Release v0.3.3

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog,
and this project adheres to Semantic Versioning.

[Unreleased]

Changed

• Planning: Phase 3 JVM/Kafka tracking consolidated into Planning/PHASE3_EPICS.md only (parity rows + P3-E1-S2d/S2e secret stories). Removed docs/java/PARITY_MATRIX.md and superseded Planning/ phase-plan markdown (PHASE1_*, Phase2_plan.md, PHASE2_EPICS.md, and related local notes) per single-tracker policy.
• README_CRATE.md: Phase 2 hero image uses raw.githubusercontent.com (same as PyPI) so crates.io and docs.rs always load the current PNG; documentation table adds PyPI, Python examples (HTML), and explicit GitHub Pages URLs for Rust rustdoc vs Python pdoc (root /rust_data_processing.html is not rustdoc; see below).
• GitHub Pages: Add docs/landing/rust_data_processing.html → site/rust_data_processing.html so https://…/rust_data_processing.html redirects to python/rust_data_processing.html (pdoc) with a visible link to rust/rust_data_processing/index.html (rustdoc). docs/DOCUMENTATION.md lists the canonical paths.

Added

• JVM Phase 3 bindings: bindings/jvm-sys (rdp_ffi_abi_version = 400, jvm_ffi feature, ffi_manifest.json); bindings/java/rust-data-processing-jvm (Maven, Gradle maven-publish, NativeAbiTest, FeatureSurfacePlaceholderTest, central-release profile); scripts/check_jvm_ffi_manifest.py, scripts/check_java_version_consistency.py; bindings/java/VERSION ↔ pom.xml ↔ gradle.properties.

• Docs: docs/adr/{004,005}-*.md, docs/java/{README,gradle,RELEASE,FFI_API_SLICE,ARROW_FFI_JVM,NATIVE_ARTIFACT_PACKAGING,SONATYPE_NAMESPACE_CHECKLIST,JIRA_INTEGRATION,JEXTRACT}.md, docs/java/MAVEN_CENTRAL_PUBLISHING.md; JVM parity rows live only in Planning/PHASE3_EPICS.md (standalone PARITY_MATRIX removed).

• CI: .github/workflows/jvm_bindings_ci.yml — Ubuntu / Windows / macOS × JDK 21, mvn verify, ./gradlew check, publishToMavenLocal, actions/cache for Maven/Gradle.

• Phase 3 planning (JVM + Kafka): Panama, Maven, Gradle, Rust↔JVM API parity, tri-surface Kafka (+ BYO connectors) in Planning/PHASE3_EPICS.md; docs/DOCUMENTATION.md, docs/adr/README.md, ADR docs/adr/003. Completed increment-1 spike crate spikes/jvm-panama-ffi (stdlib-only cdylib, rdp_ffi.h, Panama smoke). Root Cargo.toml excludes spikes/ from the crates.io tarball.

• Phase 2 (batch): ReduceOp::Median / Agg::Median; UTF-8 privacy TransformSteps (Utf8Truncate, Utf8Sha256Hex, Utf8RedactMiddle); validation Check::Utf8LenCharsBetween; modules export (JSONL + deterministic train/test indices), privacy (UTF-8 diff summaries), reports (byte-safe truncation); Arrow record_batches_to_dataset + LargeUtf8 support; Python bindings export_dataset_jsonl, privacy_summarize_utf8_changes_json, reports_truncate_utf8_bytes.

• Docs & examples: lake read ADR + user guide, SFT format guide, outreach shortlist, ML reduce gap audit, P2-E6 policy + P2-E7 deferral notes, vector export recipe, examples/{dbt,airflow,tabular_nn,llm_prep}/, notebooks/ index + starter notebooks.

0.3.3 - 2026-06-17

Changed

• (summarize this release)

0.3.2 - 2026-06-17

Changed

• (summarize this release)

0.3.1 - 2026-05-21

Changed

• (summarize this release)

0.3.0 - 2026-05-20

Changed

• (summarize this release)

0.2.2 - 2026-05-04

Changed

• (summarize this release)

0.2.1 - 2026-05-04

Fixed

• crates.io / docs.rs / PyPI README: Hero infographic now uses docs/images/phase-2-scope-overview.png and Phase 2 copy in README_CRATE.md and python-wrapper/README_PYPI.md (0.2.0 still showed the Phase 1 graphic on registries).

0.2.0 - 2026-05-04

Changed

• (summarize this release)

0.1.8 - 2026-04-14

Changed

• (summarize this release)

0.1.7 - 2026-04-14

Fixed

• docs.rs: Fix rustdoc intra-doc links that become hard errors under RUSTDOCFLAGS=-D warnings (broken Some(None)-style links, link to a private helper, redundant DataSet targets, ambiguous glob / reduce links). Republish so docs.rs can rebuild (0.1.6 docs build failed).
• docs.rs: add [package.metadata.docs.rs] with cargo-args = ["-j", "1"] so the documentation build is less likely to run out of memory while compiling Polars and the rest of the dependency graph (see Cargo.toml comments).
• PyPI / maturin: Remove duplicate docs/images/phase-1-scope-overview.png under python-wrapper/ and drop [tool.maturin] include for PNGs. The sdist merged the repo-root image (from cargo package) with the wrapper copy and failed with “was already added … can't add it from … python-wrapper/docs/images/”. The PyPI README now uses a raw.githubusercontent.com URL for that image; canonical file stays at docs/images/ in the repo.

Planned

• JVM / Java / Maven: first-class bindings (native library + Maven Central artifact, cross-platform CI) are Phase 3 scope, not Phase 2, so Phase 2 can focus on Rust + Python with smaller releases and faster cycles. Follow releases and notes here for when Java support lands.

0.1.6 - 2026-04-13

Changed

• (summarize this release)

0.1.5 - 2026-04-02

Changed

• (summarize this release)

0.1.4 - 2026-03-31

Changed

• (summarize this release)

0.1.3 - 2026-03-31

Changed

• (summarize this release)

0.1.2 - 2026-03-31

Fixed

• docs.rs / crates.io: The published crate now ships README_CRATE.md only (Rust-focused). The monorepo README.md is excluded from the .crate tarball so docs.rs no longer shows Python quick starts or mixed Python/Rust landing copy. PyPI continues to use python-wrapper/README_PYPI.md.

0.1.1 - 2026-03-30

Changed

• scripts/release_tag.ps1: optional -Comment, interactive release comment, prints last v* tag, fetches tags; clearer error text.
• CI: Documentation workflow uses astral-sh/setup-uv@v8.0.0 and drops redundant configure-pages for static Pages deploy.

0.1.0 - 2026-03-20

Added

• Initial crates.io release of rust-data-processing.
• Schema-first ingestion: CSV, JSON / NDJSON (nested dot paths), Parquet; Excel via excel feature.
• In-memory DataSet model (types) with Int64, Float64, Bool, Utf8, Null.
• processing: filter, map, reduce with ReduceOp (count, sum, min/max, mean, variance, std dev, sum-squares, L2 norm, count-distinct).
• processing::multi: feature_wise_mean_std, arg_max_row, arg_min_row, top_k_by_frequency.
• Polars-backed pipeline::DataFrame (lazy plan, collect to DataSet), group_by with Agg, joins, casts, filters.
• SQL over DataFrame (sql feature, default-on).
• execution engine: parallel filter/map, metrics, observers.
• profiling, validation, outliers, transform (TransformSpec), cdc boundary types.
• Optional db_connectorx for DB → Arrow → DataSet ingestion.