What this category's data looks like
Grade entry and credit recognition involves three types of materials: transcripts, deferred examination applications, and make-up examination arrangements. Sources are academic affairs system exports and departmental submissions. The update cadence enters in batches according to course selection, examination, and make-up examination milestones. Fields include course name, credits, grade, date, signature, and receipt. Transcripts play the role of original records, deferred examination applications play the role of approval basis, and make-up examination arrangements play the role of follow-up arrangements. After the three enter parsing and training tasks, they are first split into data rows and bound to unique data identifiers, then flow through the three index states, providing retrieval and subsequent processing to generate a unified terminology table and comparison draft. Multilingual materials and terminology processing require the same term to be consistent throughout the entire text and the translated name to be able to refer back to the original text. Document parsing and chunking determine where terms land in data rows. Chunking parameters such as chunkSize, chunkSplitter, and trainingType affect terminology segmentation and training mode selection.
Configuration and implementation checks
This step is executed by parsing and training tasks (server-side tasks), outputting data rows and index states (parsed, indexing, indexed). The "unified terminology table and comparison draft" is then completed by retrieval and subsequent processing. After parsing and chunking are completed, they are immediately persisted as visible data rows and bound to unique data identifiers. Data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Transcript | Fields, dates, and signatures must be aligned item by item with the deferred examination application and make-up examination arrangements; the same course name and credits must be written consistently across different materials | Open the data row to check whether the status is parsed, indexing, or indexed, and separately verify the file record and the data row identifier |
| Deferred examination application | The course name, quantity, and date in the approval basis must match the transcript and make-up examination arrangements; missing signature or receipt means inconsistency | Check the chunk landing points according to chunkSplitter, confirm that terms are not truncated, and inspect the data rows covered by the rebuild task |
| Make-up examination arrangements | The course, date, and signature in the follow-up arrangements must be consistent with the transcript and deferred examination application; when versions are not synchronized, the data with completed indexing shall prevail | Check the processing mode corresponding to trainingType, and confirm that changing the model only rebuilds data with completed indexing |
| Unified terminology table and comparison draft | The same term must be consistent throughout the entire text, and the translated name must be able to refer back to the original text; sample three terms to check whether their usage is consistent throughout the entire text | Check the terminology table and comparison draft in the retrieval and subsequent processing step, and confirm that the referenced text tokens and the number of array entries are recorded separately |
Three common errors
- Symptom: The course name, quantity, and date of the same record do not match between the transcript and the deferred examination application. Cause: The versions of the two materials are not synchronized, the terminology landing points differ after parsing and chunking, and no full-text consistency check was performed.
- Symptom: The make-up examination arrangements lack a signature or receipt, and grade entry and credit recognition cannot be closed-loop. Cause: Fields are missing when the materials enter parsing and training tasks; although the data rows have been persisted, the business requirements are incomplete.
- Symptom: Newly uploaded materials cannot be found, and empty data rows are left after training fails. Cause: Data pending indexing does not participate in recall, the index state has not reached indexed, and the rebuild scope only covers data with completed indexing.
How to verify the setup
Feature acceptance: Upload one material, observe the data row state transitions and the coverage scope of the rebuild task, confirm that the state transitions among parsed, indexing, and indexed, and that rebuild only covers data with completed indexing.
Business test: Sample three terms to check whether their usage is consistent throughout the entire text, and verify that the translated names of the same term in the transcript, deferred examination application, and make-up examination arrangements can refer back to the original text.
- Open the data row to check whether the status is consistent with the interface display, separately verify the file record and the data row identifier, and confirm that the unique data identifier is bound correctly.
- Inspect 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.