parquet-adapter

[juliacher]

Summary

Adds the Parquet adapter implementation, including relational and document source support, schema discovery, partition handling, query pushdown optimizations, workflow integration, tests, and final project documentation.

Fixes: N/A

Changes

• Added Parquet source integration as a Polypheny plugin for relational and document models.
• Added schema discovery for single files, multi-file folders, dataset roots, and Hive-style partitioned layouts.
• Added flat and normalized relational schema modes for nested Parquet structures.
• Added adapter-level planning and execution paths for scans, filters, joins, and aggregates.
• Added metadata-backed statistics and file pruning based on partition values and Parquet footer statistics.
• Added workflow extract/load support for Parquet files.
• Cleaned and aligned project documentation with the current implementation.

---

Features

• Relational Parquet source adapter.
• Document Parquet source adapter.
• Multi-file Parquet table discovery.
• Hive-style partition column discovery.
• Flat nested scalar field exposure.
• Normalized nested table generation with synthetic structural columns.
• Projection and filter pushdown.
• Native Parquet predicate pushdown where supported.
• Adapter-level parent/child structural joins.
• Metadata and data aggregate pushdown.
• Metadata-backed table and column statistics.
• Parquet workflow import and export activities.

Bug Fixes

N/A

---

Tests

• Added unit coverage for schema discovery, partition handling, filter translation, file pruning, primitive reads, document conversion, aggregate execution, statistics, nested join planning, IO utilities, and workflow Parquet support
• Added integration-style coverage for importing and querying Parquet-backed relational sources.

---

ToDo