What this category's data looks like
In the graduate admissions re-examination and admission scenario, materials come from documents such as re-examination plans, candidate score sheets, and transfer lists, and are updated as batches and departments change. The re-examination plan contains procedures, scoring, and dates; the score sheet contains candidate numbers, subjects, scores, and rankings; the transfer list contains names, preliminary exam scores, intentions, and signature fields. After materials enter the document parsing and chunking function, they are split into data rows according to chunkSize and chunkSplitter, and each row is bound to a unique data identifier as input for retrieval and subsequent processing; trainingType determines the vector or QA processing mode. The update cadence is driven by admissions batches, and the names, quantities, and dates of the same record across different materials must remain consistent, with signatures and receipts complete.
Configuration and implementation checks
Parsing and training tasks (server-side tasks) only output data rows and index status (parsed, indexing, indexed), while the version ledger and change records must be handled next by retrieval and subsequent processing. After parsing is completed, parsing chunks are immediately persisted as visible data rows and bound to a unique data identifier, and the index transitions through the three states parsed, indexing, and indexed.
| Object | Check | Verification action |
|---|---|---|
| Re-examination plan | The dates, signatures, and receipts of the plan clauses must be consistent with the score sheet and transfer list | Check the dates, signature fields, and receipt records in the plan, and confirm they match the score sheet and transfer list |
| Candidate score sheet | The candidate number, subjects, scores, and ranking are consistent with the same-named record in the transfer list | Separately check the candidate numbers, quantities, and dates in the score sheet and transfer list, and confirm there are no missing signatures or receipts |
| Transfer list | The names, preliminary exam scores, and intentions in the list are consistent with the records in the re-examination plan and score sheet | Check the names, quantities, dates, and signatures in the list, and confirm they are synchronized with the other two materials |
| Data row | It is in one of the three states parsed, indexing, or indexed and is consistent with the interface display | Open the data row to check the status, and check the values of chunkSize, chunkSplitter, and trainingType |
Three common errors
- Symptom: The name, quantity, or date of the same candidate in the score sheet and transfer list do not match. Cause: The two materials are updated separately, without version synchronization and cross-checking.
- Symptom: The old version of the re-examination plan is still hit by retrieval, and the current version is not unique. Cause: The version was not synchronized, and the old data rows were not marked invalid or updated with references.
- Symptom: The transfer list lacks signatures or receipts, and the materials are incomplete. Cause: Signatures and receipts were not completed together with the list update, and the version ledger did not record the changes.
How to verify the setup
Open the data row to check whether the status is consistent with the interface display, and separately check the identity identifiers of the file record and the data row.
View the data rows covered by the rebuild task, confirm that only data with completed indexing is covered, and that data pending indexing does not participate in the rebuild.
- Upload one material, observe the data row status transition from parsed to indexing and then to indexed, and check that the rebuild task only covers data with completed indexing.
- Change one piece of content, and see whether the version number and references are updated synchronously.