What this category's data looks like
In the scenario of signing research project task books, materials include three types: project task books, subject task breakdown tables, and budget books. The project task book provides research objectives, assessment indicators, and execution period, and is the main text for signing; the subject task breakdown table assigns tasks to subjects and responsible persons, with fields including subject name, task content, start and end dates, and responsible person signature; the budget book carries funding subjects, amounts, calculation basis, and approval signatures. Materials come from application system exports, offline signed scans, and revised drafts, with update rhythms concentrated in multiple rounds of countersigning before signing and change filing after signing. When the three types of materials enter document parsing and chunking, their roles differ: the task book serves as the primary data source, the breakdown table as a structured reference, and the budget book as a source for verifying amounts and subjects; all need to retain version and signature receipt clues.
Configuration and implementation checks
Parsing and training tasks (server-side tasks) only output data rows and index status (parsed, indexing, indexed); the version ledger and change records 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 the three states parsed, indexing, and indexed; data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Project task book | Whether fields, dates, and signatures are consistent with the breakdown table and budget book | Check the research objectives, execution period, and signature page; after adjusting chunking with chunkSize and chunkSplitter, see whether the data row status is parsed |
| Subject task breakdown table | Whether subject name, task content, start and end dates, and responsible person signature match the task book | Compare subject names and dates item by item; confirm that after changes the old version is traceable and references are synchronized; after data rows enter indexing, they do not participate in recall |
| Budget book | Whether funding subjects, amounts, calculation basis, and approval signatures match the task book and breakdown table | Check amounts and subjects; confirm signatures or receipts are complete; after using trainingType to distinguish processing modes, view indexed data rows |
| Version ledger and change records | The current version is unique, old versions are traceable, and references are synchronized | After changing one piece of content, see whether the version number and references are updated synchronously, and check that the rebuild task only covers data rows whose indexing has been completed |
Three common errors
- Symptom: The name, quantity, or date of the same subject does not match between the task book and the breakdown table. Cause: The two materials were revised separately, without field-level comparison and version synchronization.
- Symptom: Changes after signing are not synchronized, and the old version is still referenced. Cause: The version ledger was not updated, and reference relationships were not recalculated with the change records.
- Symptom: Signatures or receipts are missing, but the material has already entered parsing. Cause: Parsing only processes content; signature pages and receipts need separate verification and inclusion in the ledger.
How to verify the setup
Acceptance for this feature: Upload one material, observe the data row status transitions and the coverage scope of the rebuild task, and confirm that the three states parsed, indexing, and indexed are consistent with the interface display, and that rebuild only covers data rows whose indexing has been completed.
Business test: Change one piece of content and see whether the version number and references are updated synchronously.
- Open the data rows to check status, confirm that file records and data rows have separate identities, and record the values of
chunkSize,chunkSplitter, andtrainingType. - View the data rows covered by the rebuild task, confirm that it only targets data whose indexing has been completed, and that the initial training covers data that is pending indexing and indexing.