htmlsift 0.1.0__tar.gz

htmlsift-0.1.0/.gitattributes

@@ -0,0 +1,5 @@
+# Normalize line endings: LF in the repo, native in the working tree.
+* text=auto eol=lf
+
+# Treat rendered assets and goldens as text but never munge them.
+*.svg text eol=lf

htmlsift-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,24 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.10", "3.13"]   # floor + latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+      - name: Install (base deps only, no torch)
+        run: pip install -e . ruff pytest
+      - name: Lint
+        run: ruff check src tests
+      - name: Test (model goldens skip without HTMLSIFT_BUNDLE)
+        run: pytest -q

htmlsift-0.1.0/.gitignore

@@ -0,0 +1,48 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+.eggs/
+
+# Build artifacts
+build/
+dist/
+
+# Virtual environments
+.venv*/
+venv/
+
+# Secrets / local config
+.env
+
+# Tooling caches
+.pytest_cache/
+.ruff_cache/
+.mypy_cache/
+
+# Local model artifacts (weights are never committed; the package fetches them from HF)
+*.pt
+*.onnx
+
+# Local dev bundles (e.g. localbundle/mini with graphs + tokenizer.json + manifest.json)
+localbundle/
+
+# Local agent instructions (kept local, not shipped; AGENTS.md is the committed one)
+CLAUDE.md
+
+# Editors
+.vscode/
+.idea/
+
+# OS junk
+.DS_Store
+Thumbs.db
+desktop.ini
+
+# Coverage / other test tooling
+.coverage
+htmlcov/
+.tox/
+.nox/
+
+.env

htmlsift-0.1.0/AGENTS.md

@@ -0,0 +1,45 @@
+# AGENTS.md — htmlsift
+
+## What this is
+
+A pip-installable web main-content extractor. HTML string in, clean
+text / HTML / markdown out, using the trained htmlsift models. Apache-2.0.
+This is the production inference package; model training and reproduction
+live in the separate research repo (see Resources).
+
+Two modes:
+- `mini` (default) — int8 ONNX, CPU, torch-free. `pip install htmlsift`
+- `base` — 311M torch encoder, GPU-capable. `pip install htmlsift[full]`
+
+Model weights are not in the repo. They are fetched from the Hugging Face
+Hub on first use, pinned by commit SHA per release (see
+`src/htmlsift/_artifacts.py`).
+
+## Licence and dependencies
+
+Apache-2.0. Rendering and serialization are lxml-only, and the package
+keeps its dependency tree free of GPL-licensed code — that is what lets it
+stay Apache-2.0. Changes that pull in a GPL dependency (for example
+`html2text`) won't be accepted into the project.
+
+## Layout
+
+- `src/htmlsift/` — the package. Public API in `extractor.py`
+  (`Extractor`, `extract`); model assembly in `loader.py`; bundle fetch
+  in `_artifacts.py`; CLI in `cli.py`.
+- `src/htmlsift/core/` — the pipeline: parse/render, prep, features,
+  inference (`onnx_infer.py` for mini, `torch_infer.py` for base).
+- `tests/` — golden and unit tests.
+
+## Working with the code
+
+- Backends are imported lazily, so a default install never loads torch;
+  the torch/transformers imports live in `core/torch_infer.py`.
+- Lint: `ruff check src tests`
+- Test: `pytest -q`. Model golden tests skip unless `HTMLSIFT_BUNDLE`
+  points at a local bundle (e.g. `localbundle/mini`).
+
+## Resources
+
+- `htmlsift-research` — training, evaluation, and reproduction:
+  https://github.com/koivualeksi/htmlsift-research

htmlsift-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 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.

htmlsift-0.1.0/PKG-INFO

@@ -0,0 +1,89 @@
+Metadata-Version: 2.5
+Name: htmlsift
+Version: 0.1.0
+Summary: Main-content extraction for web pages
+Project-URL: Repository, https://github.com/koivualeksi/htmlsift
+Author-email: Aleksi Koivu <koivu.aleksi@gmail.com>
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: boilerplate-removal,content-extraction,html,main-content,web-scraping
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+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: Topic :: Text Processing :: Markup :: HTML
+Requires-Python: >=3.10
+Requires-Dist: huggingface-hub
+Requires-Dist: lxml
+Requires-Dist: numpy
+Requires-Dist: onnxruntime
+Requires-Dist: tokenizers
+Provides-Extra: full
+Requires-Dist: torch; extra == 'full'
+Requires-Dist: transformers; extra == 'full'
+Description-Content-Type: text/markdown
+
+# htmlsift
+
+Web main-content extraction: HTML in, clean text / HTML / markdown out. A small model,
+trained on [WebMainBench](https://github.com/opendatalab/WebMainBench), scores each rendered
+line of a page as main content or boilerplate (nav, ads, cookie banners, and footers) and
+returns just the content.
+
+Apache-2.0. Full documentation, diagrams, and benchmarks:
+https://github.com/koivualeksi/htmlsift
+
+## Install
+
+```bash
+pip install htmlsift          # default: the mini model (CPU, ONNX)
+pip install htmlsift[full]    # adds the base 311M model (torch; GPU-capable)
+```
+
+- `htmlsift` — the `mini` model only; a small install (onnxruntime + tokenizers + lxml +
+  numpy + huggingface-hub). First use downloads ~230 MB of model files.
+- `htmlsift[full]` — adds `torch` + `transformers` to run `base`. First use of `base`
+  downloads ~1 GB.
+
+Requires Python 3.10+. Model files download from the Hugging Face Hub on first use and are
+cached; run `htmlsift download <mode>` to prefetch, or set `HF_HUB_OFFLINE=1` to run from
+cache only.
+
+## Usage
+
+```python
+from htmlsift import Extractor
+
+ex = Extractor()                     # defaults to "mini"
+ex.extract(html)                     # -> str (plain text)
+ex.extract(html, output="html")      # structure, links, images kept
+ex.extract(html, output="markdown")  # clean markdown: [text](url), ![alt](src)
+ex.extract(html, with_blocks=True)   # -> (blocks, probs)
+
+import htmlsift
+htmlsift.extract(html)               # module-level convenience
+```
+
+The larger model, from `htmlsift[full]`:
+
+```python
+ex = Extractor("base")               # 311M encoder; GPU when present, else CPU (with a warning)
+ex = Extractor("base", device="cuda")  # require the GPU
+```
+
+`mini` is CPU-only; only `base` runs on the GPU.
+
+## Output modes
+
+| `output` | what you get |
+|---|---|
+| `text` (default) | selected content as plain text |
+| `html` | selected DOM, with structure / links / images kept |
+| `markdown` | clean markdown, links and images preserved |
+
+## License
+
+Apache-2.0. Model weights are distributed separately, from the Hugging Face Hub.

htmlsift-0.1.0/README.md

@@ -0,0 +1,207 @@
+# htmlsift
+
+Main-content extraction for web pages: hand it a raw HTML string and get back the main content
+ - nav, ads, cookie banners, and footers stripped
+ - outputs raw text, **HTML**, or **markdown**.
+
+Two models shipped, both trained on [WebMainBench][wmb] (WMB) — 7,825 labelled pages, custom split
+6,548 train / 732 val / 545 test:
+
+- **`mini`** (default) - a small int8-ONNX model that runs on CPU. `pip install htmlsift`
+- **`base`** - a 311M encoder, most accurate and GPU-capable. `pip install htmlsift[full]`
+
+Each model is fully re-trainable with your own annotated data and the whole research suite is available at [htmlsift-research](https://github.com/koivualeksi/htmlsift-research)
+
+[wmb]: https://github.com/opendatalab/WebMainBench
+
+## Results
+
+| model | desc | WMB test F1 | seeds | CPU (pages/s) | GPU (pages/s) | install |
+|---|---|---|---|---|---|---|
+| `base` | 311m-10 | **0.9311** | 3 | 0.4 | 24.6 | `htmlsift[full]` |
+| unpublished | 97m-6, int8 QAT | 0.9256 | 3 | 0.9 | 30.2 | — |
+| `mini` | int8 table | 0.9010 | 3 | **27.7** | — | `htmlsift` |
+
+* GPU = RTX 4090
+* CPU = Ryzen 5 3600 single thread
+
+### On CPU
+
+![htmlsift mini vs trafilatura, readability, resiliparse: F1 and ms/page on one CPU core](docs/assets/cpu_speed_accuracy.svg)
+
+One CPU core, WMB test (544 pages), extraction only. `mini` gets close to trafilatura's
+per-page time (36 vs 31 ms median) and scores ~15 F1 points higher. readability and resiliparse are faster but markedly less accurate. 
+
+### On GPU
+
+![htmlsift base sustains 21x the throughput of MinerU-HTML at matched F1 on an RTX 4090](docs/assets/gpu_speed_accuracy_lite.svg)
+
+RTX 4090, WMB test (544 pages). `base` with band attention sustains 24.6 pages/s against
+MinerU-HTML v1.1's 1.16. A 21x gap at matched accuracy (F1 0.9311 vs 0.9306).
+
+### Cross benchmark accuracy
+
+The shipped models are trained on WMB, so WMB is in-domain. The other two columns are
+**zero-shot**, the model never saw WCXB or the multilingual DAnIEL in training. Each column is
+scored with that benchmark's own metric.
+
+| model | WMB (ROUGE-5) | WCXB (word-F1, zero-shot) | DAnIEL (ROUGE-L, zero-shot) |
+|---|---|---|---|
+| `base` (311m-10) | 0.9311 | 0.8633 | 0.9175 |
+| `mini` (int8) | 0.9010 | 0.8474 | 0.8797 |
+
+All results are 3-seed test means.
+
+## Installation
+
+```bash
+pip install htmlsift          # default: the mini model (CPU, ONNX)
+pip install htmlsift[full]    # adds the base 311M model (torch; GPU-capable)
+```
+
+- **`htmlsift`** — the `mini` model only; a small install (onnxruntime + tokenizers + lxml +
+  numpy + huggingface-hub). First use downloads ~230 MB of model files.
+- **`htmlsift[full]`** — adds `torch` + `transformers` to run `base`. First use of `base`
+  downloads ~1 GB.
+
+Requires Python 3.10+. Model files download from the Hugging Face Hub on first use and are
+cached; prefetch with `htmlsift download mini` (or `base`), or set `HF_HUB_OFFLINE=1` to run
+from cache only.
+
+## Usage
+
+```python
+from htmlsift import Extractor
+
+ex = Extractor()                     # defaults to "mini"
+ex.extract(html)                     # -> str (plain text)
+ex.extract(html, output="html")      # keep structure, links, images
+ex.extract(html, output="markdown")  # clean markdown: [text](url), ![alt](src)
+ex.extract(html, with_blocks=True)   # -> (blocks, probs) for callers who want each line and its prediction individually
+```
+
+Module-level convenience (cached singleton):
+
+```python
+import htmlsift
+htmlsift.extract(html)               # default mini
+```
+
+The larger model, from the `[full]` install:
+
+```python
+ex = Extractor("base")               # 311M encoder; uses the GPU when one is present,
+                                     # else CPU (with a warning)
+ex = Extractor("base", device="cuda")  # require the GPU (errors if none is available)
+ex = Extractor("base", device="cpu", threads=4)  # force CPU; `threads` sets parallelism
+```
+
+`mini` is CPU-only (int8 ONNX); asking it for a GPU raises. Only `base` runs on the GPU.
+
+### Example
+
+Nearly 1 MB of Wikipedia HTML ([Photosynthesis](https://en.wikipedia.org/wiki/Photosynthesis))
+in — the boilerplate (nav, sidebar, citations, footer) dropped, the article kept. The verbatim
+head of the markdown (`mini` and `base` return the same here):
+
+```python
+md = Extractor().extract(html, output="markdown")
+```
+
+```markdown
+# Photosynthesis
+
+From Wikipedia, the free encyclopedia
+
+For other uses, see [Photosynthesis (disambiguation)](https://en.wikipedia.org/wiki/Photosynthesis_(disambiguation)).
+
+Schematic of photosynthesis in plants. The [carbohydrates](https://en.wikipedia.org/wiki/Carbohydrate) produced are stored in or used by the plant.
+
+Composite image showing the global distribution of photosynthesis, including both oceanic [phytoplankton](https://en.wikipedia.org/wiki/Phytoplankton) and terrestrial [vegetation](https://en.wikipedia.org/wiki/Vegetation). Dark red and blue-green indicate regions of high photosynthetic activity in the ocean and on land, respectively.
+
+**Photosynthesis** is a [system](https://en.wikipedia.org/wiki/Biological_system) of [biological processes](https://en.wikipedia.org/wiki/Biological_process) by which [photopigment](https://en.wikipedia.org/wiki/Photopigment)-bearing [autotrophic](https://en.wikipedia.org/wiki/Autotroph) [organisms](https://en.wikipedia.org/wiki/Organism), such as most [plants](https://en.wikipedia.org/wiki/Plant), [algae](https://en.wikipedia.org/wiki/Algae) and [cyanobacteria](https://en.wikipedia.org/wiki/Cyanobacteria), convert light energy—typically from sunlight—into the chemical energy necessary to fuel their metabolism…
+```
+
+### Output modes
+
+| `output` | what you get |
+|---|---|
+| `text` (default) | selected content as plain text |
+| `html` | selected DOM, with structure / links / images kept |
+| `markdown` | clean markdown, links and images preserved |
+
+
+## Concept
+
+<details>
+<summary><b>How it works</b></summary>
+
+### Line level labeling
+![htmlsift labels rendered lines, not DOM elements](docs/assets/concept.svg)
+
+Most extractors decide keep-or-drop on HTML *elements*. htmlsift first renders the page to
+markdown *lines*, then decides keep-or-drop on each line. The trade-off is in the figure above, a line-based ceiling sits below the element level gold ceiling but there are far fewer units to classify per page.
+
+### HTML 'signals' utilized
+
+- **Where it sits**: in the DOM - tag, nesting depth, whether it's inside a link.
+- **What it says**: a multilingual embedding of the line's tokens.
+- **The whole-page context**: repeated boilerplate only stands out in contrast to the rest
+  of the page, so every line is judged against all the others by a BiGRU.
+
+![The three signals, and which model tier carries each](docs/assets/signals.svg)
+
+Heuristic extractors carry the first one and a half of these; an LLM carries all three but
+pays to generate the page token by token. htmlsift carries all three in one pass, and outputs predictions without needing a decoder.
+
+![One pass over the page: base vs mini](docs/assets/architecture.svg)
+
+Both tiers share the render, the per-line pooling, and the BiGRU head. They differ only in the text processing section
+- mini converts text to tokens and mean-pools them, has no internal attention allowing fast CPU processing
+- base uses granite r2 311m embedder using only first 10 layers requiring GPU for efficient usage 
+
+
+</details>
+
+<details>
+<summary><b>Training and retraining</b></summary>
+
+Most of the accuracy arrives in the first few hundred labelled pages — the model learns a
+benchmark's labelling policy quickly, so retraining on your own labels is cheap.
+
+![Accuracy vs training-set size, three model arms](docs/assets/data-efficiency.svg)
+
+These models were trained in the **htmlsift-research** repository. To reproduce the boards or
+retrain on your own data, see [htmlsift-research](https://github.com/koivualeksi/htmlsift-research)
+and its `docs/REPRODUCE.md`.
+
+</details>
+
+<details>
+<summary><b>Known limits</b></summary>
+
+A standalone hero image that is its own block renders to no text line, so it can't be selected
+today. Inline images inside a kept block are preserved in `html` and `markdown`.
+
+</details>
+
+## License
+
+Apache-2.0. See [LICENSE](LICENSE).
+
+htmlsift does **not** depend on `html2text`, directly or transitively — rendering and
+serialization are lxml-only, which is what keeps the licence clean. Model weights are
+distributed separately, from the Hugging Face Hub.
+
+## References
+
+- **WMB — WebMainBench.** The training benchmark; ROUGE-5 main-content metric.
+  [opendatalab/WebMainBench](https://github.com/opendatalab/WebMainBench)
+- **WCXB — Web Content Extraction Benchmark.** Foley, *WCXB: A Multi-Type Web Content Extraction
+  Benchmark*. [arXiv:2605.21097](https://arxiv.org/abs/2605.21097) ·
+  [repo](https://github.com/Murrough-Foley/web-content-extraction-benchmark)
+- **DAnIEL.** Multilingual news corpus (Mutuvi et al.,
+  [LREC 2020](https://aclanthology.org/2020.lrec-1.509/)); main-content scoring follows
+  *Multilingual Evaluation of Main Content Extractors for Web Pages*, SIGIR 2025
+  ([ACM](https://dl.acm.org/doi/10.1145/3726302.3730234)).
+