What this category's data looks like
In material version and change maintenance, peer review organizations draw on sources including review invitations, expert review comments, author revision notes, and version ledgers and change records. A review invitation contains the manuscript number, expert name, invitation date, and receipt signature; expert review comments contain review dimensions, the comment body, submission date, and signature; author revision notes contain point-by-point responses, revision locations, dates, and signatures. The update cadence advances by rounds, with each round producing a new version. When these materials enter document parsing and chunking, review invitations and expert comments serve as data rows to be parsed, while author revision notes serve as reference material. Field characteristics center on the consistency of names, quantities, dates, and signatures.
Configuration and implementation checks
This stage is executed by parsing and training tasks (server-side tasks), outputting data rows and index status (parsed, indexing, indexed). The version ledger and change records are 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 |
|---|---|---|
| Review invitation | Name, date, and signature must be consistent with the version ledger | Check the manuscript name, invitation date, and receipt signature in the invitation record against the version ledger, and confirm the current version is unique |
| Expert review comments | The name, quantity, and date of the same record do not match across the two materials | Compare the number of comments, dates, and signatures in the expert comments and the author revision notes, and confirm the old version is traceable and references are synchronized |
| Author revision notes | Version not synchronized or signature missing | Check whether the version number and references in the revision notes are updated in sync, and confirm the signature and receipt are complete |
| Data row | In one of the three states and consistent with the interface display | Open the data row to check its status, and confirm that chunkSize, chunkSplitter, trainingType, and the index status transitions are consistent |
Three common errors
- Symptom: newly uploaded material cannot be found. Cause: the data row is in parsed or indexing state; data pending indexing does not participate in recall, so you must wait for indexing to complete.
- Symptom: an empty data row remains after training fails. Cause: successful parsing does not equal a retrieval hit; the index status is not completed, and failed data can be edited or deleted.
- Symptom: rebuilding takes abnormally long after switching models. Cause: rebuilding only targets data whose indexing is completed, whereas initial training covers data pending indexing and data being indexed; the two cannot be interchanged.
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, and confirm that chunkSize, chunkSplitter, trainingType, and the three states are consistent.
Business test: change one piece of content and see whether the version number and references are updated in sync.
- Check the identifiers of the file record and the data row, and confirm that their identities are distinguishable.
- Inspect the data rows covered by the rebuild task, and confirm that only data whose indexing is completed is covered.