What this category's data looks like
In the journal publication license and annual verification scenario, materials come from three categories: journal publication licenses, annual verification forms, and rectification reports. The license contains the competent authority, publication number, publication period, validity period, and seal; the annual verification form contains the verification year, publication period, sponsoring organization, signature, and receipt; the rectification report contains problem items, rectification measures, completion date, and approval signature. The update rhythm is centralized submission for annual verification, supplementary submission of rectification reports by batch, and synchronous updates when licenses are reissued. When the three types of materials enter parsing and training tasks, their roles differ: the license and verification form serve as baseline comparison materials, while the rectification report serves as the discrepancy source material. They are respectively stored as visible data rows and bound to unique data identifiers for subsequent retrieval and discrepancy localization.
Configuration and implementation checks
This step is executed by parsing and training tasks (server-side tasks), outputting data rows and index statuses (parsed, indexing, indexed). The discrepancy list is then completed by retrieval and subsequent processing.
| Object | Check | Verification action |
|---|---|---|
| Journal publication license | Whether fields, dates, and signatures are consistent with the annual verification form | Separately verify the publication number, publication period, competent authority, and validity period in the license and verification form, and compare signatures and seals; check that the data row status is parsed / indexing / indexed, and use chunkSize to record the chunking granularity |
| Annual verification form | Whether the name, quantity, and date of the same record match across the two materials | Verify the sponsoring organization, publication period, verification conclusion, and receipt date item by item according to the verification year, mark records whose names or dates do not match, and check whether file records and data row identifiers are distinguishable |
| Rectification report | Whether problem items, rectification measures, completion date, and approval signature are complete | Verify item by item the correspondence between rectification items and the problem numbers in the verification form, check items missing signatures or receipts, and check the processing mode corresponding to trainingType and chunkSplitter |
| Data rows and index status | Is in one of the three states and consistent with the interface display | Open the data rows to verify the status, confirm that data pending indexing does not participate in recall; check whether the data rows covered by the rebuild task only include data with completed indexing |
Three common errors
- Symptom: Newly uploaded materials cannot be found. Cause: Parsing is complete and has been stored as visible data rows, but indexing is still in indexing; data pending indexing is naturally isolated at the retrieval layer and is not recalled, so it is necessary to wait for the transition to indexed.
- Symptom: The name, quantity, and date of the same record do not match across the two materials. Cause: The versions of the license and annual verification form are not synchronized, or the problem numbers in the rectification report are inconsistent with the verification form, and the discrepancy list has not been written back to the reviewer in time.
- Symptom: Empty data rows remain after training fails. Cause: Training interruption leaves orphan data; failed data supports direct editing or deletion, but if not cleaned up, the status remains in indexing and requires manual verification before handling.
How to verify the setup
Acceptance of this feature: Upload one material and observe the data row status transitions and the coverage scope of the rebuild task. Check whether the status transitions according to parsed -> indexing -> indexed, whether the rebuild task only covers data with completed indexing, and whether the initial training covers data pending indexing and data being indexed.
Business test: Use one real record to find fields inconsistent with the ledger. Take one journal record, separately verify the name, quantity, and date in the license and annual verification form, mark the fields that do not match, and submit them to the reviewer for confirmation and write-back.
- Open the data rows to verify the status, confirm that the three states are consistent with the interface display, and separately verify file records and data row identifiers.
- View the data rows covered by the rebuild task, and confirm that model replacement only rebuilds data with completed indexing, and that data pending indexing does not participate in the rebuild chain.