What this category's data looks like
This category's data includes three types of materials: business cooperation agreements, advertising review forms, and distribution cooperation materials. Sources include submissions from partners, internal ledgers, and historical archives. The update cadence changes with cooperation cycles and version iterations. Field characteristics center on names, quantities, dates, signatures, and receipts. When business cooperation agreements enter the feature, they serve as the primary record, while advertising review forms and distribution cooperation materials serve as comparison materials, each taking on the role of field comparison and discrepancy localization, supporting consistency verification of the same object across the three materials.
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 discrepancy list is then completed by retrieval and subsequent processing. After parsing is complete, parsed chunks are immediately persisted as visible data rows and bound to a unique data identifier. The index transitions through three states, and data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Business cooperation agreement | Whether names, quantities, dates, signatures, and receipts are consistent across the three materials | Separately verify the file records and data rows, confirm that the data row status is in one of the three states and is consistent with the interface display, and check whether chunkSize and chunkSplitter preserve complete fields when splitting fields |
| Advertising review form | Whether names, quantities, dates, signatures, and receipts are consistent with the corresponding fields in the business cooperation agreement | Open the data row to verify the status, confirm that trainingType corresponds to the processing mode, and check fields where the version is not synchronized or signatures/receipts are missing |
| Distribution cooperation materials | Whether names, quantities, dates, signatures, and receipts are consistent with the ledger | Look at the data rows covered by the rebuild task, confirm that only data with completed indexing is covered, and verify the respective identity identifiers of the file records and data rows |
Three common errors
- Symptom: The names, quantities, and dates of the same record do not match across two materials. Cause: After parsing and chunking, fields are not aligned to the same object, and retrieval and subsequent processing do not perform discrepancy localization.
- Symptom: Version desynchronization causes old-version fields to be recalled. Cause: Pending rows are excluded from recall; an older indexed version remains available and is returned by retrieval.
- Symptom: Materials missing signatures or receipts are mistakenly judged as consistent. Cause: Material verification does not cover signature and receipt fields, and the discrepancy list does not mark missing items.
How to verify the setup
Feature acceptance: Upload one material, observe the data row status transitions and the coverage scope of the rebuild task, and confirm that the statuses are parsed, indexing, and indexed in sequence, and that the rebuild only covers data with completed indexing.
Business test: Use one real record to find fields inconsistent with the ledger, hand them to the reviewer for confirmation, and write back the discrepancy list.
- Separately verify the file records and data rows, confirm that the data row status is consistent with the interface display, and check the field splitting results related to
chunkSize,chunkSplitter, andtrainingType. - Look at the data rows covered by the rebuild task, confirm that only data with completed indexing is covered, verify the file and data row identifiers, and mark discrepancies in names, quantities, dates, signatures, and receipts.