What this category's data looks like
This page addresses the journal three-review three-proof and editorial quality scenario. The materials come from three types of editorial materials: three-review forms, proofs, and proofreading records. They are updated weekly or monthly, and the fields cover manuscript name, review round, proof round, date, quantity, signature, and receipt. Three-review forms carry the review round and signature information for initial review, re-review, and final review; proofs carry the proof round, date, and change marks; proofreading records carry proofreading entries, word count, and receipts. When the three types of materials enter knowledge base retrieval, each serves as a candidate source. Retrieval is only responsible for recalling and ranking fragments related to progress records and statistical criteria; it does not generate progress tables or statistical results. The update cadence aligns with the weekly and monthly report cycles, making it convenient to use one week's records to verify the totals.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), outputting candidate fragments and similarity scores (after pattern, merge, rerank, filter, and limit trimming). The "progress table and statistical criteria description" is completed by subsequent nodes or manually.
| Object | Check | Verification action |
|---|---|---|
| Three-review form | Whether the review round, date, and signature fields are consistent with the proofs and proofreading records, and whether the same manuscript name and quantity can be matched across the two materials | In the search test, use synonymous questions to retrieve the three-review form, record the number of candidates and scores, and verify the review round, date, and signature fields |
| Proof | Whether the proof round, date, and number of changes are synchronized with the proofreading records, and whether the version is unsynchronized | Switch searchMode to semantic, full-text, or hybrid, and compare whether the proof candidates include both semantic and full-text results |
| Proofreading record | Whether the proofreading entries, word count, and receipts are complete, and whether missing signatures or receipts can be located in the candidates | Enable and disable rerank, compare the order of the first few items, and use similarity and limit to observe the candidates after filtering and trimming |
| Progress table and statistical criteria description | Whether the record count matches the statistical count, whether the criteria are written in the header, and whether exceptions are clearly stated | Use one week's records to verify the totals, and check whether the header criteria and exception descriptions are consistent with the candidate fragments |
Three common errors
- Phenomenon: The name, quantity, and date of the same record do not match between the three-review form and the proofreading record. Cause: The two materials are entered separately, without field alignment and version synchronization.
- Phenomenon: The proof version is not synchronized, and old proof rounds and new proof rounds are mixed in the same batch of candidates. Cause: The update cadence is inconsistent, and retrieval hits fragments from multiple versions.
- Phenomenon: The proofreading record lacks a signature or receipt, and the statistical totals do not match the record count. Cause: Receipts are not archived, and the criteria do not state exceptions in the header.
How to verify the setup
This feature's acceptance: test once each with synonymously rewritten questions, compare candidates and scores, then enable and disable rerank to see the ordering of the first few items; switch searchMode to observe differences between semantic and full-text candidates, and record the number of candidates and scores under similarity and limit.
Business test: use one week's records to verify the totals, and see whether the criteria and exceptions are clearly stated; the number of citations is the array length, the citation limit verifies the tokens of the cited text, and the array item count is recorded separately.
- In the knowledge base search test, use synonymous questions to retrieve the three-review form, proofs, and proofreading records, record the number of candidates and scores, and verify the name, quantity, date, and signature fields
- Enable and disable rerank, compare the order of the first few items; adjust
similarityandlimit, and observe whether the candidate content after filtering and trimming retains key fragments