Knowledge Base Retrieval for Laboratory Hazardous Chemical Purchasing and Records: Coordinate Operations and Application Handover

Lab Chemicals: review retrieval for operational handover. Check source versions, input mappings and the values received by downstream nodes.

What this category's data looks like

This page centers on laboratory hazardous chemical procurement and ledgers. The materials include three categories: hazardous chemical procurement applications, usage registration, and inventory ledgers. Procurement applications record the product name, specification, quantity, application date, and approval signature; usage registration records the user, quantity used, date of use, and custodian signature; inventory ledgers record the balance quantity, inbound date, and inventory count signature. The three types of materials are updated by batch and shift, with name, quantity, date, and signature as common anchors across fields. When entering the knowledge base retrieval service, procurement applications and usage registration serve as candidate snippet sources, while inventory ledgers serve as cross-checking materials, each taking on the roles of recall and comparison.

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 handover-ready deployment status and operation records are completed by subsequent nodes or manually.

ObjectCheckVerification action
Hazardous chemical procurement applicationThe product name, quantity, application date, and approval signature are consistent across the three types of materialsSearch the product name in the search test, check the name, quantity, date, and signature fields of the candidate snippets, and record the number of candidates and the score
Usage registrationThe user, quantity used, date of use, and custodian signature can be matched against the inventory ledgerSwitch searchMode to semantic, full-text, or hybrid, compare whether the candidates include both semantic and full-text results, and record the number of candidates and the score
Inventory ledgerThe balance quantity, inbound date, and inventory count signature match the procurement application and usage registration item by itemCompare the ranking of the top few results before and after enabling usingReRank, check the completeness of the cited text after similarity filtering, and record the array length separately
Retrieval returned citationsThe number of citations and the tokens of the cited text are recorded separately, without equality comparisonCheck whether the cited text after limit trimming retains the key snippets of the original text, and record the number of candidates and the score

Three common errors

  • Symptom: The name and quantity of the same record do not match between the procurement application and the usage registration. Cause: The two materials were not checked against the same field definitions when entered, and retrieval candidates cannot automatically correct source discrepancies.
  • Symptom: The inventory ledger version and the procurement application version are not synchronized, and the dates do not match. Cause: The update cadence is not synchronized, and the old version remains in the candidate snippets; hits and associations must be checked first.
  • Symptom: The usage registration lacks a signature or receipt, and the handover form check finds missing items. Cause: The material archiving is incomplete, and the candidate snippets returned by retrieval cannot replace the signature and receipt themselves.

How to verify the setup

This feature acceptance: test once each with synonym-rewritten questions, compare candidates and scores, then toggle rerank on and off to inspect the ranking of the top few results.

Business test: check item by item against the handover form once, and see which record is missing.

  • In the knowledge base search test, switch searchMode and compare the number of candidates and scores under the semantic, full-text, and hybrid modes.
  • Compare the ranking of the top few results before and after enabling usingReRank, check the tokens of the cited text after similarity filtering and limit trimming, and record the array length separately.