pyrasterix 0.0.1__tar.gz

pyrasterix-0.0.1/.github/workflows/ci.yml

@@ -0,0 +1,72 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+env:
+  CARGO_TERM_COLOR: always
+  RUSTFLAGS: -D warnings
+
+jobs:
+  verify:
+    name: just verify
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install Rust toolchain
+        uses: dtolnay/rust-toolchain@stable
+        with:
+          components: rustfmt, clippy
+
+      - name: Install just
+        uses: taiki-e/install-action@v2
+        with:
+          tool: just
+
+      - name: Install cargo-deny
+        uses: taiki-e/install-action@v2
+        with:
+          tool: cargo-deny
+
+      - name: Cache cargo
+        uses: Swatinem/rust-cache@v2
+
+      - name: just verify
+        run: just verify
+
+  extras:
+    name: docs + benches compile
+    runs-on: ubuntu-latest
+    needs: verify
+    steps:
+      - uses: actions/checkout@v4
+      - uses: dtolnay/rust-toolchain@stable
+      - uses: Swatinem/rust-cache@v2
+      - name: cargo doc
+        run: cargo doc --no-deps --all-features
+      - name: cargo bench --no-run
+        run: cargo bench --all-features --no-run
+
+  fuzz-smoke:
+    name: fuzz smoke (60s per target)
+    runs-on: ubuntu-latest
+    needs: verify
+    steps:
+      - uses: actions/checkout@v4
+      - uses: dtolnay/rust-toolchain@nightly
+      - uses: Swatinem/rust-cache@v2
+      - name: Install cargo-fuzz
+        run: cargo install cargo-fuzz --locked
+      - name: parse_object
+        working-directory: fuzz
+        run: cargo +nightly fuzz run parse_object -- -max_total_time=60
+      - name: parse_xref
+        working-directory: fuzz
+        run: cargo +nightly fuzz run parse_xref -- -max_total_time=60
+      - name: parse_document
+        working-directory: fuzz
+        run: cargo +nightly fuzz run parse_document -- -max_total_time=60

pyrasterix-0.0.1/.github/workflows/wheels.yml

@@ -0,0 +1,150 @@
+name: wheels
+
+# Builds Python wheels for the pyrasterix crate across Linux,
+# macOS, and Windows. Trigger options:
+#   - push a tag like `v1.0.0` to build a full release matrix
+#   - hit "Run workflow" in the Actions tab to build on demand
+#
+# Wheels are uploaded as workflow artifacts. Download from the run page,
+# then `pip install pyrasterix-*-<platform>.whl` on the target machine.
+# (Publishing to PyPI is intentionally not wired up; flip the commented
+# `release` job below on when you have a PYPI_API_TOKEN secret set.)
+
+on:
+  push:
+    tags: ['v*']
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+jobs:
+  linux:
+    name: linux ${{ matrix.target }}
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        target: [x86_64, aarch64]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.9'
+      - name: Build wheel (manylinux2014)
+        uses: PyO3/maturin-action@v1
+        with:
+          target: ${{ matrix.target }}
+          manylinux: '2014'
+          args: --release --out dist
+          working-directory: pyrasterix
+      - name: Smoke import + bench (native x86_64 only)
+        if: matrix.target == 'x86_64'
+        run: |
+          pip install --find-links pyrasterix/dist pyrasterix fpdf2
+          python -c "import pyrasterix; print(pyrasterix.__version__)"
+          python pyrasterix/scripts/make_bench_pdf.py /tmp/bench.pdf --pages 100
+          python pyrasterix/scripts/run_bench.py /tmp/bench.pdf
+      - uses: actions/upload-artifact@v4
+        with:
+          name: wheels-linux-${{ matrix.target }}
+          path: pyrasterix/dist/*.whl
+
+  macos:
+    name: macos ${{ matrix.target }}
+    runs-on: macos-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        target: [x86_64, aarch64]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.9'
+      - name: Build wheel
+        uses: PyO3/maturin-action@v1
+        with:
+          target: ${{ matrix.target }}
+          args: --release --out dist
+          working-directory: pyrasterix
+      - name: Smoke import + bench (arm64 only - macos-latest is Apple Silicon)
+        if: matrix.target == 'aarch64'
+        run: |
+          pip install --find-links pyrasterix/dist pyrasterix fpdf2
+          python -c "import pyrasterix; print(pyrasterix.__version__)"
+          python pyrasterix/scripts/make_bench_pdf.py /tmp/bench.pdf --pages 100
+          python pyrasterix/scripts/run_bench.py /tmp/bench.pdf
+      - uses: actions/upload-artifact@v4
+        with:
+          name: wheels-macos-${{ matrix.target }}
+          path: pyrasterix/dist/*.whl
+
+  windows:
+    name: windows x64
+    runs-on: windows-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.9'
+          architecture: x64
+      - name: Build wheel
+        uses: PyO3/maturin-action@v1
+        with:
+          target: x64
+          args: --release --out dist
+          working-directory: pyrasterix
+      - name: Smoke import + bench
+        run: |
+          pip install --find-links pyrasterix/dist pyrasterix fpdf2
+          python -c "import pyrasterix; print(pyrasterix.__version__)"
+          python pyrasterix/scripts/make_bench_pdf.py D:/bench.pdf --pages 100
+          python pyrasterix/scripts/run_bench.py D:/bench.pdf
+      - uses: actions/upload-artifact@v4
+        with:
+          name: wheels-windows-x64
+          path: pyrasterix/dist/*.whl
+
+  sdist:
+    name: source distribution
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Build sdist
+        uses: PyO3/maturin-action@v1
+        with:
+          command: sdist
+          args: --out dist
+          working-directory: pyrasterix
+      - uses: actions/upload-artifact@v4
+        with:
+          name: sdist
+          path: pyrasterix/dist/*.tar.gz
+
+  # Publishes all built wheels + sdist to PyPI on tag push. Uses PyPI
+  # trusted publishing (OIDC), no API tokens stored. You need to:
+  #
+  # 1. Create a Pending Publisher on https://pypi.org for the
+  #    `pyrasterix` project, pointing at this repo + workflow file
+  #    + the `pypi` environment name below.
+  # 2. Create a `pypi` GitHub Environment under Settings -> Environments.
+  # 3. Push a tag like `v0.0.1` and this job fires.
+  release:
+    name: publish to pypi
+    needs: [linux, macos, windows, sdist]
+    runs-on: ubuntu-latest
+    if: startsWith(github.ref, 'refs/tags/v')
+    environment:
+      name: pypi
+      url: https://pypi.org/p/pyrasterix
+    permissions:
+      id-token: write
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          path: dist
+          merge-multiple: true
+      - uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: dist

pyrasterix-0.0.1/.gitignore

@@ -0,0 +1,21 @@
+/target
+Cargo.lock.bak
+**/*.rs.bk
+
+# Fuzz crash dumps, coverage, and build artifacts are machine-specific.
+# Inside /fuzz/corpus/* only files named `*.pdf` / `*.obj` are committed
+# (curated seeds); libfuzzer's auto-generated inputs are not tracked.
+/fuzz/artifacts
+/fuzz/coverage
+/fuzz/target
+/fuzz/corpus/*/*
+!/fuzz/corpus/*/*.pdf
+!/fuzz/corpus/*/*.obj
+
+# Editor/OS junk
+.DS_Store
+.idea/
+.vscode/
+
+# Claude Code per-machine runtime state.
+/.claude

pyrasterix-0.0.1/CHANGELOG.md

@@ -0,0 +1,54 @@
+# Changelog
+
+## Unreleased
+
+Quality gate closed (mean corpus SSIM 0.944, target was 0.94).
+Perf at 1.84x a comparable Python library's MP pool cold, target was 2x. Honest writeup in
+`docs/BENCHMARKS.md`.
+
+Two bugs in this batch landed the quality gate. Both took five
+minutes to fix once the actual cause was traced:
+
+- Image XObjects with `/ColorSpace [/Separation /Black …]` were
+  being painted as raw luminance instead of as tint (so photos
+  came out inverted). Fixed in T-216. Worst-page SSIM went 0.68
+  to 0.95 on the Springer figure pages.
+- The page-tree walker only resolved `/Resources` when stored as
+  an indirect reference. Some PDFs (e.g. the No-Starch Rust book)
+  store `/Resources` inline as a Dict on every page, so no fonts
+  ever got registered and everything fell back to Helvetica.
+  Fixed in T-218. Rust book SSIM went 0.65 to 0.94 on the worst
+  pages.
+
+Perf work this round:
+
+- Page-level work stealing across PDFs (T-302b). Previously batch
+  ran rayon at PDF granularity so one worker would stall on a
+  700-page textbook while others sat idle. Flat page-level
+  `par_iter` over the merged page list pushed throughput from
+  270 to 303 p/s wall.
+- Separate PNG encode pool fed by a crossbeam channel (T-303).
+  No measurable win at 10 workers (encode is 3% of single-thread
+  time) but it's the SPEC §3.1 architectural shape.
+- Optional `mtpng` parallel PNG encoder behind the
+  `fast-encoders` feature flag (T-304). Slower at 10 workers due
+  to oversubscription. Off by default.
+
+Spec coverage this round: ICC profiles for image XObjects via
+qcms (T-406), CCITT Fax G3/G4 (T-404), tiling patterns (T-402),
+`/SMask` alpha (T-401 step 2), `/Indexed` color space (T-208).
+
+Tried and rejected: multi-process worker pool (T-301b). LPT-
+balanced 10 child processes gave 226 p/s vs rayon's 303 p/s.
+Static partitioning loses to dynamic work-stealing when per-page
+render cost varies.
+
+## 0.0.0
+
+The first 200-ish commits. PDF parser, page tree, content-stream
+interpreter, Standard 14 fonts via Liberation TTFs, DCT/Flate
+image decoding, JPEG output, Form XObject recursion, embedded
+TrueType / CFF / Type 1 via stet-fonts, ToUnicode CMaps for
+Identity-H.
+
+Full history in `docs/PROGRESS.md`.