Vector Models and Indexes for Research Project Task Agreement Signing: Maintain Document Versions and Changes

Research Agreements: review vector indexes for version tracking. Check source versions, input mappings and the values received by downstream nodes.

What this category's data looks like

This page addresses the scenario of signing research project task books. Materials include the project task book, the subject task breakdown table, and the budget book. The task book defines research objectives, assessment indicators, and start/end dates; the breakdown table splits subjects, undertaking units, and responsible persons; the budget book lists subject amounts, calculation bases, and annual allocations. When these three types of materials enter the knowledge base, they respectively serve as the master outline, division of labor, and funding basis. Fields involve project name, subject number, amount, date, and signature fields. Materials are updated with signing milestones, budget adjustments, and subject changes, and versions change frequently. The same record often has inconsistent names, quantities, and dates across different materials, so it is necessary to verify, against the version ledger and change records, that the current version is unique, old versions are traceable, and references are synchronized.

Configuration and implementation checks

This step is performed by the indexing service and vector database, outputting vectors and index status (whether recall is possible, whether data in the same group is merged). The "version ledger and change records" are then completed by subsequent AI nodes, code nodes, or manual work (retrieval and reference trimming only take supporting materials).

ObjectCheckVerification action
Project task bookCheck whether the project name, subject number, start/end dates, signature fields, and receipt are consistent with the breakdown table and budget bookOpen the knowledge base settings and check the index model; sample one data row to see whether the index field contains multiple vector identifiers, and confirm the status is indexed
Subject task breakdown tableCheck whether the subject name, undertaking unit, responsible person, quantity, and dates correspond to the task book and budget bookRun a retrieval once to see whether data in the same group is merged; check the data scope covered by the rebuild task record
Budget bookCheck whether the subject amounts, calculation bases, annual allocations, and signature fields are consistent with the task book and breakdown tableAfter changing one piece of content, re-index and check whether the version number and references are updated synchronously
Data row index fieldWhether multiple vector identifiers for one data row are all present, and whether data in the same group is merged by taking the highest scoreCheck the reference text budget against limit, and record the array length separately

Three common errors

  • Symptom: duplicate entries appear in retrieval results. Inspect the following: the name, quantity, and date of the same record do not match across two materials, and the vector identifiers are not merged into the same data group.
  • Symptom: recall quality drops after changing models. Inspect the following: the rebuild task record only covers the data scope for which indexing has been completed, and old-version references are not updated synchronously.
  • Symptom: a data row has content but cannot be retrieved. Inspect the following: the index field is missing vector identifiers, or signatures and receipts are incomplete, so the material does not enter the index.

How to verify the setup

This feature acceptance: sample one data row to inspect the index field, then run a retrieval once to see whether the same group is merged; check that the status is indexed, and inspect the coverage scope of the rebuild task record.

Business test: change one piece of content and see whether the version number and references are updated synchronously; the number of references is the array length, quoteMaxToken is the reference content budget, the two are not compared for equality, and the tokens of the reference text are checked separately.

  • In the knowledge base settings, check that the index model is consistent with the vector model actually used by the knowledge base.
  • Open a data row and check the index field to confirm that multiple vector identifiers are all present, and inspect the rebuild task record.