What this category's data looks like
Kindergarten home-school co-education communication materials come from class notices, observation records, parent requests, and handover forms, with an update cadence that rolls daily or weekly. Field characteristics include committed items and response times, descriptions of child performance, categories of parent requests, notice delivery status, responsible person, and version date. When the materials enter the knowledge base retrieval service, each plays a different role: notices and observation records serve as candidate fragment sources, handover forms serve as the basis for responsibility fields, and parent requests serve as classification labels. The retrieval node and search test are responsible for recalling these fragments, and subsequent nodes or humans use them to draft issuable materials and handover records.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), which outputs candidate fragments and similarity scores (after pattern, merge, rerank, filter, and limit trimming). The issuable material drafts and handover records are then completed by subsequent nodes or humans.
| Object | Check | Verification action |
|---|---|---|
| Committed items and response times | Whether the commitment is responded to within the agreed time, and whether the response content corresponds to the original commitment | In the search test, use synonymous questions to retrieve the commitment field, switch searchMode to semantic, full-text, or hybrid, and compare the number of candidates and the similarity score |
| Descriptions of child performance | Whether the description is consistent with the observation record, and whether there is subjective exaggeration or omission of key behaviors | Retrieve once each with the original words and rewritten words from the observation record, compare the ranking of the top few items before and after enabling usingReRank, and record the number of candidates |
| Categories of parent requests | Whether requests are classified into categories such as notice delivery, activity arrangement, or individual communication, and whether the classification labels are complete | Retrieve by classification label, check whether the returned candidates include both semantic and full-text results, and verify the tokens of the quoted text after limit trimming |
| Handover records | Whether the responsible person, version, and date are complete, and whether the receipt field is clear | Retrieve by handover form number, confirm that the candidate fragments include the responsible person and date fields, record the number of array items, and separately verify the tokens of the quoted text |
Three common errors
- Symptom: committed items are not responded to in the materials, and are only supplemented after the parent follows up. Cause: when retrieving, the commitment field and response time were not used as filter conditions, so the candidate fragments lack the corresponding records.
- Symptom: descriptions of child performance do not match the observation records, and the wording is subjective. Cause: the recalled fragments come from different channels and were not reranked, so the top few rankings place weakly relevant descriptions first.
- Symptom: parent requests are not categorized, and responsibilities are unclear at handover. Cause: the classification labels did not enter the retrieval conditions, so the candidate fragments do not cover the request category field.
How to verify the setup
Acceptance for this feature: test once each with synonymously rewritten questions, compare the candidates and similarity scores, then toggle usingReRank on and off to see the ranking of the top few items; switch searchMode and record the number of candidates after limit trimming.
Business test: draft one according to the template and complete one handover, and see whether missing items are blocked.
- In the knowledge base search test, fix the question, switch
searchModeto semantic, full-text, and hybrid, and record the differences in the number of candidates and scores. - Enable and disable
usingReRank, compare the order of the top few candidates, and confirm whether relevant fragments move forward after reranking.