What this category's data looks like
This page addresses the enrollment consultation conversion scenario for training institutions. Materials come from three categories: enrollment brochures, consultation records, and trial class schedules. Enrollment brochures provide course names, class types, start dates, fee standards, and signatures/seals; consultation records carry parent intent, communication times, follow-up staff signatures, and acknowledgments; trial class schedules record trial times, teaching instructors, attendance numbers, and confirmation signatures. Materials enter the knowledge base as files, and after parsing and chunking they each become data rows, respectively serving the roles of course basis, communication basis, and trial basis. The update rhythm rolls with enrollment batches and trial schedules. Field characteristics are the cross-occurrence of four types: names, quantities, dates, and signatures. The same record is often registered repeatedly across two materials, and currently valid entries must be distinguished by version and signature status.
Configuration and implementation checks
Parsing and training tasks (server-side tasks) only output data rows and index status (parsed, indexing, indexed). Basis entries that can be traced back to the original text must be handled by retrieval and subsequent processing. After parsing is completed, parsed chunks are immediately persisted as visible data rows and bound to a unique data identifier. The index transitions through three states, and data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Enrollment brochure | Whether the course name, class type, start date, fee standard, and signature/seal are consistent with the current version, and whether the name, quantity, and date of the same record can be matched between the brochure and the consultation record | Open the data row to check whether the status is parsed, indexing, or indexed, and separately verify the identity identifiers of the file record and the data row |
| Consultation record | Whether parent intent, communication time, follow-up staff signature, and acknowledgment are complete, and whether the date and number of people for the same record are consistent between the consultation record and the trial class schedule | Check the training mode by trainingType, and confirm that the entry can be traced back to a specific material paragraph |
| Trial class schedule | Whether the trial time, teaching instructor, attendance number, and confirmation signature are complete, and whether entries with unsynchronized versions or missing signature acknowledgments can be identified | Verify whether the data rows covered by the rebuild task fall only on data with completed indexing, and check the chunk boundaries corresponding to chunkSize and chunkSplitter |
| Data rows and index status | Whether the data row is in one of the three states and consistent with the interface display, and whether data pending indexing is not recalled by retrieval | Look at the data rows covered by the rebuild task, and confirm that changing the model only triggers rebuilds for data with completed indexing |
Three common errors
- Symptom: The name, quantity, and date of the same record do not match between the enrollment brochure and the consultation record. Cause: The two materials are registered separately and versions are not synchronized, lacking a unified traceback field.
- Symptom: The trial class schedule lacks a signature or acknowledgment, and the basis for attendance cannot be confirmed when replying to parents. Cause: Material versions are not synchronized, the old version overwrites the new version, and the signature status is not retained.
- Symptom: Newly uploaded materials cannot be found. Cause: The data row is in parsed or indexing status and indexing has not yet completed, so the retrieval layer naturally isolates it and it is not recalled.
How to verify the setup
Feature acceptance: Upload one material, observe the data row status transitions and the coverage scope of the rebuild task, confirm that the status correctly switches among parsed, indexing, and indexed, and that rebuilds only cover data with completed indexing.
Business test: Ask a basis question that spans materials, see whether the answer provides a traceable source, and verify that the number of citations and the tokens of the cited text are recorded separately.
- Open the data row to check whether the status is consistent with the interface display, and separately verify the identity identifiers of the file record and the data row.
- View the data rows covered by the rebuild task, and confirm that changing the model only rebuilds data with completed indexing, and that data pending indexing does not participate in the rebuild chain.