Overview
The core model behind Reconify runs.
Reconify has four core concepts:
- Sources describe input files and parser settings.
- Parsers map source-specific columns into normalized transactions.
- Pairs define which two sources to compare and which matching rules to apply.
- Outputs report matches, differences, duplicates, unmatched rows, and summary totals.
Data flow
- Reconify loads
reconify.yaml. - It validates config shape and source references.
- It resolves input files from explicit flags or source glob patterns.
- It parses rows into normalized transactions.
- It builds an index for the right source.
- It streams the left source and emits reconciliation events.
- It writes output as JSON, streaming JSON, NDJSON, CSV, or table.
Where to configure behavior
Parsing behavior belongs under sources.<name>.parser.
Matching behavior belongs under pairs.<name>.
Large-file memory behavior belongs under index.