What this category's data looks like
This category's materials for pre-departure training and overseas safety reminders include three types: pre-departure training courseware, safety reminder checklists, and accommodation contact forms. The courseware contains fields for course name, teaching date, and instructor signature; the checklist contains fields for safety items, responsible department, and confirmation date; the contact form contains fields for accommodation address, contact person, emergency phone number, and receipt signature. The material sources are the project team, the international office, and accommodation institutions, and the update cadence is centralized revision before each session starts, temporary supplementation after overseas safety incidents, and fine-tuning as needed during the semester. After entering the document parsing and chunking function, the courseware is split into knowledge fragments as the main training material, the checklist is split item by item as safety entries, and the contact form is parsed row by row as a structured table; together the three support knowledge base retrieval and synchronization of external documents.
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). The "version ledger and change records" is then completed by retrieval and subsequent processing.
| Object | Check | Verification action |
|---|---|---|
| Pre-departure training courseware | The courseware name, teaching date, and instructor signature must be consistent across two materials for the same session, and the current version must be unique | Check the courseware record and the data row separately, confirm there are no missing items in the signature field and date field, and open the data row to check the status |
| Safety reminder checklist | Safety item entries must correspond one by one with the responsible department and confirmation date, and the old version must be traceable | Compare the checklist entries with the indexed data rows, and check whether the chunkSize and chunkSplitter chunking results cover all entries |
| Accommodation contact form | The accommodation address, contact person, emergency phone number, and receipt signature must be complete, and the name and quantity in the same record must be consistent | Check the identity of the table row and the data row, confirm the receipt signature field exists, and check the processing mode corresponding to trainingType |
| Data row status | It is in one of the three states and consistent with the interface display; data pending indexing does not participate in recall | Open the data row to check the status, and confirm that the parsed, indexing, and indexed transitions are consistent with the interface |
Three common errors
- Symptom: The name, quantity, and date of the same accommodation record do not match between the contact form and the courseware appendix. Cause: The two materials are maintained separately and the fields were not compared against the version ledger.
- Symptom: The checklist has been updated but the knowledge base still recalls old entries. Cause: The version was not synchronized, the old version was not marked invalid, and references were not updated with the version number.
- Symptom: The accommodation contact form lacks a receipt signature or confirmation date. Cause: The signature field was not validated during material collection, and it directly entered parsing and chunking.
How to verify the setup
Acceptance for this function: upload one material, observe the data row status transitions and the coverage scope of the rebuild task, confirm that the three states parsed, indexing, and indexed are consistent with the interface, and that the rebuild only covers data whose indexing has been completed.
Business test: change one piece of content and see whether the version number and references are updated synchronously
- Check the file record and the data row separately, confirm that the identifiers of the file and the data row are distinguishable, and that the data row status is consistent with the interface display
- View the data rows covered by the rebuild task, and confirm that it only targets data whose indexing has been completed; data pending indexing does not participate in the rebuild chain