What this category's data looks like
Materials for the establishment and annual reports of Chinese-Foreign cooperative education institutions include establishment applications, cooperation agreements, articles of association, annual report statements, and receipts. Sources are approval systems and institution archives, updated according to academic years and approval milestones. Field characteristics: institution name, education level, enrollment scale, cooperating party, signing date, signature and seal, version number. When materials enter the indexing service and vector database, the establishment application provides the establishment basis and original field values, the cooperation agreement provides the rights and obligations of both parties and dates, the articles of association provide the governance structure and signature page, and the annual report provides annual data and receipts. Each material is stored as a data row, and the same record may correspond to multiple vector identifiers. During retrieval, vectors are first recalled and then merged back into the original data.
Configuration and implementation checks
This step is executed by the indexing service and vector database, outputting vectors and index status (whether recall is possible, whether data in the same group is merged). "Basis entries that can be traced back to the original text" are subsequently completed by AI nodes, code nodes, or manual work (retrieval and citation trimming only take supporting materials).
| Object | Check | Verification action |
|---|---|---|
| Establishment application | Fields include institution name, education level, enrollment scale, cooperating party, signing date, and signature and seal; the version must be current | Check that the name is consistent with the cooperation agreement and articles of association; check the date against the signature receipt; check that the index model is consistent with the knowledge base settings |
| Cooperation agreement | Fields include cooperating party name, rights and obligations clauses, signing date, and signature page; the version must be current | Check whether the name, quantity, and date of the same record match across the two materials; mark if the signature or receipt is missing |
| Articles of association | Fields include governance structure, decision-making procedures, and signature page; the version must be synchronized with the annual report | Check that the version of the articles of association is consistent with the version cited in the annual report; open the data row to check whether the index field contains multiple vector identifiers |
| Data row index field | Multiple vector identifiers for one piece of data are all present, and the status is indexed | Open a data row to check the index field; view the coverage of the rebuild task record; check the limit, and quoteMaxToken budgets |
Three common errors
- Symptom: Duplicate entries appear in retrieval results. Inspect the following: The name, quantity, and date of the same record do not match across the two materials, and the vectors are not merged back into the original data.
- Symptom: Recall quality drops after changing models. Inspect the following: The version is not synchronized, and data rows that have not completed indexing do not participate in recall; indexing must be completed first and then rebuilt.
- Symptom: A data row has content but cannot be retrieved. Inspect the following: Materials missing signatures or receipts have not completed indexing, the status is not indexed, and the index field is empty.
How to verify the setup
Acceptance of this feature: sample a data row to check the index field, then retrieve once to see whether the same group is merged; check the limit, and quoteMaxToken budgets, and record the number of citations separately.
Business test: ask a basis question that spans materials and see whether the answer provides a traceable source.
- In the knowledge base settings, check the index model and ensure it is consistent with the vector model actually used by the knowledge base.
- View the data range covered by the rebuild task record; after changing models, trigger only for data that has completed indexing.