What this category's data looks like
The progress records and statistical verification of peer review organizations draw on three types of records: review invitations, expert review comments, and author revision notes, updated weekly and archived monthly. Review invitations carry the invitation sent and reply dates, expert name and affiliation, and invitation status; expert review comments carry the comment body, review completion date, and signature; author revision notes carry revision items, revision dates, and receipts. When these three types of materials enter the database and operations function, each plays a different role: review invitations serve as the starting point of progress, expert review comments serve as the basis for progress completion, and author revision notes serve as the closing evidence of progress. Field characteristics center on four types: names, quantities, dates, and signatures. The same record must remain cross-referenceable across the three materials, and the latest archive is authoritative for versions. Statistical criteria are written in the table header, and exceptional cases are listed separately in notes, so they can be cited directly in weekly or monthly reports.
Configuration and implementation checks
This step is performed by the connection configuration of the deployment environment (server side), not by the capability execution of a model node, and it outputs whether the connection is available (service startup result and connection logs). The "progress table and statistical criteria description" is then completed by the deployment handover and subsequent business processing. Connection addresses are configured via environment variables: the main business database, log database, cache, and vector database can point to different instances respectively; if the log database is not configured separately, it reuses the main database.
| Object | Check | Verification action |
|---|---|---|
| Review invitation | Whether the invitation sent date, reply date, invitation status, and expert signature are complete, and whether they are checked item by item against the dates in the expert review comments | In the deployment environment, verify that the value of the main business database connection variable matches the database address, connect once to the cache with a client, and check the connection logs |
| Expert review comments | Whether the comment body, review completion date, and signature match the item names, quantities, and dates in the author revision notes | Against the deployment instructions, verify the selected vector database and address, check the database to which logs are written, and confirm whether the log database is configured separately |
| Author revision notes | Whether the revision items, revision dates, and receipts are complete, and whether the version is synchronized with the latest archive | Summarize the record count weekly, check it item by item against the statistical count, and write exceptional cases into the criteria table header |
| Progress table and statistical criteria description | Whether the table header states the statistical criteria and exceptions, and whether the record count and statistical count are consistent | Use one week of records to check the total, verify the tokens of the quoted text, and record the array length separately |
Three common errors
- Symptom: the expert name and date for the same review record do not match between the review invitation and the expert review comments. Cause: the two materials are registered separately, are not merged by the same record key, and versions are not synchronized.
- Symptom: the number of items in the author revision notes is inconsistent with the number of items listed in the expert review comments. Cause: revision notes are submitted in batches, old versions are not replaced, and receipts or signatures are missing.
- Symptom: after startup, the connection logs continuously report connection timeouts, and retrieval reports vector database errors. Cause: the instance pointed to by the environment variable is unreachable, or the log database and main database are shared but were not left blank for reuse as instructed.
How to verify the setup
This function's acceptance: connect once to the main business database, log database, cache, and vector database item by item, check the connection results in the startup logs, and verify that the three values MONGODB_URI, REDIS_URL, and PG_URL match the deployment instructions.
Business test: use one week of records to check the total, see whether the criteria and exceptions are clearly stated, and whether the record count and statistical count match.
- In the deployment environment, verify the values of the main business database and log database connection variables, and confirm that the database to which logs are written matches the instructions
- Connect once to the cache and vector database with a client, verify the returned results, and check the connection results in the startup logs