pdfluent-extract 0.1.0__tar.gz

pdfluent_extract-0.1.0/.gitignore

@@ -0,0 +1,187 @@
+/target
+pdfluent-examples/dotnet/**/bin/
+pdfluent-examples/dotnet/**/obj/
+.env
+.env.*
+*.pem
+*.key
+*.p12
+*.pfx
+*.swp
+*.swo
+.DS_Store
+*.pdf
+!/tests/fixtures/*.pdf
+!/fixtures/formcalc/*.pdf
+!/fixtures/xfa-layout/*.pdf
+!/fixtures/xfa-data/*.pdf
+!/fixtures/scanned/*.pdf
+# Hostile PDFs: small, hand-built files that must not crash the renderer (#262).
+!/fixtures/pdf/*.pdf
+# The document every binding's structured-extraction suite reads (#573). It is
+# built and pinned by crates/pdf-engine/tests/the_api_over_the_block_model.rs;
+# the older files beside it are tracked from before `*.pdf` was ignored at all,
+# which is why only this one needs saying.
+!/fixtures/structured.pdf
+# Three text pages under a two-level outline, which the chunk suites of both
+# bindings read for per-page toc_items. Built by
+# scripts/make-outlined-fixture.py, byte-deterministic.
+!/fixtures/outlined.pdf
+# The four-page report with a running head and page numbers that the markdown
+# furniture suites read. Built and pinned by
+# crates/pdf-engine/tests/markdown_furniture_and_separators.rs.
+!/fixtures/running-head.pdf
+# One page with two hand-built image XObjects (7x5 RGB Flate, 4x3 Gray raw)
+# that every binding's known-images suite reads. Built by
+# fixtures/gen/known_images.py, byte-deterministic (#643).
+!/fixtures/known-images.pdf
+!/tests/corpus-mini/*.pdf
+!/crates/pdf-java/src/test/resources/*.pdf
+!/crates/visual-regression/fixtures/*.pdf
+# One DocLayNet train page, registered in corpus_herkomst.py: the operator
+# page of the list-splitting model (#599).
+!/crates/pdf-engine/tests/fixtures/doclaynet/*.pdf
+# PubTables-1M train pages, registered in corpus_herkomst.py (#574).
+!/crates/pdf-engine/tests/fixtures/pubtables/*.pdf
+metadata.sqlite
+/reports/
+/renders/
+/crates/xfa-wasm/pkg/
+/fuzz/artifacts/
+/fuzz/corpus/*/
+!/fuzz/corpus/*/seed_*
+node_modules/
+package-lock.json
+!/test-data/xfa-golden/*.pdf
+/benchmarks/weekly/benchmark-*.json
+/benchmarks/weekly/benchmark-*-report.md
+/benchmarks/weekly/regression-*.md
+/benchmarks/stress-fixtures/
+# Any __pycache__, at any depth. The narrower `scripts/__pycache__/` did not match
+# scripts/ci/__pycache__/, so bytecode from the CI helpers kept turning up staged.
+__pycache__/
+*.pyc
+.cache/
+out/
+
+# M8-CLEANUP-04 (#1380): VPS/local expensive validation reports
+validation-reports/
+
+# Build artifacts: Python bindings (maturin/PyO3 compiled extensions).
+# Machine-specific, must never be committed.
+crates/pdf-python/python/**/*.so
+crates/pdf-python/python/**/*.pyd
+crates/pdf-python/dist/
+# Project-local Python environment (`maturin develop` installs here).
+.venv/
+
+# Build artifacts: Node.js bindings (napi-rs compiled .node files).
+crates/pdf-node/index.*.node
+crates/pdf-node/artifacts/
+
+# Build artifacts: the @pdfluent/extract wrapper (tsc emit, never source).
+integrations/pdfluent-extract-js/dist/
+
+# Belt-and-suspenders for wasm-pack output within pkg/ (already covered by /crates/xfa-wasm/pkg/ above).
+crates/xfa-wasm/pkg/*.js
+crates/xfa-wasm/pkg/*.wasm
+crates/xfa-wasm/pkg/*.d.ts
+
+# .NET and Java build outputs (added wave1-release)
+bindings/dotnet/**/bin/
+bindings/dotnet/**/obj/
+bindings/java/target/
+crates/pdf-java/target/
+
+# .NET runtime-specific native bundles (regenerated at release time)
+bindings/dotnet/**/runtimes/
+
+# Java JNA native bundles (built by release pipeline, not committed)
+bindings/java/src/main/resources/native/
+
+# .NET NuGet packages
+*.nupkg
+
+# Doxygen generated output
+crates/pdf-capi/doxygen-out/
+
+# Compiled C example binaries (built by examples/Makefile)
+crates/pdf-capi/examples/license_activate
+
+# M9-A hygiene: generated publish dry-run logs embed machine-local /Users/ paths
+benchmarks/runs/**/dry_run_logs*/
+benchmarks/runs/**/dry_run_logs/
+**/audit_rust_post_commit.log
+
+# Corpus PDFs: large public-domain input files; not needed for CI builds.
+# Exception: small test fixtures referenced by non-ignored integration tests.
+# Note: only ignore the PDFs/JSONs, not the directory itself, so exceptions work.
+corpus/*.pdf
+!/corpus/PDFBOX-4322-3.pdf
+!/corpus/f1040.pdf
+!/corpus/sf15.pdf
+!/corpus/sf181.pdf
+!/corpus/fw7.pdf
+corpus/*.json
+# The gate corpus is a manifest, not a directory of PDFs: URLs pinned to an
+# upstream commit plus a SHA-256 each. Text, so it publishes safely (#215),
+# and it is the thing the pull-request gates are pinned to (#276).
+!/corpus/GATE_CORPUS_MANIFEST.json
+!/corpus/GATE_CORPUS_KNOWN_CRASHES.json
+# Where fetch_gate_corpus.py materialises that manifest. Never committed.
+/corpus/gate/
+# ...but the gate manifests are inputs, not outputs. They are tracked, and
+# `corpus/*.json` above would silently drop a re-added one -- the others
+# survive only because they predate that line. Declared, so the next person
+# who deletes and restores one does not lose it without being told.
+!/corpus/SSIM_GATE_MANIFEST.json
+
+benchmarks/runs/xfa_enterprise_plan/sprint2/heatmaps_demo/
+crates/pdf-capi/tests/test_text_blocks
+crates/pdf-capi/tests/test_structured_extraction
+crates/pdf-capi/tests/test_journal
+crates/pdf-capi/tests/test_form_write
+crates/pdf-capi/tests/test_plain_images
+benchmarks/runs/xfa_enterprise_plan/custom_defects/xfa_test_flatten/artifacts/
+/dist/
+
+# ── Fuzz build outputs ────────────────────────────────────────────────────────
+# The root /target rule only covers the workspace target. Every fuzz crate builds
+# its own target/ under its fuzz/ subdirectory and these are never source.
+/fuzz/target/
+**/fuzz/target/
+
+# ── WASM pkg-release directory ────────────────────────────────────────────────
+# /crates/xfa-wasm/pkg/ is already covered above; pkg-release/ is the optimised
+# release variant produced by wasm-pack --release and is equally reproducible.
+/crates/xfa-wasm/pkg-release/
+
+# ── Native binary artifacts (belt-and-suspenders) ─────────────────────────────
+# Specific directories are already covered above (bindings/**/runtimes/, etc.).
+# These global patterns guard against any new crate that adds a native output
+# path not yet covered by a directory rule.
+# Exceptions: none expected — no legitimate source file has these extensions.
+*.dylib
+*.dylib.dSYM/
+**/*.dSYM/
+*.dll
+*.so
+# *.rlib and *.rmeta live inside target/ (already ignored) but cover edge cases
+# where cargo or fuzz builds emit them outside the workspace target directory.
+*.rlib
+*.rmeta
+
+# ── pdfluent-examples example node_modules ───────────────────────────────────
+# The top-level node_modules/ rule covers workspace root, but examples that run
+# npm install inside their own subdirectory need this explicit pattern.
+pdfluent-examples/**/node_modules/
+crates/pdf-node/node_modules/
+
+# ── local measurement output ───────────────────────────────────────────────
+# Jobs write evidence under meas-out/<name>/ inside the workspace; the
+# launcher copies it out afterwards. Never committed, and invisible to the
+# clean-tree gates (a run that refused itself over its own output dir).
+# Measurement outputs: local proof logs, never tracked (#574).
+meas-out/
+# Measurement evidence, same rule: train Q carried meas/p4l-3f/ into the repo.
+meas/