What this category's data looks like
Admissions enrollment and filing materials for regular higher education institutions mainly consist of enrollment plans, filing rules, and admission rosters. The sources are provincial admissions offices and higher education institution admissions departments, updated by batch and round. Fields include institution code, major group, planned number, filing cutoff score, admission date, and signature column. In multilingual materials and terminology handling, the enrollment plan serves as the source of the planned scope, the filing rules serve as the source of rule clauses, and the admission roster serves as the source of result records. These three types of materials enter the document parsing and chunking stage, respectively assuming the roles of scope, rules, and results, for subsequent retrieval and terminology comparison.
Configuration and implementation checks
Parsing and training tasks (server-side tasks) only output data rows and index status (parsed, indexing, indexed). The unified terminology table and comparison draft must be handled next 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. Data pending indexing does not participate in recall. Changing the model only rebuilds data that has completed indexing.
| Object | Check | Verification action |
|---|---|---|
| Enrollment plan | Institution code, major group, planned number, date, and signature are consistent across language versions | Check the identity of the file record and the data row separately, inspect chunk boundaries according to chunkSize and chunkSplitter, and confirm that enrollment plan fields have not been truncated |
| Filing rules | The translated names of rule clauses can refer back to the original text, and the same term is written consistently throughout | Check the rule clause chunks under the processing mode corresponding to trainingType, and sample three terms to check whether their writing is consistent throughout |
| Admission roster | When the name, quantity, or date of the same record does not match between two materials, it is necessary to refer back to the original text for verification | Open the data row to check the status, and confirm that the roster record is in one of parsed / indexing / indexed and is consistent with the interface display |
| Data row | It is in one of the three states and is consistent with the interface display; the rebuild scope only covers data that has completed indexing | Look at the data rows covered by the rebuild task, and distinguish data pending indexing from data that has completed indexing |
Three common errors
- Phenomenon: The name, quantity, or date of the same record does not match between the enrollment plan and the admission roster. Cause: After multilingual materials are chunked, fields are not aligned, and the terminology table is not unified.
- Phenomenon: The filing rule version is not synchronized, and old-version clauses are still hit by retrieval. Cause: Changing the model for rebuild only covers data that has completed indexing, and data pending indexing is not included.
- Phenomenon: The admission roster lacks a signature or receipt, and the published draft cannot refer back to the original text. Cause: Successful parsing does not mean retrieval will definitely hit; the index status must also be checked.
How to verify the setup
Acceptance for this feature: Upload one material, observe the status transitions of data rows and the coverage scope of the rebuild task, confirm that the data row is in one of parsed / indexing / indexed, and check the actual values of chunkSize, chunkSplitter, and trainingType in chunking and training.
Business test: Sample three terms to check whether their writing is consistent throughout
- Open the data row to check the status, and confirm that the three states are consistent with the interface display
- Check the file record and the data row separately, and confirm that their identities can be distinguished