What this category's data looks like
This page is for modern apprenticeship and order class management scenarios. The materials include three types: school-enterprise cooperation agreements, order class training plans, and master apprenticeship records. School-enterprise cooperation agreements include the cooperating enterprise name, position, number of people, start and end dates, and signatures of both parties; order class training plans include courses, semesters, lists of enterprise masters and school teachers, and assessment methods; master apprenticeship records include apprenticeship dates, processes, work hours, and master-signed receipts. The three types of materials are updated by semester or project milestone. It is common for the same record to have inconsistent names, quantities, or dates across two materials, unsynchronized versions, and missing signatures or receipts. When entering this function, cooperation agreements and training plans serve as basis sources, and apprenticeship records serve as process evidence. They are split into data rows respectively, and the vector model generates vectors and index fields to support subsequent retrieval and citation trimming for supporting materials.
Configuration and implementation checks
This step is executed by the index service and vector database, outputting vectors and index status (whether recall is possible and whether data in the same group is merged). "Basis entries that can be traced back to the original text" are then completed by subsequent AI nodes, code nodes, or manual work (retrieval and citation trimming only take supporting materials).
| Object | Check | Verification action |
|---|---|---|
| Index model | Consistent with the vector model actually used by the knowledge base; after changing the model, only trigger rebuild task records for data that has already completed indexing | Check the index model in the knowledge base settings and view the data scope covered by the rebuild task |
| Index fields of the data row | Multiple vector identifiers for one piece of data are all present; limit constrains the token budget of retrieved reference text, and do not change the number of index fields | Open a data row to check the index fields and confirm the status is indexed |
| School-enterprise cooperation agreement | Fields include the cooperating enterprise name, position, number of people, start and end dates, and signatures of both parties; dates and signatures must match the order class training plan | Check the enterprise name, number of people, start and end dates, and signature page of each agreement against the training plan |
| Master apprenticeship record | Fields include apprenticeship date, process, work hours, and master-signed receipt; dates and number of people must match the training plan | Sample one apprenticeship record and compare the date, number of people, and signed receipt with the training plan |
Three common errors
- Symptom: duplicate entries appear in retrieval results. Inspect the following: the same group of data corresponds to multiple vectors; after recall, they are not merged back to the original data by index field, and the score is not taken as the highest.
- Symptom: recall quality drops after changing the model. Inspect the following: rebuild task; data rows that have not completed indexing do not participate in recall, so indexing must be completed before rebuilding.
- Symptom: a data row has content but cannot be retrieved. Inspect the following: the index field is missing or the status has not reached indexed, the vector has not been written to the vector database, and it cannot be obtained during the recall stage.
How to verify the setup
Acceptance for this function: sample one data row and check the index fields, confirming that multiple vector identifiers are all present and the status is indexed; then retrieve once and see whether the same group of data is merged and whether the score is taken as the highest.
Business test: ask a basis question across materials, for example whether the number of people in an order class and the apprenticeship date are consistent with the agreement, and see whether the answer gives a traceable source.
- Check the index model in the knowledge base settings and view the data scope covered by the rebuild task records, confirming that after changing the model, rebuild is triggered only for data that has already completed indexing.
- Open the data row to check the index fields and status, confirming that
limitconstrains the token budget of retrieved reference text, and that the number of citations and citation text tokens are recorded separately.