What this category's data looks like
This page addresses the vocational training subsidy application scenario. Materials include training class filing forms, attendance records, completion assessment forms, and subsidy application receipts. These materials are submitted by training institutions in batches before class starts, during training, and after completion, with an update cadence that rolls with class sessions. Fields center on trainee name, ID number, training occupation, class start and completion dates, number of class hours, assessment score, signature, and receipt number. When the materials enter the document parsing and chunking feature, each plays a different role: the class filing form provides the subject and course basis, attendance records provide attendance facts, completion assessment forms provide the result basis, and receipts provide submission proof. After parsing and chunking, each material becomes a visible data row bound to a unique data identifier for subsequent retrieval and comparison.
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. After parsing is complete, parsing and chunking immediately persist to the database as visible data rows and bind a unique data identifier. The index transitions through three states, and data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Training class filing form | Fields include institution name, training occupation, class start date, trainee list, and signature; the date and signature must match the attendance records and completion assessment form | Check whether the number of trainees and class start date in the filing form and attendance records are consistent, and check whether signatures and receipts are complete; confirm the data row status is parsed, indexing, or indexed |
| Attendance records | Fields include attendance date, trainee name, number of class hours, and attendance signature; the name, quantity, and date in the same record across different materials must be consistent | Compare trainee names, class hours, and completion dates in the attendance records and completion assessment form; check for missing signatures or receipts; check whether the data row status matches the interface display |
| Completion assessment form | Fields include assessment date, score, assessment signature, and completion conclusion; the version must be current, and the date and signature must correspond to the filing form and attendance records | Check whether the assessment form version is current, and check whether the same trainee's name, quantity, and date in the two materials do not match; confirm that the file record and data row identifier are distinguishable |
| Data rows and index status | It is in one of the three states and matches the interface display; the rebuild scope only covers data that has completed indexing | Open the data row to check the status, and view the data rows covered by the rebuild task; check whether the configuration of chunkSize, chunkSplitter, and trainingType matches the actual chunking result |
Three common errors
- Symptom: The same trainee's name, class hours, and date do not match between the attendance records and the completion assessment form. Cause: The two materials were filled in at different stages, and no field-level comparison or traceback check was performed.
- Symptom: The filing form version was not synchronized, and the old class start date conflicts with the current attendance records. Cause: After the materials were updated, parsing and chunking were not redone, and the data row still points to the old version content.
- Symptom: The completion assessment form is missing the assessment signature or the subsidy application receipt is missing. Cause: When the materials were submitted, the completeness of signatures and receipts was not checked, and after parsing, missing items were not marked in the data row.
How to verify the setup
Feature acceptance: Upload one material and observe the data row status transitions and the coverage scope of the rebuild task. The data row should first become a visible data row and be bound to a unique data identifier, with status transitioning through parsed, indexing, and indexed; when switching models, the rebuild only covers data that has completed indexing, and data pending indexing does not participate in the rebuild. Check whether chunkSize, chunkSplitter, and trainingType match the actual chunking result.
Business test: Ask a cross-material basis question, for example whether a trainee's completion date corresponds to the class start date in the filing form, and see whether the answer provides a traceable source, and check whether the source can be traced back to the original text entry.
- Open the data row to check whether the status is parsed, indexing, or indexed, and compare it with the interface display; separately check the identifiers of the file record and the data row to confirm that the two identities are distinguishable.
- View the data rows covered by the rebuild task, confirm that only data that has completed indexing is covered; check the configuration of
chunkSize,chunkSplitter, andtrainingType, and check that the tokens of the referenced text and the number of array entries are recorded separately.