What this category's data looks like
Research ethics review materials center on ethics review applications, informed consent form templates, and risk assessment forms. Their sources are institutional ethics committees and research management departments, and their update cadence rolls forward with regulatory revisions, project changes, and annual continuing review. When these three types of materials enter the document parsing and chunking feature, their roles differ: the ethics review application carries the project name, review category, date, and signature fields; the informed consent form template carries the version number, revision date, and receipt/acknowledgment section; the risk assessment form carries the risk level, quantity, and signature confirmation. The same record often appears in two materials with inconsistent names, quantities, or dates, versions not synchronized, or missing signature receipts. Therefore, a version ledger and change records are needed to ensure the current version is unique, old versions are traceable, and references stay synchronized.
Configuration and implementation checks
This stage is executed by parsing and training tasks (server-side tasks), outputting data rows and index statuses (parsed, indexing, indexed). The "version ledger and change records" are then completed by retrieval and subsequent processing.
| Object | Check | Verification action |
|---|---|---|
| Ethics review application | The project name, review category, date, and signature fields must align item by item with the informed consent form template and the risk assessment form | Open the data row to check the status, verify the file record and the data row separately, confirm the trainingType and chunkSplitter configuration, then check reference synchronization |
| Informed consent form template | The version number, revision date, and receipt/acknowledgment section must correspond uniquely to the current version, and old versions must be traceable | Verify the identity of the file record and the data row, confirm chunkSize and chunkSplitter chunking, then check whether the version number and references are synchronized |
| Risk assessment form | The risk level, quantity, date, and signature must be consistent with the ethics review application, and no signature or receipt may be missing | View the data rows covered by the rebuild task, confirm that only data with completed indexing is covered, and that the status is parsed / indexing / indexed |
Three common errors
- Symptom: a newly uploaded material cannot be found. Cause: the data row is still in parsed or indexing, and data pending indexing is naturally isolated at the retrieval layer and is not recalled.
- Symptom: the name, quantity, or date of the same record does not match across two materials. Cause: versions are not synchronized, old versions are not traced, and references are not updated with changes.
- Symptom: empty data rows are left after a training failure. Cause: the initial training covers data pending indexing and data being indexed, while rebuild only targets data with completed indexing; the two scopes are not clearly distinguished.
How to verify the setup
This feature's acceptance: upload one material and observe the data row status transitions and the coverage scope of the rebuild task.
Business test: change one piece of content and see whether the version number and references are updated synchronously.
- Open the data row to check the status, and confirm it is in one of the three states parsed / indexing / indexed and is consistent with what the interface displays.
- Verify the file record and the data row separately, confirm that the identifiers of the file and the data row are distinguishable, and that the rebuild task only covers data with completed indexing.