Match vendor SKUs to inventory items using PostgreSQL trigram indexes.
npx skills add https://github.com/dadbodgeoff/drift --skill fuzzy-matching
Please help me install this Agent Skill. Skill: fuzzy-matching Source: https://github.com/dadbodgeoff/drift/tree/main/drift%20v1%20depreciated/skills/fuzzy-matching Command: npx skills add https://github.com/dadbodgeoff/drift --skill fuzzy-matching
This Skill solves the challenge of accurately matching and reconciling entities (like inventory items or products) that have variations in their names, formats, or details, preventing duplicates and ensuring data consistency.
Use the fuzzy-matching skill to find similar items for 'BNLS CHKN BRST 10LB' in the 'poultry' category for user 'user_123'.
High-intent search queries and answers about installing and using this skill.
To match vendor SKUs to inventory items with inconsistent names, use multi-stage fuzzy matching. This process applies PostgreSQL trigram pre-filtering, salient overlap checks, and weighted similarity scoring for accurate entity reconciliation.
Deduplicating products across sources requires multi-stage fuzzy matching in PostgreSQL. It uses trigram indexes for efficient pre-filtering, then applies multi-factor similarity scoring with configurable thresholds to identify and reconcile duplicate entities.
Trigram indexing accelerates fuzzy matching by breaking strings into character sequences for rapid PostgreSQL pre-filtering. This narrows down potential matches before executing computationally intensive multi-factor similarity scoring on the full dataset.
Yes, entity reconciliation supports configurable thresholds for similarity scoring. You can adjust these parameters to control match strictness, balancing precision and recall when reconciling entities with varying name formats or details.
Yes, multi-stage fuzzy matching requires Python alongside PostgreSQL. The pipeline combines Python's processing capabilities with PostgreSQL trigram indexes to execute pre-filtering, overlap checks, and weighted similarity scoring efficiently.
Exact matching fails for entity reconciliation because it cannot handle abbreviations or formatting differences. Fuzzy matching solves this by using trigram overlap and weighted similarity scoring to match strings like 'BNLS CHKN BRST' to 'Boneless Chicken Breast'.
cwenselSearch indexed collections with semantic and full-text queries.
GeneralReasoningMatch similar strings across datasets using Levenshtein-based similarity algorithms.
princegarg001Reconcile inventory items using PostgreSQL trigram pre-filtering and multi-factor similarity scoring.
SufficientDaikonDetect and deduplicate text content using normalization, SHA-256 fingerprinting, and Jaccard similarity.
