What this category's data looks like
This page focuses on authorized access and external services for compulsory education dropout prevention and retention. The materials come from three types of ledger documents: dropout risk ledgers, persuasion-and-return records, and home visit records. The update cadence is per semester, with rolling supplementary entries as persuasion-and-return and home visit progress advances. Field characteristics center on student name, grade and class, risk level, persuasion-and-return date, home visit date, and signature and receipt columns. When the materials enter the knowledge base retrieval and recall function, the ledger serves as the semantic source for risk clues and school attendance status, the persuasion-and-return records serve as the semantic source for the handling process and responsible parties, and the home visit records serve as the semantic source for household visit frequency and parent feedback. Together, the three form a pool of candidate snippets that can be retrieved, for subsequent nodes or humans to generate the external entry point and authorization scope accordingly.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), outputting candidate snippets and similarity scores (after pattern, merge, rerank, filter, and limit trimming). The external entry point and authorization scope are then completed by subsequent nodes or humans.
| Object | Check | Verification action |
|---|---|---|
| Dropout risk ledger | Check whether student name, grade and class, risk level, persuasion-and-return date, and signature column are consistent across the three types of materials | In the search test, search once each by name and by grade, compare the names, quantities, and dates in the candidate snippets, and record the number of candidates and scores |
| Persuasion-and-return records | Check whether the persuasion-and-return date, responsible party, receipt status, and home visit records match up | Switch searchMode to semantic, full-text, and hybrid once each, see whether the candidates include both semantic and full-text results, and verify the receipt and signature |
| Home visit records | Check whether the home visit date, household visit frequency, parent feedback, and persuasion-and-return records are version-synchronized | Test once each with rerank enabled and disabled, compare the order of the top few entries, record the number of candidates and scores, and confirm that key snippets have not been trimmed out |
| External entry point and authorization scope | Check whether the visitor identity matches the visible scope, and that unauthorized access is not visible | Access once with a restricted identity, see whether the visible scope matches the setting, and verify the tokens and array count of the quoted text and record them separately |
Three common errors
- Symptom: The same student's name, quantities, and dates do not match between the ledger and the persuasion-and-return records. Cause: The two materials are supplemented separately, lacking student-keyed field alignment and receipt verification.
- Symptom: The home visit records have been updated, but the persuasion-and-return records remain at the old version. Cause: Versions are not synchronized, retrieval candidates hit both old and new snippets at the same time, and the higher-ranked one is not necessarily the latest version.
- Symptom: The persuasion-and-return records lack a signature or receipt, yet the home visit records show it as closed-loop. Cause: The signature and receipt columns are not backfilled along with the process, and the material status is disconnected from the handling status.
How to verify the setup
This function's acceptance: test once each with synonym-rewritten questions, compare candidates and scores, then toggle rerank on and off to see the ordering of the top few entries; at the same time record the number of candidates and scores, and verify the values of similarity, limit, and usingReRank in the test.
Business test: access once with a restricted identity, and see whether the visible scope matches the setting.
- In the knowledge base search test, search once each by student name and by grade, record the number of candidates and scores, and compare the names, quantities, and dates of the three types of materials.
- Switch
searchModeto semantic, full-text, and hybrid once each, and enable and disable rerank once each, to verify the impact ofsimilarityandlimiton the number of candidates and ordering.