What this category's data looks like
This page addresses the kindergarten play activity organization scenario. Materials come from class observation records, area activity plans, material placement ledgers, and weekly reflection forms, with an update cadence of daily records, weekly plans, and monthly summaries. When these materials enter the document parsing and chunking feature, each plays a different role: observation records provide evidence of children's behavior and goal attainment, area plans provide activity goals and material presets, material placement ledgers provide quantity and date baselines, and reflection forms provide clues for next week's adjustments. Field characteristics include activity name, material name, placement quantity, activity date, goal description, and reflection points. The name, quantity, and date of the same object in two materials must be consistent. Discrepancy localization means finding fields inconsistent with the ledger, forming a discrepancy list for the reviewer to confirm and write back.
Configuration and implementation checks
This step is executed by the parsing and training task (server-side task), outputting data rows and index status (parsed, indexing, indexed). The discrepancy list is then completed by retrieval and subsequent processing. After parsing is complete, parsing and chunking immediately persist to the database as visible data rows and bind a unique data identifier. The index transitions through three states, and data pending indexing does not participate in recall.
| Object | Check | Verification action |
|---|---|---|
| Observation records correspond to area plans | The same activity name and date are consistent in both materials, and the goal matches the observed behavior | Check the file records and data rows separately, and check whether the activity name falls in the same data row after chunking by chunkSize and chunkSplitter, and whether the status is parsed / indexing / indexed |
| Material placement quantity | The material name and placement quantity are consistent between the ledger and the plan, with no material shortage | Check the data rows of the material placement ledger and the area plan, record the quantity discrepancy fields, and confirm the processing mode corresponding to trainingType |
| Activity goal attainment | The behavioral evidence in the observation records can support the activity goal, with no mismatch between goal and observation | Compare the goal description with the observed behavior one by one, mark the mismatched fields, and write them into the discrepancy list |
| Reflection converted to next week's plan | The reflection points have been transferred to next week's plan, and places where the plan has not been updated can be located | Check the data rows of the reflection form and next week's plan, and check the plan update date and material adjustment items |
Three common errors
- Symptom: newly uploaded materials cannot be found. Cause: successful parsing does not mean retrieval will definitely hit; the index status must also be checked, and data pending indexing does not participate in recall.
- Symptom: material shortage is not reflected in the plan. Cause: the material placement quantity and the area plan were not checked in sync, and the discrepancy fields were not written into the discrepancy list.
- Symptom: a mismatch between goal and observation is still carried forward. Cause: the correspondence between observation records and area plans was not compared one by one, and evidence of goal attainment is missing.
How to verify the setup
This feature's acceptance: upload one material and observe the data row status transitions and the coverage scope of the rebuild task; the status should be parsed / indexing / indexed, the rebuild should only cover data whose indexing is completed, and the initial training covers data that is pending indexing and indexing.
Business test: use one real record to find fields inconsistent with the ledger, form a discrepancy list for the reviewer to confirm and write back.
- Open the data rows to check the status, confirm that the three states are consistent with the interface display, and check the file records and data rows separately
- View the data rows covered by the rebuild task, confirm that only data whose indexing is completed is covered, and record the discrepancy fields