What this category's data looks like
Enterprise compliance and security training materials come from three sources: compliance training courseware, security operating procedures, and exam records, updated quarterly or upon system revision. Courseware contains clause numbers and effective dates; procedures contain operating steps, responsible persons, and signature fields; exam records contain candidates, scores, dates, and receipts. When materials enter the document parsing and chunking function, courseware and procedures serve as basis sources and are split into retrievable chunks, while exam records are stored synchronously as supporting materials, each bound to a unique data identifier. Field characteristics center on names, quantities, dates, signatures, and version numbers. The same record often shows inconsistent names or dates across different materials, versions are not synchronized, or signature receipts are missing. After parsing, verification must be performed against data row status to ensure the basis comes from the institution’s materials and the version is current.
Configuration and implementation checks
This step is executed by the parsing and training task (server-side task), outputting data rows and index status (parsed, indexing, indexed). Basis entries that can be traced back to the original text are then completed by retrieval and subsequent processing.
| Object | Check | Verification action |
|---|---|---|
| Compliance training courseware | Clause numbers, effective dates, version numbers, and signature fields are complete and consistent with the current system | Open the data row to check whether the status is parsed or indexed, check whether the courseware name, clause count, and dates match another material, and verify the values of chunkSize and chunkSplitter |
| Security operating procedures | Operating steps, responsible persons, dates, and receipt signatures are complete, and the version is current | Verify the file record and the data row separately, confirm that the procedure version number matches the names, quantities, and dates in the exam records, mark any missing signature or receipt, and verify the trainingType field |
| Exam records | Candidates, scores, dates, and receipt signatures can be traced back to the original text and are synchronized with the courseware version | Verify the data row status transitions, confirm that the dates in the records do not conflict with the courseware effective dates, record missing receipts separately, and recognize only parsed, indexing, and indexed as statuses |
| Data row | Is in one of the three states parsed / indexing / indexed and consistent with the interface display | Check the data rows covered by the rebuild task, confirm that only indexed data is covered, that initial training covers pending-index and indexing data, and verify the chunkSize, chunkSplitter, and trainingType fields |
Three common errors
- Symptom: The names, quantities, and dates of the same record do not match between the courseware and the exam records. Cause: The two materials were entered independently, and no field-level comparison or trace-back verification was performed after parsing.
- Symptom: The procedure version is not synchronized, and old clauses are still treated as the current basis. Cause: Material updates did not trigger index rebuild, and pending-index data is mixed with completed-index data.
- Symptom: Exam records lack signatures or receipts, and basis entries cannot be traced back to the original text. Cause: Parsing only stores visible data rows and does not validate the completeness of signature fields and receipt fields.
How to verify the setup
This function's acceptance: upload one material, observe the data row status transitions and the coverage scope of the rebuild task, confirm that rebuild is triggered only for indexed data, that pending-index data does not participate in recall, and verify the three fields chunkSize, chunkSplitter, and trainingType.
Business test: ask a cross-material basis question, see whether the answer provides traceable sources, and verify whether the citation count and citation text tokens are recorded separately.
- Open the data row to check whether the status is parsed, indexing, or indexed, and whether it is consistent with the interface display.
- Verify the file record and the data row separately, confirm that the rebuild task covers only completed-index data, and that initial training covers pending-index and indexing data.