What this category's data looks like
In the initial review scenario for academic journal submissions, materials include three categories: submitted manuscripts, author information forms, and plagiarism check reports. Submitted manuscripts are provided by authors and contain the title, author bylines, affiliations, funding information, and body structure; author information forms are collected using the editorial office template and contain names, ordering, affiliations, contact details, and signature fields; plagiarism check reports are exported by the detection system and contain detection time, similarity rate range, and comparison sources. The three material categories update at different rhythms: manuscripts may undergo multiple rounds of revision, information forms are refilled as bylines change, and plagiarism check reports are regenerated per version. In terms of field characteristics, names, quantities, dates, signatures, and receipts are verification anchors; the same object is often written differently across two materials, so parsing and chunking must come first before discrepancies can be localized. When materials enter this feature, the manuscript and information form serve as sources of object identity and attributes, while the plagiarism check report serves as time and version corroboration; together the three support discrepancy localization.
Configuration and implementation checks
This step is executed by parsing and training tasks (server-side tasks), outputting data rows and index states (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; data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Submitted manuscript | Name, quantity, date, and signature receipt are consistent across the three materials | Verify the title, author count, receipt date, and signature field, and confirm the data row state corresponding to the chunking parameters chunkSize, chunkSplitter, and trainingType |
| Author information form | Name, affiliation, ordering, and contact details are consistent with the manuscript | Compare the information form with the manuscript field by field, record name-writing discrepancies and missing signatures, and check whether the corresponding data row is in parsed or indexing |
| Plagiarism check report | Detection time, version, and receipt number are consistent with the ledger | Verify the report date and version number, confirm whether the report data row has completed indexing, and if not, mark it for re-review |
| Data row state | Is in one of the three states parsed, indexing, indexed and is consistent with the interface display | Open the data row to verify its state, confirm that data pending indexing does not participate in recall, and that rebuilds only cover data that has completed indexing |
Three common errors
- Symptom: a newly uploaded material cannot be found. Cause: the data row is in parsed or indexing; data pending indexing is naturally isolated at the retrieval layer, and successful parsing does not equal a retrieval hit.
- Symptom: the name, quantity, and date of the same record do not match across two materials. Cause: the manuscript and information form versions are not synchronized, and field-writing discrepancies were not captured by discrepancy localization.
- Symptom: empty data rows remain after a training failure. Cause: interrupted training leaves orphan data behind, and materials missing signatures or receipts are not cleaned up in time, affecting subsequent verification.
How to verify the setup
Feature acceptance: upload one material, observe the data row state transitions and the coverage scope of the rebuild task, confirm that the three states parsed, indexing, and indexed are consistent with the interface, and that rebuilds only cover data that has completed indexing.
Business test: use one real record to find fields inconsistent with the ledger, record name, quantity, date, or signature discrepancies, hand them to the reviewer for confirmation, and write them back.
- Verify file records and data rows separately, confirm that file and data row identifiers are distinguishable, and check the configurations corresponding to
chunkSize,chunkSplitter, andtrainingType - Check the data rows covered by the rebuild task, confirm that it targets only data that has completed indexing, and that data pending indexing and data being indexed do not participate in the rebuild