What this category's data looks like
This category covers pre-departure training and overseas safety reminders. Materials include pre-departure training courseware, safety reminder checklists, and accommodation contact lists. Sources are the international exchange department and the overseas affairs office. The update cadence follows semester batches and destination policy adjustments. Fields include course name, safety item, accommodation address, contact person, effective date, and signed receipt. After courseware enters parsing and chunking, it serves as the main basis data row. Safety reminder checklists are split by item and bound to a unique data identifier. Accommodation contact lists are stored as visible data rows by table row. During retrieval, the three types of materials each serve as a basis source role. The name, quantity, and date of the same record must match across materials. The version and signed receipt determine whether it is currently valid.
Configuration and implementation checks
This step is executed by parsing and training tasks (server-side tasks), outputting data rows and index status (parsed, indexing, indexed). Basis items that can be traced back to the original text are then completed by retrieval and subsequent processing. Successful parsing does not mean retrieval will definitely hit; it also depends on the index status.
| Object | Check | Verification action |
|---|---|---|
| Pre-departure training courseware | Fields include course name, batch, effective date, and signature; the name and date of the same record in two materials must match | Check the chunkSize and chunkSplitter chunking results of the courseware data row, confirm the status is parsed, indexing, or indexed, and trace back to the original text paragraph |
| Safety reminder checklist | Safety items, quantity, and version must be consistent with the current materials; missing signature or receipt is considered not effective | Check the checklist data row status and original item number, confirm the processing mode corresponding to trainingType, and check whether data pending indexing is isolated |
| Accommodation contact list | Accommodation address, contact person, and date fields must match the courseware and checklist; if the version is not synchronized, it cannot be used as a basis | Check the file record and data row identifier separately, confirm that rebuild only covers data that has completed indexing, and view the data rows covered by the rebuild task |
Three common errors
- Symptom: The name, quantity, and date of the same accommodation record do not match between the courseware and the contact list. Cause: The two materials are parsed and stored separately, without cross-material field alignment and version comparison.
- Symptom: The safety reminder checklist has been updated, but retrieval still returns old items. Cause: The version is not synchronized, data pending indexing does not participate in recall, and the old data row is still in the completed indexing state.
- Symptom: Pre-departure training courseware is still cited as a basis even though it lacks a signature or receipt. Cause: The signature field is not included in the data row check, and successful parsing is mistakenly regarded as directly usable as current material.
How to verify the setup
This feature acceptance: Upload one material, observe the data row status transitions and the coverage scope of the rebuild task, and confirm that the status only changes among parsed, indexing, and indexed, and that rebuild only targets data that has completed indexing.
Business test: Ask a cross-material basis question, see whether the answer provides a traceable source, and check that the citation count and the tokens budget of the cited text are recorded separately.
- Open the data row to check the status, confirm it is consistent with the interface display, and check the identifiers of the file record and data row separately.
- View the data rows covered by the rebuild task, confirm that the initial training covers data pending indexing and data being indexed, and that changing the model only rebuilds data that has completed indexing.