What this category's data looks like
In course selection, scheduling, and classroom resource allocation scenarios, materials come from the course offering plans, scheduling tables, and course selection rules exported from the academic affairs system, as well as confirmation documents and signed receipts returned by each department. The update cadence follows semester rounds and course selection phases, and the fields mainly consist of course name, course code, class size, classroom capacity, week and period, effective date, and approval signature. The course offering plan plays the baseline role for courses and credits, the scheduling table plays the role of time and space placement, the course selection rules play the role of admission and capacity constraints, and the receipts play the role of version and signature confirmation. After materials enter the document parsing and chunking feature, they are first split by structure into retrievable data rows, each row is bound to a unique data identifier and enters the three-state index flow, for subsequent retrieval and discrepancy localization calls.
Configuration and implementation checks
This step is executed by the parsing and training task (server-side task), outputting data rows and index states (parsed, indexing, indexed). The discrepancy list is then completed by retrieval and subsequent processing. After parsing is completed, parsing and chunking are immediately persisted as visible data rows, and the index flows from parsed to indexing to indexed; data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Course offering plan | The course name, course code, credits, and course start date are consistent in both materials, and the signature column has a corresponding approval record | Check the file records and data rows separately, inspect whether the fields after chunkSize and chunkSplitter splitting fall into the same data row, and confirm the status is indexed |
| Scheduling table | The class size, classroom capacity, week and period, and effective date of the same course match between the ledger and the scheduling table | Open the data rows to check the status, view rows in parsed or indexing state in read-only mode, and compare the fields under the processing mode corresponding to trainingType |
| Course selection rules | The admission conditions, capacity limit, course selection time window, and version number are consistent in both materials, and the receipts are complete | Look at the data rows covered by the rebuild task, confirm that only data with completed indexing is covered, and list unsynchronized versions separately |
| Signed receipt | The course scope, date, and signatory on the receipt are consistent with the records in the course offering plan and scheduling table | Check the file records and data rows separately, and inspect whether the receipt fields are completely split and enter the indexed state |
Three common errors
- Symptom: The name, class size, or effective date of the same course does not match between the course offering plan and the scheduling table. Cause: The two materials are exported from different stages, the field definitions and versions are not synchronized, and the discrepancy is not exposed before review.
- Symptom: The corresponding data row cannot be retrieved immediately after the material is uploaded. Cause: The data row is in parsed or indexing state, and data pending indexing is isolated at the retrieval layer and has not yet entered indexed.
- Symptom: After changing models, the rebuild takes abnormally long, or empty data rows are left after a training failure. Cause: Rebuild only targets data with completed indexing, data pending indexing does not participate in rebuild, and failed data must be edited or deleted before retrying.
How to verify the setup
Feature acceptance: Upload a material, observe the data row status flow and the coverage scope of the rebuild task, confirm that after parsing is completed it is immediately persisted as visible data rows and bound to a unique data identifier, the status passes through parsed, indexing, and indexed in sequence, and rebuild only covers data with completed indexing.
Business test: Use one real record to find fields inconsistent with the ledger, compare name, quantity, date, and signed receipt item by item, form a discrepancy list, submit it to the reviewer for confirmation, and write it back.
- Open the data rows to check the status, confirm it is consistent with the interface display, and check the respective identity identifiers of the file records and data rows separately.
- Look at the data rows covered by the rebuild task, and confirm that only data with completed indexing is covered and that data pending indexing does not participate in rebuild.