warp-ingest 1.0.0__tar.gz

warp_ingest-1.0.0/LICENSE.txt

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

warp_ingest-1.0.0/NOTICE.txt

@@ -0,0 +1,13 @@
+Copyright 2024 Ambika Sukla, Nlmatics Corp.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.

warp_ingest-1.0.0/PKG-INFO

@@ -0,0 +1,245 @@
+Metadata-Version: 2.4
+Name: warp-ingest
+Version: 1.0.0
+Summary: Pure-Python, RAG-friendly document parsers (PDF/HTML/text/XML/Markdown). Java- and Tika-free.
+Author: Ambika Sukla, Jonathan, John Scrudato
+Author-email: Ambika Sukla <ambika.sukla@nlmatics.com>, Jonathan <jonathan.hilgart@gmail.com>, John Scrudato <scrudato@umich.edu>
+License-Expression: Apache-2.0
+License-File: LICENSE.txt
+License-File: NOTICE.txt
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Legal Industry
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Requires-Dist: beautifulsoup4>=4.12.0,<5.0.0
+Requires-Dist: fastapi>=0.115,<1.0
+Requires-Dist: nltk>=3.9.1
+Requires-Dist: numpy>=1.24,<3
+Requires-Dist: pdfplumber>=0.11.4,<0.12.0
+Requires-Dist: pypdfium2>=5
+Requires-Dist: python-multipart>=0.0.9
+Requires-Dist: uvicorn>=0.30,<1.0
+Requires-Dist: onnxruntime>=1.17.0,<1.23 ; python_full_version == '3.10.*' and extra == 'ocr'
+Requires-Dist: onnxruntime>=1.27.0 ; python_full_version >= '3.11' and python_full_version < '3.15' and extra == 'ocr'
+Requires-Dist: opencv-python-headless>=4.13.0.92 ; extra == 'ocr'
+Requires-Dist: pyclipper>=1.4.0 ; extra == 'ocr'
+Requires-Dist: pyyaml>=6.0.3 ; extra == 'ocr'
+Requires-Dist: rapidocr-onnxruntime>=1.3.0,<2.0.0 ; extra == 'ocr'
+Requires-Dist: shapely>=2.1.2 ; extra == 'ocr'
+Requires-Dist: six>=1.17.0 ; extra == 'ocr'
+Requires-Dist: pymupdf4llm>=0.0.17 ; extra == 'tables'
+Requires-Dist: markdown2>=2.4.0 ; extra == 'tables'
+Requires-Python: >=3.10, <3.15
+Project-URL: Repository, https://github.com/Open-Source-Legal/Warp-Ingest
+Project-URL: Issues, https://github.com/Open-Source-Legal/Warp-Ingest/issues
+Provides-Extra: ocr
+Provides-Extra: tables
+Description-Content-Type: text/markdown
+
+<img width="620" height="423" alt="Warp-Ingest" align="center" src="https://github.com/user-attachments/assets/dffc8c8b-d7fc-4b64-8442-6ecce16a244f" />
+
+# Warp-Ingest
+
+**Warp-Ingest is a state-of-the-art, deterministic PDF parser.** It turns a PDF into
+layout-aware structure — accurate word- and block-level bounding boxes, structural
+labels (section header, paragraph, list item, table row), and the relationships
+between them (the section/sub-section/paragraph **parent ↔ child** hierarchy) — and
+renders it as blocks, JSON, HTML, or an [OpenContracts](https://github.com/JSv4/OpenContracts)
+structural export.
+
+It is **rule-based, not model-based**: structure comes from text coordinates,
+graphics, and font data — no GPU, no per-page rasterization, no training set. That
+makes it fast and predictable on long, text-layer documents (hundreds of pages), which
+is where vision parsers are slowest and least stable. See
+[Rule-based vs. model-based](#rule-based-vs-model-based) for the rationale.
+
+Warp-Ingest is a pure-Python rewrite of the nlmatics `nlm-ingestor` engine, with the
+Java/Apache-Tika and Tesseract dependencies removed.
+
+## How the pipeline works
+
+```
+PDF ──pdfplumber──► per-word boxes + fonts ──► Tika-format XHTML ──► visual_ingestor ──► blocks / JSON / HTML / OpenContracts
+       (scanned pages ──rapidocr──► OCR words ──► same XHTML ──┘)
+```
+
+- **`pdfplumber`** (MIT) extracts each word's real bounding box and font data in
+  absolute, top-left-origin PDF points.
+- A scanned or sparse page is detected automatically and routed to the optional
+  **`rapidocr-onnxruntime`** OCR backend (Apache-2.0, no Tesseract binary, no GPU),
+  which emits the **same** word-box format — so a scanned page and a born-digital page
+  flow through the identical layout engine.
+- The front-end emits an intermediate **Tika-format XHTML** (one `<p>` per visual line,
+  carrying per-word positions and fonts). `visual_ingestor` — the ~6,000-line rule
+  engine — groups those lines into typed blocks, detects tables from rule-line
+  graphics, strips repeating headers/footers, removes watermarks, and fixes reading
+  order.
+
+The word-level boxes use the same technique as OpenContracts (see
+[docs/bbox_architecture.md](docs/bbox_architecture.md)).
+
+## What the parser produces
+
+1. Sections and sub-sections with their nesting levels.
+2. Paragraphs (lines joined into coherent blocks).
+3. The parent ↔ child links between sections and paragraphs.
+4. Tables, with the section each table sits in.
+5. Lists and nested lists.
+6. Content joined across page breaks.
+7. Removal of repeating headers and footers.
+8. Watermark removal.
+9. OCR with bounding boxes for scanned pages.
+10. An [OpenContracts structural export](docs/opencontracts_export_format.md): PAWLS
+    word tokens, one structural annotation per block, and the `parent_id` heading
+    hierarchy as explicit relationships.
+
+## Benchmarks
+
+Warp-Ingest is scored on the official
+[LlamaIndex ParseBench](https://github.com/run-llama/ParseBench) — **2,078
+human-verified pages** of real enterprise documents — by running it *through the
+official framework* with deterministic, rule-based metrics (no LLM-as-a-judge). Among
+the **8 deterministic, local, no-API parsers**, Warp-Ingest is a **top-2 result
+(2nd overall)**, it is the **only** local parser that carries real **visual
+grounding**, and it **leads on Charts**. Full numbers, methodology, and the
+reproduction commands are in [benchmarks/parsebench/RESULTS.md](benchmarks/parsebench/RESULTS.md).
+
+## Installation
+
+```bash
+# install the project and dev tools with uv
+uv sync --group dev
+
+# include the optional OCR backend for scanned PDFs
+uv sync --group dev --extra ocr
+
+# one-time NLTK data download
+uv run python -m nltk.downloader punkt punkt_tab stopwords
+```
+
+## Running the service
+
+```bash
+python -m warp_ingest.ingestion_daemon      # or: ./run.sh   (FastAPI/uvicorn, port 5001)
+```
+
+The launcher budgets concurrency automatically from the CPUs actually available
+(CPU affinity **and** the container cgroup quota, so a docker `--cpus` / K8s
+`limits.cpu` deployment uses exactly its slice): one uvicorn worker per
+effective CPU, with the front-end page-striping pool and the OCR session
+threads sized so the layers never oversubscribe the box. Every knob can be
+overridden via environment variables:
+
+| Env var | Default | Effect |
+|---|---|---|
+| `WARP_API_KEY` | `abc123` | API key required by `/api/parse` (send as `X-API-Key` or `Authorization: Bearer`) |
+| `WARP_WEB_WORKERS` (or `WEB_CONCURRENCY`) | effective CPUs | uvicorn worker processes |
+| `WARP_FE_WORKERS` | `max(1, min(8, cpus // workers))` | per-worker front-end page-striping pool (≤1 = serial) |
+| `WARP_OCR_THREADS` | `max(1, cpus // (workers × fe_workers))` | onnxruntime intra-op threads per OCR session |
+| `WARP_WORKER_PARSE_SLOTS` | `1` | concurrent parses allowed inside one worker |
+| `WARP_HOST` / `WARP_PORT` (or `PORT`) | `0.0.0.0` / `5001` | bind address |
+
+`POST /api/parse` with a `file` form field and the API key; the response body is
+the parse result (`{"page_dim": ..., "num_pages": ..., "result": ...}`) and
+errors are standard `{"detail": ...}` bodies. Interactive OpenAPI docs at `/docs`.
+
+```
+curl -H "X-API-Key: abc123" -F file=@document.pdf \
+  "http://localhost:5001/api/parse?render_format=all"
+```
+
+Query parameters (booleans accept `true`/`false`, `1`/`0`, `yes`/`no`, `on`/`off`):
+
+| Param | Values | Effect |
+|---|---|---|
+| `render_format` | `all` \| `json` \| `html` \| `opencontracts` | output rendering |
+| `apply_ocr` | bool | force OCR on every page (scanned pages are OCR'd automatically regardless) |
+| `disable_ocr` | bool | keep every page on its embedded text layer — no OCR for this request |
+| `semantic_units` | bool | append the additive Semantic-Unit clause layer (`render_format=opencontracts`) |
+
+`GET /` and `GET /healthz` are unauthenticated health endpoints (`/healthz`
+reports the resolved concurrency settings and OCR availability). Warp-Ingest
+parses **PDF only**; a non-PDF upload returns HTTP 415.
+
+### Docker
+
+```bash
+docker build -t warp-ingest .
+docker run -p 5010:5001 -e WARP_API_KEY=change-me warp-ingest
+```
+
+Every GitHub release publishes a versioned multi-arch image to GHCR
+(`ghcr.io/open-source-legal/warp-ingest:X.Y.Z` / `:X.Y` / `:X` / `:latest`,
+cosign-signed) via `.github/workflows/docker-publish.yml`.
+
+### Library use
+
+```python
+from warp_ingest.ingestor import pdf_ingestor
+
+export = pdf_ingestor.parse_to_opencontracts("document.pdf")   # OpenContracts export
+ingestor = pdf_ingestor.PDFIngestor("document.pdf", {"render_format": "all"})
+blocks = ingestor.blocks
+```
+
+The notebook
+[pdf_visual_ingestor_step_by_step.ipynb](notebooks/pdf_visual_ingestor_step_by_step.ipynb)
+walks the whole pipeline on a sample PDF.
+
+## Testing
+
+No Java or Tika is needed.
+
+```bash
+make test                  # full pytest suite (unit + fixture parsing, incl. OCR)
+uv run pytest tests/       # same, directly
+```
+
+Beyond unit tests, the suite includes cross-engine regression against the original
+Java/Tika engine (S-1), an OpenContracts-export regression, a Docling layout oracle,
+and vision-adjudicated structural-correctness suites — all floored against committed
+baselines so engine changes can only improve, never silently regress. See
+[CLAUDE.md](CLAUDE.md) for the map of which suite guards what.
+
+## Rule-based vs. model-based
+
+Over four years the nlmatics team evaluated many options, including a YOLO-based vision
+parser, and settled on the rule-based approach for these reasons:
+
+1. **Speed.** It is ~100× faster than a vision parser, which must rasterize every page
+   (even text-layer ones). A vision parser is the better tool for scanned PDFs with no
+   text layer or small form-like documents; for large text-layer PDFs spanning hundreds
+   of pages, a rule-based parser is far more practical.
+2. **No special hardware.** It runs on CPU.
+3. **Fixable.** Vision-parser errors are fixed either by adding training examples (which
+   can degrade previously-correct behavior) or by layering on rules anyway — at which
+   point you are writing rules again.
+
+## Credits
+
+The PDF parser
+[visual_ingestor](warp_ingest/ingestor/visual_ingestor/visual_ingestor.py) and its
+indent parsers were
+written by Ambika Sukla, with contributions from Reshav Abraham, Tom Liu (the original
+Indent Parser), and Kiran Panicker (parsing speed, table-parsing, indent-parsing, and
+reordering accuracy). The core
+[line_parser](warp_ingest/ingestor/line_parser.py) was written by Ambika Sukla.
+
+Thanks to the `pdfplumber` / `pdfminer.six`, `pypdfium2`, and `RapidOCR` open-source
+communities, and to the Apache PDFBox and Tika developers whose XHTML format the engine
+is built around.
+
+## History
+
+Earlier versions depended on an nlmatics-modified Apache Tika (`nlm-tika`) on the JVM
+plus Tesseract for OCR. That is gone — the pure-Python front-end
+([`pdf_plumber_parser.py`](warp_ingest/file_parser/pdf_plumber_parser.py) and
+[`ocr_parser.py`](warp_ingest/file_parser/ocr_parser.py)) reproduces the same
+intermediate XHTML contract, so the layout engine did not have to change.

warp_ingest-1.0.0/README.md

@@ -0,0 +1,200 @@
+<img width="620" height="423" alt="Warp-Ingest" align="center" src="https://github.com/user-attachments/assets/dffc8c8b-d7fc-4b64-8442-6ecce16a244f" />
+
+# Warp-Ingest
+
+**Warp-Ingest is a state-of-the-art, deterministic PDF parser.** It turns a PDF into
+layout-aware structure — accurate word- and block-level bounding boxes, structural
+labels (section header, paragraph, list item, table row), and the relationships
+between them (the section/sub-section/paragraph **parent ↔ child** hierarchy) — and
+renders it as blocks, JSON, HTML, or an [OpenContracts](https://github.com/JSv4/OpenContracts)
+structural export.
+
+It is **rule-based, not model-based**: structure comes from text coordinates,
+graphics, and font data — no GPU, no per-page rasterization, no training set. That
+makes it fast and predictable on long, text-layer documents (hundreds of pages), which
+is where vision parsers are slowest and least stable. See
+[Rule-based vs. model-based](#rule-based-vs-model-based) for the rationale.
+
+Warp-Ingest is a pure-Python rewrite of the nlmatics `nlm-ingestor` engine, with the
+Java/Apache-Tika and Tesseract dependencies removed.
+
+## How the pipeline works
+
+```
+PDF ──pdfplumber──► per-word boxes + fonts ──► Tika-format XHTML ──► visual_ingestor ──► blocks / JSON / HTML / OpenContracts
+       (scanned pages ──rapidocr──► OCR words ──► same XHTML ──┘)
+```
+
+- **`pdfplumber`** (MIT) extracts each word's real bounding box and font data in
+  absolute, top-left-origin PDF points.
+- A scanned or sparse page is detected automatically and routed to the optional
+  **`rapidocr-onnxruntime`** OCR backend (Apache-2.0, no Tesseract binary, no GPU),
+  which emits the **same** word-box format — so a scanned page and a born-digital page
+  flow through the identical layout engine.
+- The front-end emits an intermediate **Tika-format XHTML** (one `<p>` per visual line,
+  carrying per-word positions and fonts). `visual_ingestor` — the ~6,000-line rule
+  engine — groups those lines into typed blocks, detects tables from rule-line
+  graphics, strips repeating headers/footers, removes watermarks, and fixes reading
+  order.
+
+The word-level boxes use the same technique as OpenContracts (see
+[docs/bbox_architecture.md](docs/bbox_architecture.md)).
+
+## What the parser produces
+
+1. Sections and sub-sections with their nesting levels.
+2. Paragraphs (lines joined into coherent blocks).
+3. The parent ↔ child links between sections and paragraphs.
+4. Tables, with the section each table sits in.
+5. Lists and nested lists.
+6. Content joined across page breaks.
+7. Removal of repeating headers and footers.
+8. Watermark removal.
+9. OCR with bounding boxes for scanned pages.
+10. An [OpenContracts structural export](docs/opencontracts_export_format.md): PAWLS
+    word tokens, one structural annotation per block, and the `parent_id` heading
+    hierarchy as explicit relationships.
+
+## Benchmarks
+
+Warp-Ingest is scored on the official
+[LlamaIndex ParseBench](https://github.com/run-llama/ParseBench) — **2,078
+human-verified pages** of real enterprise documents — by running it *through the
+official framework* with deterministic, rule-based metrics (no LLM-as-a-judge). Among
+the **8 deterministic, local, no-API parsers**, Warp-Ingest is a **top-2 result
+(2nd overall)**, it is the **only** local parser that carries real **visual
+grounding**, and it **leads on Charts**. Full numbers, methodology, and the
+reproduction commands are in [benchmarks/parsebench/RESULTS.md](benchmarks/parsebench/RESULTS.md).
+
+## Installation
+
+```bash
+# install the project and dev tools with uv
+uv sync --group dev
+
+# include the optional OCR backend for scanned PDFs
+uv sync --group dev --extra ocr
+
+# one-time NLTK data download
+uv run python -m nltk.downloader punkt punkt_tab stopwords
+```
+
+## Running the service
+
+```bash
+python -m warp_ingest.ingestion_daemon      # or: ./run.sh   (FastAPI/uvicorn, port 5001)
+```
+
+The launcher budgets concurrency automatically from the CPUs actually available
+(CPU affinity **and** the container cgroup quota, so a docker `--cpus` / K8s
+`limits.cpu` deployment uses exactly its slice): one uvicorn worker per
+effective CPU, with the front-end page-striping pool and the OCR session
+threads sized so the layers never oversubscribe the box. Every knob can be
+overridden via environment variables:
+
+| Env var | Default | Effect |
+|---|---|---|
+| `WARP_API_KEY` | `abc123` | API key required by `/api/parse` (send as `X-API-Key` or `Authorization: Bearer`) |
+| `WARP_WEB_WORKERS` (or `WEB_CONCURRENCY`) | effective CPUs | uvicorn worker processes |
+| `WARP_FE_WORKERS` | `max(1, min(8, cpus // workers))` | per-worker front-end page-striping pool (≤1 = serial) |
+| `WARP_OCR_THREADS` | `max(1, cpus // (workers × fe_workers))` | onnxruntime intra-op threads per OCR session |
+| `WARP_WORKER_PARSE_SLOTS` | `1` | concurrent parses allowed inside one worker |
+| `WARP_HOST` / `WARP_PORT` (or `PORT`) | `0.0.0.0` / `5001` | bind address |
+
+`POST /api/parse` with a `file` form field and the API key; the response body is
+the parse result (`{"page_dim": ..., "num_pages": ..., "result": ...}`) and
+errors are standard `{"detail": ...}` bodies. Interactive OpenAPI docs at `/docs`.
+
+```
+curl -H "X-API-Key: abc123" -F file=@document.pdf \
+  "http://localhost:5001/api/parse?render_format=all"
+```
+
+Query parameters (booleans accept `true`/`false`, `1`/`0`, `yes`/`no`, `on`/`off`):
+
+| Param | Values | Effect |
+|---|---|---|
+| `render_format` | `all` \| `json` \| `html` \| `opencontracts` | output rendering |
+| `apply_ocr` | bool | force OCR on every page (scanned pages are OCR'd automatically regardless) |
+| `disable_ocr` | bool | keep every page on its embedded text layer — no OCR for this request |
+| `semantic_units` | bool | append the additive Semantic-Unit clause layer (`render_format=opencontracts`) |
+
+`GET /` and `GET /healthz` are unauthenticated health endpoints (`/healthz`
+reports the resolved concurrency settings and OCR availability). Warp-Ingest
+parses **PDF only**; a non-PDF upload returns HTTP 415.
+
+### Docker
+
+```bash
+docker build -t warp-ingest .
+docker run -p 5010:5001 -e WARP_API_KEY=change-me warp-ingest
+```
+
+Every GitHub release publishes a versioned multi-arch image to GHCR
+(`ghcr.io/open-source-legal/warp-ingest:X.Y.Z` / `:X.Y` / `:X` / `:latest`,
+cosign-signed) via `.github/workflows/docker-publish.yml`.
+
+### Library use
+
+```python
+from warp_ingest.ingestor import pdf_ingestor
+
+export = pdf_ingestor.parse_to_opencontracts("document.pdf")   # OpenContracts export
+ingestor = pdf_ingestor.PDFIngestor("document.pdf", {"render_format": "all"})
+blocks = ingestor.blocks
+```
+
+The notebook
+[pdf_visual_ingestor_step_by_step.ipynb](notebooks/pdf_visual_ingestor_step_by_step.ipynb)
+walks the whole pipeline on a sample PDF.
+
+## Testing
+
+No Java or Tika is needed.
+
+```bash
+make test                  # full pytest suite (unit + fixture parsing, incl. OCR)
+uv run pytest tests/       # same, directly
+```
+
+Beyond unit tests, the suite includes cross-engine regression against the original
+Java/Tika engine (S-1), an OpenContracts-export regression, a Docling layout oracle,
+and vision-adjudicated structural-correctness suites — all floored against committed
+baselines so engine changes can only improve, never silently regress. See
+[CLAUDE.md](CLAUDE.md) for the map of which suite guards what.
+
+## Rule-based vs. model-based
+
+Over four years the nlmatics team evaluated many options, including a YOLO-based vision
+parser, and settled on the rule-based approach for these reasons:
+
+1. **Speed.** It is ~100× faster than a vision parser, which must rasterize every page
+   (even text-layer ones). A vision parser is the better tool for scanned PDFs with no
+   text layer or small form-like documents; for large text-layer PDFs spanning hundreds
+   of pages, a rule-based parser is far more practical.
+2. **No special hardware.** It runs on CPU.
+3. **Fixable.** Vision-parser errors are fixed either by adding training examples (which
+   can degrade previously-correct behavior) or by layering on rules anyway — at which
+   point you are writing rules again.
+
+## Credits
+
+The PDF parser
+[visual_ingestor](warp_ingest/ingestor/visual_ingestor/visual_ingestor.py) and its
+indent parsers were
+written by Ambika Sukla, with contributions from Reshav Abraham, Tom Liu (the original
+Indent Parser), and Kiran Panicker (parsing speed, table-parsing, indent-parsing, and
+reordering accuracy). The core
+[line_parser](warp_ingest/ingestor/line_parser.py) was written by Ambika Sukla.
+
+Thanks to the `pdfplumber` / `pdfminer.six`, `pypdfium2`, and `RapidOCR` open-source
+communities, and to the Apache PDFBox and Tika developers whose XHTML format the engine
+is built around.
+
+## History
+
+Earlier versions depended on an nlmatics-modified Apache Tika (`nlm-tika`) on the JVM
+plus Tesseract for OCR. That is gone — the pure-Python front-end
+([`pdf_plumber_parser.py`](warp_ingest/file_parser/pdf_plumber_parser.py) and
+[`ocr_parser.py`](warp_ingest/file_parser/ocr_parser.py)) reproduces the same
+intermediate XHTML contract, so the layout engine did not have to change.