goldenmatch dedupe FILE [...] | Deduplicate one or more CSV / Parquet / Excel files. |
goldenmatch match TARGET --against REF | Match a target against a reference file. |
goldenmatch sync --table TABLE --connection-string "$DB" | Incremental Postgres database sync. |
goldenmatch watch --table TABLE --config CONFIG | Live stream mode: poll a table and match new records continuously. |
goldenmatch serve FILE [...] | Start the REST API server (default 127.0.0.1:8080). |
goldenmatch mcp-serve FILE [...] | Start the MCP server (stdio or HTTP transport). |
goldenmatch interactive FILE [...] | Launch the interactive TUI. |
goldenmatch serve-ui [--project DIR] | Start the web UI workbench. |
goldenmatch setup | Interactive setup wizard (GPU, API keys, database). |
goldenmatch demo | Run the built-in demo with sample data. |
goldenmatch init | Interactive config wizard. |
goldenmatch profile FILE | Profile data quality. |
goldenmatch evaluate FILE --gt GT.csv | Evaluate matching accuracy against ground truth. |
goldenmatch incremental BASE --new NEW | Match new records against an existing base. |
goldenmatch analyze-blocking FILE | Analyze data and suggest blocking strategies. |
goldenmatch label FILE --config --gt | Interactively label pairs to build ground truth. |
goldenmatch review --config CONFIG | Guided review loop: walk borderline pairs (and the stale-correction queue) one at a time. Decisions feed Learning Memory. |
goldenmatch rollback RUN_ID | Undo a previous merge run. |
goldenmatch unmerge RECORD_ID --clusters FILE | Remove a record from its cluster and re-cluster the rest; writes the updated clusters CSV. Use --shatter to break the whole cluster apart. |
goldenmatch runs | List previous runs for rollback. |
goldenmatch memory stats|learn|export|import|show | Manage the Learning Memory store. |
goldenmatch config save|load|list|show | Manage config presets. |
goldenmatch import-splink SETTINGS.json -o CONFIG.yaml [--model-out MODEL.json] [--upgrade DATA] | Convert a Splink settings or trained-model JSON to a GoldenMatch config; --model-out persists trained m/u probabilities for reuse. --upgrade DATA.parquet additionally runs the data-aware upgrade pass (TF tables, measured distance thresholds, fan-out defenses: NE suggestion + cluster-guard tuning, threshold calibration) and writes the faithful baseline alongside as *.baseline.*. |
goldenmatch compare-clusters A B | Cluster comparison (unchanged / merged / partitioned / overlapping). |
goldenmatch sensitivity FILE | Parameter sweep over threshold, blocking, and matchkey. |
goldenmatch pprl link FILE_A FILE_B | Privacy-preserving record linkage. |