature, and date. When entering document parsing and chunking, admissions plans take the role of verifying quota and major group fields, filing rules take the role of verifying process and ratio fields, and admission rosters take the role of verifying lists and signed receipts; the three categories of materials are respectively stored as data rows, bound to unique data identifiers, for retrieval and subsequent processing to assemble external documents according to the version ledger and change records.
Configuration and implementation checks
This step is executed by parsing and training tasks (server-side tasks), outputting data rows and index status (parsed, indexing, indexed). The version ledger and change records are then completed by retrieval and subsequent processing.
| Object | Check | Verification action |
|---|---|---|
| Admissions plan | Institution code, major group, planned number, date, and signature are unique in the current version and traceable to the old version | Open the data row to check whether the status is parsed, indexing, or indexed, and compare the planned number and date fields |
| Filing rules | Filing ratio, batch date, and signed receipt are consistent in the current version, and old-version references are updated synchronously | Separately verify the file record and the data row identifier, and check whether the chunks corresponding to trainingType and chunkSize cover the rule clauses |
| Admission roster | Candidate number, name, quantity, date, and signature match between the two materials, and missing signatures or receipts can be located | Check whether the data rows covered by the rebuild task contain only data with completed indexing, and verify the roster and planned quantities |
| Version ledger | The current version is unique, the old version is traceable, references are synchronized, and after changing one piece of content the version number and references are updated synchronously | After changing one piece of content, verify whether the version number and references are synchronized, and record the number of entries in the reference array and the tokens of the reference text |
Three common errors
- Symptom: The name, quantity, and date of the same record do not match between the two materials. Cause: After chunking, the admissions plan and admission roster are not aligned by the same data identifier, and old-version references are not synchronized.
- Symptom: A newly uploaded material cannot be found. Cause: The data row is in parsed or indexing, and data awaiting indexing is naturally isolated at the retrieval layer and not recalled; successful parsing does not equal a retrieval hit.
- Symptom: After a training failure, an empty data row remains, or rebuilding after changing models takes abnormally long. Cause: Failed training data was not edited or deleted, and the rebuild did not target only data with completed indexing.
How to verify the setup
Acceptance for this feature: Upload one material and observe the data row status transitions and the coverage scope of the rebuild task. Upload one admissions plan, observe the data row status go from parsed to indexing to indexed, and confirm that data awaiting indexing is not recalled; after changing models, check that the rebuild task covers only data rows with completed indexing, and verify that chunkSplitter, chunkSize, trainingType, and the chunking results are consistent.
Business test: Change one piece of content and see whether the version number and references are updated synchronously.
- Open the data row to check the status, confirm that the three states are consistent with the interface display, and separately verify the file record and the data row identifier.
- Check the data rows covered by the rebuild task, confirm that only data with completed indexing is covered, and record the number of entries in the reference array and the tokens of the reference text.