What this category's data looks like
This page addresses the research ethics review scenario. The material sources are ethics review applications, informed consent form templates, risk assessment forms, and their multilingual counterparts, with an update cadence that rolls with review batches and version revisions. Field characteristics center on subject ID, research project name, material version number, signing date, and signature and receipt columns. When materials enter the parsing and chunking function, the ethics review application carries the role of holding project basic information and review elements, the informed consent form template carries the role of holding multilingual clauses and terminology counterparts, and the risk assessment form carries the role of holding risk levels and disposition records. The three each bind to independent data rows and file records, and immediately after parsing is completed they are persisted as visible data rows for subsequent retrieval and terminology handling.
Configuration and implementation checks
Parsing and training tasks (server-side tasks) only output data rows and index status (parsed, indexing, indexed); the unified terminology table and counterparts must be handled next by retrieval and subsequent processing. After parsing is completed, parsing and chunking immediately persist as visible data rows and bind a unique data identifier; the index transitions through three states, data pending indexing does not participate in recall, and changing models only rebuilds data whose indexing is completed.
| Object | Check | Verification action |
|---|---|---|
| Ethics review application | Subject ID, project name, version number, signing date, and signature column are written consistently in the multilingual counterparts, and can refer back to the corresponding fields in the original text | Open the data row and check whether the status is parsed, indexing, or indexed; check the file record and the data row respectively; compare the names, quantities, and dates in the two materials |
| Informed consent form template | Terminology translations in clauses of each language are consistent throughout; the version number and signing date correspond to the application materials; the signature and receipt columns have no missing items | Use chunkSize and chunkSplitter to check whether chunk boundaries cut off terminology; sample three terms and check whether their writing is consistent throughout |
| Risk assessment form | Risk level, disposition records, and dates correspond across the three materials; versions are synchronized and signatures and receipts are complete | Check the processing mode corresponding to trainingType; confirm that the rebuild task only covers data rows whose indexing is completed |
Three common errors
- Symptom: the same subject ID is written differently in the ethics review application and the informed consent form template. Cause: multilingual material terminology is not unified, and after chunking each is independently persisted, lacking a counterpart that refers back to the original text.
- Symptom: the risk assessment form version number is not synchronized with the application material version, and the dates do not match. Cause: materials are uploaded in batches, and old-version data rows are still in the parsed state and have not been replaced.
- Symptom: the informed consent form template is missing the signature or receipt column, yet retrieval can still hit it. Cause: successful parsing does not equal a guaranteed retrieval hit; index status and field completeness are not verified in sync.
How to verify the setup
This feature's acceptance: upload one material, observe the data row status transitions and the coverage scope of the rebuild task, and confirm the statuses are parsed, indexing, and indexed in order, and that the rebuild only covers data rows whose indexing is completed.
Business test: sample three terms and check whether their writing is consistent throughout, and verify that the translations can refer back to the corresponding fields in the original text.
- Open the data row and check the status; check the file record and the data row respectively; confirm that the two identities are distinguishable and that the status is consistent with what the interface displays.
- View the data rows covered by the rebuild task, and confirm that it targets only data whose indexing is completed; the initial training covers data that is pending indexing and indexing.