What this category's data looks like
Annual progress and acceptance of research projects involve three types of materials: annual progress reports, mid-term review materials, and acceptance reports. The sources are documents and attachments submitted by the project team at milestones, and the update cadence rolls with the milestones. Field characteristics include project name, undertaking unit, person in charge, start and end dates, funding, indicators, list of achievements, signatures, and receipts. In the multilingual materials and terminology handling scenario, annual progress reports provide stage-specific terminology usage, mid-term review materials supplement revised translated names, and acceptance reports confirm the final reference draft. When the three types of materials enter this feature, each plays the role of terminology source, comparison baseline, and finalization basis respectively, and the vector model and indexing are responsible for making the same term recallable and mergeable across materials in different languages and different versions.
Configuration and implementation checks
This step is executed by the indexing service and the vector database, and outputs vectors and index status (whether recall is possible, whether data in the same group is merged). "Unified terminology glossary and reference draft" is then completed by subsequent AI nodes, code nodes, or manual work (retrieval and citation trimming only take supporting materials).
| Object | Check | Verification action |
|---|---|---|
| Annual progress report | Project name, person in charge, start and end dates, funding, and indicators must be aligned item by item with the mid-term review materials and acceptance report; the name, quantity, and date of the same record must not appear in two sets of wording | In the knowledge base settings, check that the indexing model is consistent with the vector model actually used by the knowledge base; open a data row and check whether the index field records multiple vector identifiers; view the data scope covered by the rebuild task |
| Mid-term review materials | Term translated names must be able to refer back to the source text, and the same term must be written consistently throughout; when versions are not synchronized, the latest signed version prevails | Sample a data row to look at the index field, then search once to see whether the same group is merged; check the budget of limit in the cited text; confirm that the status value is indexed or that a rebuild task record exists |
| Acceptance report | Signatures and receipts are complete, and the quantities and dates in the list of achievements are consistent with the annual progress report and mid-term review materials; missing signatures or receipts are regarded as incomplete | Search for the recall of the same term in materials in different languages; check whether data in the same group is merged and whether the vector score takes the highest score; view whether the rebuild task record covers data that has completed indexing |
| Unified terminology glossary and reference draft | The same term is consistent throughout, and the translated name can refer back to the source text; sample three terms to check whether the wording is consistent throughout | The terminology glossary and reference draft are completed by subsequent AI nodes, code nodes, or manual work; this feature only provides recallable vectors and index status as supporting materials |
Three common errors
- Symptom: the name, quantity, and date of the same record do not match between the annual progress report and the acceptance report. Inspect the following: the two materials each maintain their own fields, versions are not synchronized, and a unified terminology glossary and reference draft are lacking.
- Symptom: duplicate entries appear in search results. Inspect the following: compare the relevant source fields, document versions and downstream processing records.
- Symptom: after changing the model, recall quality drops, and a data row has content but cannot be retrieved. Inspect the following: rebuild is triggered only for data that has completed indexing; data rows not covered have not completed indexing and do not participate in recall.
How to verify the setup
Feature acceptance: sample a data row to see whether the index field records multiple vector identifiers, then search once to see whether the same group is merged and whether the vector score takes the highest score; check defaultToken, maxToken, the status value indexed, and the rebuild task record.
Business test: sample three terms to check whether the wording is consistent throughout, confirm that the translated names can refer back to the source text, and that signatures and receipts are complete.
- In the knowledge base settings, check that the indexing model is consistent with the vector model actually used by the knowledge base, and view the data scope covered by the rebuild task.
- Open a data row and check the index field to confirm that multiple vector identifiers are all present, and after retrieval that data in the same group is merged and takes the highest score.