sciwrite-lint 0.2.0__tar.gz → 0.2.2__tar.gz

{sciwrite_lint-0.2.0 → sciwrite_lint-0.2.2}/PKG-INFO

@@ -1,43 +1,50 @@
 Metadata-Version: 2.4
 Name: sciwrite-lint
-Version: 0.2.0
+Version: 0.2.2
 Summary: A linter for scientific manuscripts — reference verification, consistency checking, and structural validation.
 Author: Sergey Samsonau
+Maintainer-email: Authentic Research Partners LLC <info@arpconnect.com>
 License-Expression: MIT
-Project-URL: Homepage, https://github.com/authentic-research-partners/sciwrite-lint
+Project-URL: Homepage, https://arpconnect.com
 Project-URL: Repository, https://github.com/authentic-research-partners/sciwrite-lint
+Project-URL: Documentation, https://github.com/authentic-research-partners/sciwrite-lint#readme
 Project-URL: Issues, https://github.com/authentic-research-partners/sciwrite-lint/issues
-Keywords: linter,scientific-writing,citation-verification,latex
+Project-URL: Paper, https://arxiv.org/abs/2604.08501
+Keywords: linter,latex,bibtex,pdf,scientific-writing,academic-writing,citation-checker,reference-verification,claim-verification,manuscript,peer-review,arxiv,vllm,qwen
 Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
 Classifier: Intended Audience :: Science/Research
-Classifier: Topic :: Text Processing :: Markup :: LaTeX
+Classifier: Topic :: Scientific/Engineering
+Classifier: Topic :: Software Development :: Quality Assurance
+Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Typing :: Typed
 Requires-Python: <3.14,>=3.13
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: httpx>=0.27
-Requires-Dist: tenacity>=8.0
-Requires-Dist: trafilatura>=1.8
-Requires-Dist: aiosqlite>=0.20
+Requires-Dist: httpx>=0.28
+Requires-Dist: tenacity>=9.0
+Requires-Dist: trafilatura>=2.0
+Requires-Dist: aiosqlite>=0.22
 Requires-Dist: loguru>=0.7
 Requires-Dist: anyascii>=0.3
-Requires-Dist: rich>=13.0
+Requires-Dist: rich>=14.0
 Requires-Dist: defusedxml>=0.7
-Requires-Dist: pydantic>=2.0
+Requires-Dist: pydantic>=2.10
 Requires-Dist: rapidfuzz>=3.0
 Requires-Dist: bibtexparser>=1.4
-Requires-Dist: pdfplumber>=0.10
-Requires-Dist: openai>=1.0
+Requires-Dist: pdfplumber>=0.11
+Requires-Dist: openai>=2.0
 Requires-Dist: grobid-tei-xml>=0.1
-Requires-Dist: sentence-transformers>=3.0
-Requires-Dist: numpy>=1.24
+Requires-Dist: sentence-transformers>=5.0
+Requires-Dist: numpy>=2.0
 Requires-Dist: sqlite-vec>=0.1.6
-Requires-Dist: torch>=2.0
-Requires-Dist: transformers>=4.45
-Requires-Dist: Pillow>=10.0
-Requires-Dist: pypdf>=4.0
-Requires-Dist: pypdfium2>=4.0
-Requires-Dist: pdf2image>=1.16
+Requires-Dist: torch>=2.5
+Requires-Dist: transformers>=4.50
+Requires-Dist: Pillow>=12.0
+Requires-Dist: pypdf>=6.0
+Requires-Dist: pypdfium2>=5.0
+Requires-Dist: pdf2image>=1.17
 Provides-Extra: dev
 Requires-Dist: pytest>=8.0; extra == "dev"
 Requires-Dist: pytest-xdist>=3.0; extra == "dev"
@@ -45,6 +52,8 @@ Dynamic: license-file
 
 # sciwrite-lint
 
+[![arXiv](https://img.shields.io/badge/arXiv-2604.08501-b31b1b.svg)](https://arxiv.org/abs/2604.08501)
+
 A linter for scientific manuscripts. Checks that your references exist, your metadata is accurate, your citations support the claims you make, and your cited papers' own bibliographies are real. Works on LaTeX and PDF. Runs entirely on your machine. Produces a **SciLint Score**.
 
 The only open-source tool that combines reference verification, claim checking, manuscript consistency, figure analysis, and bibliography verification in one pipeline — powered entirely by a local LLM on your own GPU.
@@ -123,7 +132,7 @@ sciwrite-lint contributions paper.pdf --format json
 
 ### Optimizations
 
-Three models — a vision model (Qwen3-VL-2B), an embedding model (Arctic Embed), and an 8B reasoning LLM (Qwen3 via vLLM) — share a single consumer GPU. Models run in separate pipeline stages; on WSL2, CUDA memory virtualization pages idle allocations to system RAM, letting all three share physical VRAM. FP8 weights and KV cache (Ada Lovelace+) and per-paper SQLite caching with hash-based invalidation are baseline. On top of that:
+Three models — a vision model (Qwen3-VL-2B default, or 8B FP8 via `--vision-backend vllm` for +15% accuracy), an embedding model (Arctic Embed), and an 8B reasoning LLM (Qwen3 via vLLM) — share a single consumer GPU. Models run in separate pipeline stages; on WSL2, CUDA memory virtualization pages idle allocations to system RAM, letting all three share physical VRAM. FP8 weights and KV cache (Ada Lovelace+) and per-paper SQLite caching with hash-based invalidation are baseline. On top of that:
 
 - **Semantic section filtering** — embedding-based KNN retrieval sends only the ~5 most relevant sections per claim to the LLM, reducing LLM calls ~4x
 - **Prefix-first prompt structure** — shared context placed before variable content in all prompts, maximizing vLLM's automatic prefix caching
@@ -134,11 +143,11 @@ Three models — a vision model (Qwen3-VL-2B), an embedding model (Arctic Embed)
 - **8-source full-text cascade** — early exit on first successful download across arXiv, Semantic Scholar, OpenAlex, PubMed Central, Europe PMC, Unpaywall, bioRxiv, CORE
 - **Live monitoring** *(advanced)* — `sciwrite-lint containers monitor` shows service health, VRAM usage, and KV cache utilization in a terminal dashboard
 
-Full pipeline on a 50-reference paper: ~30 minutes initial (dominated by downloads and claim verification), minutes on cached reruns. On native Linux, embedding and vision default to CPU (tens of times slower) due to limited VRAM on consumer GPUs and no transparent paging of inactive vLLM container allocations; force GPU via config if VRAM permits.
+Full pipeline on a 50-reference paper: ~30 minutes initial (dominated by downloads and claim verification), minutes on cached reruns. On native Linux, the pipeline automatically swaps vLLM containers to free GPU for embedding and vision stages (~50x faster than CPU). *(Native Linux GPU swap is preliminary — tested on WSL2 only; expected to work, may need minor fixes.)*
 
 ## Install
 
-**Assumed setup:** A workstation with an NVIDIA GPU (16+ GB VRAM). Developed and tested on Windows (WSL2). Native Linux is likely to work with GPU memory allocation tuning (see [docs/services.md](docs/services.md#embedding-device)). Not tested on macOS.
+**Assumed setup:** A workstation with an NVIDIA GPU (16+ GB VRAM). Developed and tested on Windows (WSL2). Native Linux is likely to work with GPU memory allocation tuning (see [docs/services.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/services.md#embedding-device)). Not tested on macOS.
 
 Requires [uv](https://docs.astral.sh/uv/getting-started/installation/), a container runtime (podman or docker), CUDA drivers, and [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html).
 
@@ -148,6 +157,32 @@ uv tool install sciwrite-lint --python 3.13
 
 uv downloads Python 3.13 automatically (does not affect any Python you may already have) and installs sciwrite-lint as a globally available command.
 
+### Reproducible environment (optional)
+
+`pyproject.toml` declares lower bounds for dependencies, so `uv tool install` / `pip install` may resolve newer versions as upstream releases ship. If you want the exact versions the maintainer develops and tests against — for reproducing benchmarks, debugging version-drift issues, or CI — install from [`requirements-pinned.txt`](requirements-pinned.txt) instead.
+
+**What's in it:** top-level packages declared in `pyproject.toml` pinned to `==` versions. Transitive dependencies are not pinned — pip resolves them within each top-level package's own constraints. This is a deliberate middle ground: it locks the packages sciwrite-lint actually imports and tests against, without over-constraining the dep tree.
+
+```bash
+# Clone the public repo (required — uv tool install can't read a local file)
+git clone https://github.com/authentic-research-partners/sciwrite-lint.git
+cd sciwrite-lint
+
+# With uv (recommended — fast, isolated, manages Python itself)
+uv venv --python 3.13 .venv
+source .venv/bin/activate
+uv pip install -r requirements-pinned.txt
+uv pip install -e . --no-deps    # install sciwrite-lint source; deps already pinned
+
+# Or with plain pip
+python3.13 -m venv .venv
+source .venv/bin/activate
+pip install -r requirements-pinned.txt
+pip install -e . --no-deps
+```
+
+Activate the venv in each new terminal with `source .venv/bin/activate`. For most users, the plain `uv tool install sciwrite-lint` above is simpler — use the pinned file only when you need exact version reproduction.
+
 ## Setup
 
 ```bash
@@ -157,11 +192,22 @@ sciwrite-lint containers start         # start GROBID + vLLM (needs GPU for vLLM
 sciwrite-lint containers monitor       # live dashboard: service health, VRAM, KV cache
 ```
 
-![Monitor dashboard](docs/monitor.png)
+![Monitor dashboard](https://github.com/authentic-research-partners/sciwrite-lint/raw/main/docs/monitor.png)
 
 `init` detects `.tex` files and their `.bib` references and generates a `.sciwrite-lint.toml` config. Review to confirm the right files were detected.
 
-**Paywalled references:** drop PDFs into `local_pdfs/` with filenames matching the reference title. The tool fuzzy-matches filenames against your `.bib` titles and uses local copies instead of downloading.
+**Providing PDFs manually (`local_pdfs/`):** Drop any reference PDF into the `local_pdfs/` directory to skip the download step entirely. This is useful when you already have the file, when the paper is paywalled, or when the publisher blocks automated downloads (Cell, Springer, some PMC pages). Name files by the reference title:
+
+```
+local_pdfs/
+├── States of Curiosity Modulate Hippocampus-Dependent Learning.pdf
+├── Mind in Society.pdf
+└── The System of Professions.pdf
+```
+
+The tool fuzzy-matches filenames against your `.bib` titles (threshold 0.80) and copies matches to the workspace. On the next run, matched references upgrade from T2 to T1 and go through GROBID parsing and claim verification like any other full-text reference.
+
+**Finding what to download:** `sciwrite-lint verify` lists T2 references that are confirmed open access with direct URLs — open in your browser and save to `local_pdfs/`. For paywalled papers, use your institution's library access.
 
 **Optional API keys** increase rate limits for Semantic Scholar, NCBI, and CORE:
 
@@ -170,7 +216,7 @@ sciwrite-lint config show              # see what's configured
 sciwrite-lint config set-key semantic-scholar YOUR_KEY  # dedicated rate limit
 ```
 
-See [docs/services.md](docs/services.md) for GPU requirements, all external APIs, and API key details.
+See [docs/services.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/services.md) for GPU requirements, all external APIs, and API key details.
 
 ## Usage
 
@@ -256,12 +302,27 @@ Output formats: terminal (default) or `--format json`.
 
 - `sciwrite-lint checks` — list all checks
 - `sciwrite-lint <command> --help` — detailed usage for any command
-- [docs/services.md](docs/services.md) — GROBID, vLLM, external APIs, configuration
+- [docs/services.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/services.md) — GROBID, vLLM, external APIs, configuration
 
 For contributors and advanced users:
 
-- [docs/evals.md](docs/evals.md) — detection evaluation framework, adding test cases
-- [docs/calibration.md](docs/calibration.md) — SciLint Score calibration against ground-truth papers
+- [docs/evals.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/evals.md) — detection evaluation framework, adding test cases
+- [docs/calibration.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/calibration.md) — SciLint Score calibration against ground-truth papers
+
+## Citation
+
+If you use sciwrite-lint in your research, please cite the accompanying paper ([arXiv:2604.08501](https://arxiv.org/abs/2604.08501)):
+
+```bibtex
+@article{samsonau2026sciwritelint,
+  title   = {sciwrite-lint: Verification Infrastructure for the Age of Science Vibe-Writing},
+  author  = {Samsonau, Sergey V.},
+  journal = {arXiv preprint arXiv:2604.08501},
+  year    = {2026},
+}
+```
+
+GitHub's "Cite this repository" button (powered by [`CITATION.cff`](CITATION.cff)) will also generate this for you.
 
 ## License
 

sciwrite_lint-0.2.0/sciwrite_lint.egg-info/PKG-INFO → sciwrite_lint-0.2.2/README.md

@@ -1,50 +1,7 @@
-Metadata-Version: 2.4
-Name: sciwrite-lint
-Version: 0.2.0
-Summary: A linter for scientific manuscripts — reference verification, consistency checking, and structural validation.
-Author: Sergey Samsonau
-License-Expression: MIT
-Project-URL: Homepage, https://github.com/authentic-research-partners/sciwrite-lint
-Project-URL: Repository, https://github.com/authentic-research-partners/sciwrite-lint
-Project-URL: Issues, https://github.com/authentic-research-partners/sciwrite-lint/issues
-Keywords: linter,scientific-writing,citation-verification,latex
-Classifier: Development Status :: 3 - Alpha
-Classifier: Intended Audience :: Science/Research
-Classifier: Topic :: Text Processing :: Markup :: LaTeX
-Classifier: Programming Language :: Python :: 3.13
-Requires-Python: <3.14,>=3.13
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: httpx>=0.27
-Requires-Dist: tenacity>=8.0
-Requires-Dist: trafilatura>=1.8
-Requires-Dist: aiosqlite>=0.20
-Requires-Dist: loguru>=0.7
-Requires-Dist: anyascii>=0.3
-Requires-Dist: rich>=13.0
-Requires-Dist: defusedxml>=0.7
-Requires-Dist: pydantic>=2.0
-Requires-Dist: rapidfuzz>=3.0
-Requires-Dist: bibtexparser>=1.4
-Requires-Dist: pdfplumber>=0.10
-Requires-Dist: openai>=1.0
-Requires-Dist: grobid-tei-xml>=0.1
-Requires-Dist: sentence-transformers>=3.0
-Requires-Dist: numpy>=1.24
-Requires-Dist: sqlite-vec>=0.1.6
-Requires-Dist: torch>=2.0
-Requires-Dist: transformers>=4.45
-Requires-Dist: Pillow>=10.0
-Requires-Dist: pypdf>=4.0
-Requires-Dist: pypdfium2>=4.0
-Requires-Dist: pdf2image>=1.16
-Provides-Extra: dev
-Requires-Dist: pytest>=8.0; extra == "dev"
-Requires-Dist: pytest-xdist>=3.0; extra == "dev"
-Dynamic: license-file
-
 # sciwrite-lint
 
+[![arXiv](https://img.shields.io/badge/arXiv-2604.08501-b31b1b.svg)](https://arxiv.org/abs/2604.08501)
+
 A linter for scientific manuscripts. Checks that your references exist, your metadata is accurate, your citations support the claims you make, and your cited papers' own bibliographies are real. Works on LaTeX and PDF. Runs entirely on your machine. Produces a **SciLint Score**.
 
 The only open-source tool that combines reference verification, claim checking, manuscript consistency, figure analysis, and bibliography verification in one pipeline — powered entirely by a local LLM on your own GPU.
@@ -123,7 +80,7 @@ sciwrite-lint contributions paper.pdf --format json
 
 ### Optimizations
 
-Three models — a vision model (Qwen3-VL-2B), an embedding model (Arctic Embed), and an 8B reasoning LLM (Qwen3 via vLLM) — share a single consumer GPU. Models run in separate pipeline stages; on WSL2, CUDA memory virtualization pages idle allocations to system RAM, letting all three share physical VRAM. FP8 weights and KV cache (Ada Lovelace+) and per-paper SQLite caching with hash-based invalidation are baseline. On top of that:
+Three models — a vision model (Qwen3-VL-2B default, or 8B FP8 via `--vision-backend vllm` for +15% accuracy), an embedding model (Arctic Embed), and an 8B reasoning LLM (Qwen3 via vLLM) — share a single consumer GPU. Models run in separate pipeline stages; on WSL2, CUDA memory virtualization pages idle allocations to system RAM, letting all three share physical VRAM. FP8 weights and KV cache (Ada Lovelace+) and per-paper SQLite caching with hash-based invalidation are baseline. On top of that:
 
 - **Semantic section filtering** — embedding-based KNN retrieval sends only the ~5 most relevant sections per claim to the LLM, reducing LLM calls ~4x
 - **Prefix-first prompt structure** — shared context placed before variable content in all prompts, maximizing vLLM's automatic prefix caching
@@ -134,11 +91,11 @@ Three models — a vision model (Qwen3-VL-2B), an embedding model (Arctic Embed)
 - **8-source full-text cascade** — early exit on first successful download across arXiv, Semantic Scholar, OpenAlex, PubMed Central, Europe PMC, Unpaywall, bioRxiv, CORE
 - **Live monitoring** *(advanced)* — `sciwrite-lint containers monitor` shows service health, VRAM usage, and KV cache utilization in a terminal dashboard
 
-Full pipeline on a 50-reference paper: ~30 minutes initial (dominated by downloads and claim verification), minutes on cached reruns. On native Linux, embedding and vision default to CPU (tens of times slower) due to limited VRAM on consumer GPUs and no transparent paging of inactive vLLM container allocations; force GPU via config if VRAM permits.
+Full pipeline on a 50-reference paper: ~30 minutes initial (dominated by downloads and claim verification), minutes on cached reruns. On native Linux, the pipeline automatically swaps vLLM containers to free GPU for embedding and vision stages (~50x faster than CPU). *(Native Linux GPU swap is preliminary — tested on WSL2 only; expected to work, may need minor fixes.)*
 
 ## Install
 
-**Assumed setup:** A workstation with an NVIDIA GPU (16+ GB VRAM). Developed and tested on Windows (WSL2). Native Linux is likely to work with GPU memory allocation tuning (see [docs/services.md](docs/services.md#embedding-device)). Not tested on macOS.
+**Assumed setup:** A workstation with an NVIDIA GPU (16+ GB VRAM). Developed and tested on Windows (WSL2). Native Linux is likely to work with GPU memory allocation tuning (see [docs/services.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/services.md#embedding-device)). Not tested on macOS.
 
 Requires [uv](https://docs.astral.sh/uv/getting-started/installation/), a container runtime (podman or docker), CUDA drivers, and [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html).
 
@@ -148,6 +105,32 @@ uv tool install sciwrite-lint --python 3.13
 
 uv downloads Python 3.13 automatically (does not affect any Python you may already have) and installs sciwrite-lint as a globally available command.
 
+### Reproducible environment (optional)
+
+`pyproject.toml` declares lower bounds for dependencies, so `uv tool install` / `pip install` may resolve newer versions as upstream releases ship. If you want the exact versions the maintainer develops and tests against — for reproducing benchmarks, debugging version-drift issues, or CI — install from [`requirements-pinned.txt`](requirements-pinned.txt) instead.
+
+**What's in it:** top-level packages declared in `pyproject.toml` pinned to `==` versions. Transitive dependencies are not pinned — pip resolves them within each top-level package's own constraints. This is a deliberate middle ground: it locks the packages sciwrite-lint actually imports and tests against, without over-constraining the dep tree.
+
+```bash
+# Clone the public repo (required — uv tool install can't read a local file)
+git clone https://github.com/authentic-research-partners/sciwrite-lint.git
+cd sciwrite-lint
+
+# With uv (recommended — fast, isolated, manages Python itself)
+uv venv --python 3.13 .venv
+source .venv/bin/activate
+uv pip install -r requirements-pinned.txt
+uv pip install -e . --no-deps    # install sciwrite-lint source; deps already pinned
+
+# Or with plain pip
+python3.13 -m venv .venv
+source .venv/bin/activate
+pip install -r requirements-pinned.txt
+pip install -e . --no-deps
+```
+
+Activate the venv in each new terminal with `source .venv/bin/activate`. For most users, the plain `uv tool install sciwrite-lint` above is simpler — use the pinned file only when you need exact version reproduction.
+
 ## Setup
 
 ```bash
@@ -157,11 +140,22 @@ sciwrite-lint containers start         # start GROBID + vLLM (needs GPU for vLLM
 sciwrite-lint containers monitor       # live dashboard: service health, VRAM, KV cache
 ```
 
-![Monitor dashboard](docs/monitor.png)
+![Monitor dashboard](https://github.com/authentic-research-partners/sciwrite-lint/raw/main/docs/monitor.png)
 
 `init` detects `.tex` files and their `.bib` references and generates a `.sciwrite-lint.toml` config. Review to confirm the right files were detected.
 
-**Paywalled references:** drop PDFs into `local_pdfs/` with filenames matching the reference title. The tool fuzzy-matches filenames against your `.bib` titles and uses local copies instead of downloading.
+**Providing PDFs manually (`local_pdfs/`):** Drop any reference PDF into the `local_pdfs/` directory to skip the download step entirely. This is useful when you already have the file, when the paper is paywalled, or when the publisher blocks automated downloads (Cell, Springer, some PMC pages). Name files by the reference title:
+
+```
+local_pdfs/
+├── States of Curiosity Modulate Hippocampus-Dependent Learning.pdf
+├── Mind in Society.pdf
+└── The System of Professions.pdf
+```
+
+The tool fuzzy-matches filenames against your `.bib` titles (threshold 0.80) and copies matches to the workspace. On the next run, matched references upgrade from T2 to T1 and go through GROBID parsing and claim verification like any other full-text reference.
+
+**Finding what to download:** `sciwrite-lint verify` lists T2 references that are confirmed open access with direct URLs — open in your browser and save to `local_pdfs/`. For paywalled papers, use your institution's library access.
 
 **Optional API keys** increase rate limits for Semantic Scholar, NCBI, and CORE:
 
@@ -170,7 +164,7 @@ sciwrite-lint config show              # see what's configured
 sciwrite-lint config set-key semantic-scholar YOUR_KEY  # dedicated rate limit
 ```
 
-See [docs/services.md](docs/services.md) for GPU requirements, all external APIs, and API key details.
+See [docs/services.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/services.md) for GPU requirements, all external APIs, and API key details.
 
 ## Usage
 
@@ -256,12 +250,27 @@ Output formats: terminal (default) or `--format json`.
 
 - `sciwrite-lint checks` — list all checks
 - `sciwrite-lint <command> --help` — detailed usage for any command
-- [docs/services.md](docs/services.md) — GROBID, vLLM, external APIs, configuration
+- [docs/services.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/services.md) — GROBID, vLLM, external APIs, configuration
 
 For contributors and advanced users:
 
-- [docs/evals.md](docs/evals.md) — detection evaluation framework, adding test cases
-- [docs/calibration.md](docs/calibration.md) — SciLint Score calibration against ground-truth papers
+- [docs/evals.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/evals.md) — detection evaluation framework, adding test cases
+- [docs/calibration.md](https://github.com/authentic-research-partners/sciwrite-lint/blob/main/docs/calibration.md) — SciLint Score calibration against ground-truth papers
+
+## Citation
+
+If you use sciwrite-lint in your research, please cite the accompanying paper ([arXiv:2604.08501](https://arxiv.org/abs/2604.08501)):
+
+```bibtex
+@article{samsonau2026sciwritelint,
+  title   = {sciwrite-lint: Verification Infrastructure for the Age of Science Vibe-Writing},
+  author  = {Samsonau, Sergey V.},
+  journal = {arXiv preprint arXiv:2604.08501},
+  year    = {2026},
+}
+```
+
+GitHub's "Cite this repository" button (powered by [`CITATION.cff`](CITATION.cff)) will also generate this for you.
 
 ## License
 

{sciwrite_lint-0.2.0 → sciwrite_lint-0.2.2}/pyproject.toml

@@ -4,49 +4,56 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sciwrite-lint"
-version = "0.2.0"
+version = "0.2.2"
 description = "A linter for scientific manuscripts — reference verification, consistency checking, and structural validation."
 requires-python = ">=3.13,<3.14"
 license = "MIT"
 authors = [{ name = "Sergey Samsonau" }]
+maintainers = [{ name = "Authentic Research Partners LLC", email = "info@arpconnect.com" }]
 readme = "README.md"
-keywords = ["linter", "scientific-writing", "citation-verification", "latex"]
+keywords = ["linter", "latex", "bibtex", "pdf", "scientific-writing", "academic-writing", "citation-checker", "reference-verification", "claim-verification", "manuscript", "peer-review", "arxiv", "vllm", "qwen"]
 classifiers = [
     "Development Status :: 3 - Alpha",
+    "Intended Audience :: Developers",
     "Intended Audience :: Science/Research",
-    "Topic :: Text Processing :: Markup :: LaTeX",
+    "Topic :: Scientific/Engineering",
+    "Topic :: Software Development :: Quality Assurance",
+    "Programming Language :: Python :: 3",
     "Programming Language :: Python :: 3.13",
+    "Typing :: Typed",
 ]
 dependencies = [
-    "httpx>=0.27",
-    "tenacity>=8.0",
-    "trafilatura>=1.8",
-    "aiosqlite>=0.20",
+    "httpx>=0.28",
+    "tenacity>=9.0",
+    "trafilatura>=2.0",
+    "aiosqlite>=0.22",
     "loguru>=0.7",
     "anyascii>=0.3",
-    "rich>=13.0",
+    "rich>=14.0",
     "defusedxml>=0.7",
-    "pydantic>=2.0",
+    "pydantic>=2.10",
     "rapidfuzz>=3.0",
     "bibtexparser>=1.4",
-    "pdfplumber>=0.10",
-    "openai>=1.0",
+    "pdfplumber>=0.11",
+    "openai>=2.0",
     "grobid-tei-xml>=0.1",
-    "sentence-transformers>=3.0",
-    "numpy>=1.24",
+    "sentence-transformers>=5.0",
+    "numpy>=2.0",
     "sqlite-vec>=0.1.6",
-    "torch>=2.0",
-    "transformers>=4.45",
-    "Pillow>=10.0",
-    "pypdf>=4.0",
-    "pypdfium2>=4.0",
-    "pdf2image>=1.16",
+    "torch>=2.5",
+    "transformers>=4.50",
+    "Pillow>=12.0",
+    "pypdf>=6.0",
+    "pypdfium2>=5.0",
+    "pdf2image>=1.17",
 ]
 
 [project.urls]
-Homepage = "https://github.com/authentic-research-partners/sciwrite-lint"
+Homepage = "https://arpconnect.com"
 Repository = "https://github.com/authentic-research-partners/sciwrite-lint"
+Documentation = "https://github.com/authentic-research-partners/sciwrite-lint#readme"
 Issues = "https://github.com/authentic-research-partners/sciwrite-lint/issues"
+Paper = "https://arxiv.org/abs/2604.08501"
 
 [project.optional-dependencies]
 dev = [

{sciwrite_lint-0.2.0 → sciwrite_lint-0.2.2}/sciwrite_lint/__init__.py

@@ -1,3 +1,3 @@
 """sciwrite-lint: a linter for scientific manuscripts."""
 
-__version__ = "0.2.0"
+__version__ = "0.2.2"

{sciwrite_lint-0.2.0 → sciwrite_lint-0.2.2}/sciwrite_lint/__main__.py

@@ -18,6 +18,10 @@ from loguru import logger
 
 from sciwrite_lint import __version__
 from sciwrite_lint.config import LintConfig, PaperConfig, load_config
+from sciwrite_lint.vllm.vllm_server import MODELS as _VLLM_MODELS
+
+_ALL_MODELS = sorted(_VLLM_MODELS.keys())
+_TEXT_MODELS = [k for k, v in _VLLM_MODELS.items() if v["kind"] == "text"]
 
 
 # ---------------------------------------------------------------------------
@@ -179,6 +183,12 @@ def main(argv: list[str] | None = None) -> int:
         "limit. Above 4, the live monitor may fail to track progress and "
         "vLLM/API throughput plateaus.",
     )
+    p_check.add_argument(
+        "--vision-backend",
+        choices=["transformers", "vllm"],
+        default=None,
+        help="Vision backend: transformers (2B, default) or vllm (8B FP8 on port 5002)",
+    )
     p_check.set_defaults(func=run_check)
 
     # --- checks ---
@@ -296,7 +306,7 @@ def main(argv: list[str] | None = None) -> int:
     )
     p_vc.add_argument(
         "--model",
-        choices=["qwen3", "gemma3"],
+        choices=_TEXT_MODELS,
         default="",
         help="vLLM model preset (default: qwen3)",
     )
@@ -381,14 +391,20 @@ def main(argv: list[str] | None = None) -> int:
     # --- vision ---
     p_vision = sub.add_parser(
         "vision",
-        help="Extract and describe manuscript figures via Qwen3-VL-2B (also runs automatically in check pipeline).",
+        help="Extract and describe manuscript figures (transformers 2B or vLLM 8B FP8).",
     )
     p_vision.add_argument("--paper", required=True, help="Paper name from [[papers]]")
     p_vision.add_argument(
         "--device",
         choices=["auto", "cpu", "cuda"],
-        default="auto",
-        help="Device for VL inference (default: auto — CUDA if available)",
+        default=None,
+        help="Device for VL inference (default: auto — CUDA if available, transformers only)",
+    )
+    p_vision.add_argument(
+        "--backend",
+        choices=["transformers", "vllm"],
+        default=None,
+        help="Vision backend: transformers (2B, default) or vllm (8B FP8 on port 5002)",
     )
     p_vision.add_argument(
         "--fresh",
@@ -407,9 +423,7 @@ def main(argv: list[str] | None = None) -> int:
         choices=["start", "stop", "restart", "status", "monitor"],
         help="start/stop/restart/status/monitor of all containers",
     )
-    p_containers.add_argument(
-        "--model", choices=["qwen3", "gemma3"], help="vLLM model to use"
-    )
+    p_containers.add_argument("--model", choices=_ALL_MODELS, help="vLLM model to use")
     p_containers.add_argument(
         "--update",
         action="store_true",
@@ -427,6 +441,11 @@ def main(argv: list[str] | None = None) -> int:
         action="store_true",
         help="Remove and recreate containers (applies config changes like memory limits)",
     )
+    p_containers.add_argument(
+        "--vision",
+        action="store_true",
+        help="Also start vision vLLM (qwen3-vl-8b-fp8 on port 5002)",
+    )
     p_containers.add_argument("--config", help="Path to .sciwrite-lint.toml")
     p_containers.set_defaults(func=run_containers)
 
@@ -452,9 +471,7 @@ def main(argv: list[str] | None = None) -> int:
     p_vllm_status.set_defaults(func=run_vllm)
 
     p_vllm_start = vllm_sub.add_parser("start", help="Start vLLM server")
-    p_vllm_start.add_argument(
-        "--model", choices=["qwen3", "gemma3"], help="Model to serve"
-    )
+    p_vllm_start.add_argument("--model", choices=_ALL_MODELS, help="Model to serve")
     p_vllm_start.add_argument(
         "--update",
         action="store_true",
@@ -464,13 +481,11 @@ def main(argv: list[str] | None = None) -> int:
     p_vllm_start.set_defaults(func=run_vllm)
 
     p_vllm_stop = vllm_sub.add_parser("stop", help="Stop vLLM server")
-    p_vllm_stop.add_argument(
-        "--model", choices=["qwen3", "gemma3"], help="Model to stop"
-    )
+    p_vllm_stop.add_argument("--model", choices=_ALL_MODELS, help="Model to stop")
     p_vllm_stop.set_defaults(func=run_vllm)
 
     p_vllm_logs = vllm_sub.add_parser("logs", help="Show vLLM container logs")
-    p_vllm_logs.add_argument("--model", choices=["qwen3", "gemma3"])
+    p_vllm_logs.add_argument("--model", choices=_ALL_MODELS)
     p_vllm_logs.add_argument(
         "-f", "--follow", action="store_true", help="Follow log output"
     )
@@ -478,7 +493,7 @@ def main(argv: list[str] | None = None) -> int:
     p_vllm_logs.set_defaults(func=run_vllm)
 
     p_vllm_rm = vllm_sub.add_parser("rm", help="Remove vLLM container")
-    p_vllm_rm.add_argument("--model", choices=["qwen3", "gemma3"])
+    p_vllm_rm.add_argument("--model", choices=_ALL_MODELS)
     p_vllm_rm.add_argument(
         "--force", action="store_true", help="Force remove even if running"
     )

{sciwrite_lint-0.2.0 → sciwrite_lint-0.2.2}/sciwrite_lint/checks/cross_section_consistency.py

@@ -44,6 +44,9 @@ For each item, set is_genuine to true ONLY if the values or claims \
 CONFLICT. Set is_genuine to false if the passages agree or discuss \
 different topics.
 
+Report at most 4 contradictions. If you find more, return only the 4 \
+most clear-cut cases.
+
 Reply with JSON: {"contradictions": [\
 {"type": "number|claim|framing", "section_a_says": "...", \
 "section_b_says": "...", "explanation": "...", "is_genuine": true/false}]}