madmatcher-pro documentation

MadMatcher’s premium entity-matching stack: Sparkly (TF/IDF blocking), Delex (DAG multi-strategy blocking), MatchFlow (ML matching with active learning), and semantic (dense / embedding) blocking, with built-in crash recovery and a live progress dashboard.

If you are new here, start with a tutorial: each is a step-by-step walkthrough from loading data to saving blocking or matching output. Run the matching demo first to confirm your install, then read the tutorial to understand each step. The API reference documents every class and function you call, organized by area.

Tutorials

• Sparkly Tutorial
• Delex Tutorial
• MatchFlow Technical Guide
• Similarity Functions and Tokenizers in MatchFlow
• Semantic Blocking Tutorial

Guides

• Troubleshooting Guide
• Spark Tuning Guide

Demos

• Demos

API reference

• Session and progress tracking
  • MadMatcherSession
• Sparkly (lexical blocking)
  • Search
  • Index
  • Query spec
  • Index optimization and config generation (Sparkly Auto)
  • Table validation
• Delex (DAG multi-strategy blocking)
  • Running a blocking program
  • Blocking programs and rules
  • Predicates
  • Table validation
  • Advanced: extension base classes
• MatchFlow (ML matching + active learning)
  • Pipeline functions
  • Model wrappers
  • Labelers
  • Advanced: extension base classes
• Semantic (dense / embedding blocking)
  • Core pipeline
  • Embedding providers
  • Serializers
  • Advanced