What this category's data looks like
In the journal three-review-three-proof and editing quality scenario, materials mainly come from three types: review sheets, proofs, and proofreading records. Review sheets carry the fields and signatures of initial review, re-review, and final review; proofs carry the final presentation of body text, figures, and layout; proofreading records carry each proofreading round, change descriptions, and receipts. The update rhythm advances with proofreading rounds; each round of proofs and proofreading records is produced in pairs, and review sheets are completed at review-stage nodes. Field characteristics center on the name, quantity, date, signature, and version number of the same record. When the three types of materials enter document parsing and chunking, each is stored as an independent data row, bound to a unique data identifier, providing a traceable draft 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 status (parsed, indexing, indexed). The discrepancy list is then completed by retrieval and subsequent processing. After parsing is completed, parsed chunks are immediately stored as visible data rows, and the index transitions through the three states parsed, indexing, and indexed.
| Object | Check | Verification action |
|---|---|---|
| Review sheet | Whether the name, quantity, date, and signature of the same record are consistent across review stages | Open the data row to check the status, compare the review-stage fields and signature column, and confirm that trainingType corresponds to the processing mode |
| Proof | Whether the name, quantity, date, and version number of the same record are consistent with the ledger | Check the file record and the data row separately, and inspect whether chunkSize and the chunk boundaries cover the figure and layout fields |
| Proofreading record | Whether the date, change description, and receipt of each proofreading round are complete and version-synchronized | Check the receipt and signature, and confirm that after splitting by chunkSplitter, each leaf node can still be traced back to the same source record |
| Data row | Is in one of the three states parsed, indexing, or indexed and consistent with the interface display | View the data rows covered by the rebuild task, and confirm that rebuild is triggered only for indexed data |
Three common errors
- Symptom: a newly uploaded review sheet cannot be found in proof comparison. Cause: the data row is in parsed or indexing; data awaiting indexing does not participate in recall, so you must wait for indexing to complete.
- Symptom: the name, quantity, and date of the same record do not match between two materials. Cause: versions are not synchronized, old proofs and new proofreading records are mixed, and fields are not written back to the ledger.
- Symptom: empty data rows remain after training fails. Cause: successful parsing does not equal retrieval hits; failed data is not cleaned up, and missing signatures or receipts prevent the record from closing the loop.
How to verify the setup
This feature acceptance: upload one material, observe the data row status transitions and the coverage scope of the rebuild task, confirm the transition from parsed to indexing to indexed, and verify that changing the model rebuilds only indexed data.
Business test: use one real record to find fields inconsistent with the ledger, locate the name, quantity, or date discrepancy, and submit it to the reviewer for confirmation and write-back.
- Check the file record and the data row separately, confirm that the identifiers of the file and the data row are distinguishable, and then compare the fields and signatures of the review sheet, proof, and proofreading record.
- View the data rows covered by the rebuild task, and confirm that the initial training covers data awaiting indexing and data being indexed, while rebuild targets only data whose indexing has been completed.