DetX

DetX is a modular software framework designed to support automated data extraction, transformation, and validation in information systems. It emphasizes determinism and reproducibility, offering a domain-specific language and a plugin architecture to build end-to-end data pipelines that can be versioned and audited.

The core comprises four layers: connectors, processors, orchestrator, and sink adapters. Connectors ingest from databases, files, APIs, and message queues; processors apply transformations, enrichments, and rules; the orchestrator manages tasks and dependency graphs; sink adapters write results to data warehouses or lakes and downstream systems. A declarative DSL lets users define extraction rules, transformation pipelines, and validation checks.

Key features include schema inference and evolution, data lineage and provenance, reproducible runs via configuration hash/versioning, and a test harness for unit and integration tests. It supports both batch and streaming processing, with pluggable backends for storage and compute.

Applications include data integration and quality assurance, compliance auditing, and supply chain transparency, where reproducibility and audit trails matter. The project is maintained by an open, community-driven effort with governance and semantic versioning.

Limitations include a learning curve for the DSL and potential overhead for very large pipelines, along with the need for disciplined environment governance to maintain reproducibility across stages.

See also ETL, data provenance, data lineage, reproducible research.