What this category's data looks like
This page is intended for peer review organizations. The materials come from three types of manuscript correspondence records: review invitations, expert review comments, and author revision notes. The update cadence follows each round of review. Fields include manuscript number, reviewer name, review round, comment items, revision notes, date, and signed receipt. After these materials enter the document parsing and chunking feature, they each take on different roles: review invitations provide reviewer identity and invitation date, expert review comments provide comment items and professional terminology, and author revision notes provide item-by-item responses and terminology translations. The difficulty in multilingual materials and terminology handling lies in aligning the same term across different rounds and different language materials. After parsing and chunking, they are stored as data rows for retrieval and subsequent processing to generate a unified terminology glossary and comparison draft.
Configuration and implementation checks
Parsing and training tasks (server-side tasks) only output data rows and index status (parsed, indexing, indexed). The unified terminology glossary and comparison draft must be handled next by retrieval and subsequent processing. After parsing is complete, parsed chunks are immediately stored 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 |
|---|---|---|
| Review invitation | Whether the reviewer name, invitation date, manuscript number are consistent with the corresponding records in expert review comments and author revision notes | Separately verify the reviewer name, invitation date, and manuscript number in the review invitation and expert review comments, confirm that the same record's name and date match, and check whether signatures or receipts are complete |
| Expert review comments | Whether the number of comment items and review rounds correspond to the response items in the author revision notes, and whether terminology translations can refer back to the original text | Compare the number of comment items and rounds in expert review comments and author revision notes one by one, sample three terms to check whether their spelling is consistent throughout the document, and verify whether terms are split after chunking by chunkSize and chunkSplitter |
| Author revision notes | Whether the dates and signatures in the revision notes are synchronized with the versions of the review invitation and expert review comments | Verify the dates and signatures in the author revision notes, confirm that the version is synchronized with the review invitation and expert review comments, check whether the data row status is parsed, indexing, or indexed, and record the processing mode corresponding to trainingType |
| Data row | Whether it is in one of the three states and consistent with the interface display, whether data pending indexing is not recalled, and whether rebuilding only covers data with completed indexing | Open the data row to verify the status, separately verify the file record and the data row identifier, and check whether the data rows covered by the rebuild task include only data with completed indexing |
Three common errors
- Symptom: The name and date of the same review record do not match between the review invitation and expert review comments. Cause: The two materials were entered separately, were not aligned by manuscript number and round, and the versions were not synchronized.
- Symptom: The author revision notes lack signatures or receipts, and the number of comment items is inconsistent with the expert review comments. Cause: The revision notes circulated separately, receipts were not collected, and item additions or deletions were not written back.
- Symptom: Newly uploaded materials cannot be found, and failed training leaves empty data rows. Cause: Successful parsing does not equal retrieval hits, the index status has not reached completed, and failed data has not been cleaned up.
How to verify the setup
Acceptance for this feature: Upload one material, observe the data row status transitions and the coverage scope of the rebuild task; confirm that after parsing is complete it is immediately stored as visible data rows and bound to a unique data identifier, that the status transitions through parsed, indexing, and indexed, that data pending indexing does not participate in recall, and that changing models only rebuilds data with completed indexing, with the rebuild scope excluding data pending indexing.
Business test: Sample three terms to check whether their spelling is consistent throughout the document, whether translations can refer back to the original text, verify the chunking and processing modes corresponding to chunkSize, chunkSplitter, and trainingType, and separately record the number of citations and the tokens of the cited text.
- Open the data row to verify the status, confirm that it is one of parsed, indexing, or indexed and consistent with the interface display, and separately verify the file record and the data row identifier
- View the data rows covered by the rebuild task, confirm that only data with completed indexing is covered, that data pending indexing does not participate in rebuilding, and check whether the dates and signatures of the review invitation, expert review comments, and author revision notes are complete