quire-grammar-model 0.2.0__tar.gz

quire_grammar_model-0.2.0/PKG-INFO

@@ -0,0 +1,44 @@
+Metadata-Version: 2.4
+Name: quire-grammar-model
+Version: 0.2.0
+Summary: Bundled phase-2 edit-tagger for quire-grammar (data only, no code).
+Author: Roger Cooper
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://github.com/jutreuter/quire-grammar
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+
+# quire-grammar-model
+
+The bundled phase-2 edit-tagger for
+[quire-grammar](https://github.com/jutreuter/quire-grammar) — **data only, no
+code**.
+
+```
+pip install quire-grammar[model]     # pulls this + onnxruntime + tokenizers
+```
+
+`quire_grammar.check()` runs rules-only without it and activates the model
+when it's present — "off if missing", like a spell dictionary.
+
+## What's inside
+
+| file | what |
+|---|---|
+| `tagger.onnx` | GECToR-style token tagger, int8, ~66 MB |
+| `tokenizer.json` | the matching `tokenizers` (WordPiece) definition |
+| `tag_vocab.json` | the edit-tag label list |
+
+`quire_grammar_model.path()` returns the directory; `available()` says whether
+every file is present.
+
+## Provenance
+
+Fine-tuned from `distilbert-base-cased` (Apache-2.0; pretrained on Wikipedia +
+BooksCorpus) on synthetic native-writer errors injected over public-domain
+prose (Project Gutenberg). No non-commercial or research-only corpus is in the
+training set. Licensed **Apache-2.0**.
+
+The model file is regenerated from the training pipeline in the main repo
+(`train/` → `train.export --out-dir packages/quire-grammar-model/src/quire_grammar_model/`),
+not edited by hand.

quire_grammar_model-0.2.0/README.md

@@ -0,0 +1,34 @@
+# quire-grammar-model
+
+The bundled phase-2 edit-tagger for
+[quire-grammar](https://github.com/jutreuter/quire-grammar) — **data only, no
+code**.
+
+```
+pip install quire-grammar[model]     # pulls this + onnxruntime + tokenizers
+```
+
+`quire_grammar.check()` runs rules-only without it and activates the model
+when it's present — "off if missing", like a spell dictionary.
+
+## What's inside
+
+| file | what |
+|---|---|
+| `tagger.onnx` | GECToR-style token tagger, int8, ~66 MB |
+| `tokenizer.json` | the matching `tokenizers` (WordPiece) definition |
+| `tag_vocab.json` | the edit-tag label list |
+
+`quire_grammar_model.path()` returns the directory; `available()` says whether
+every file is present.
+
+## Provenance
+
+Fine-tuned from `distilbert-base-cased` (Apache-2.0; pretrained on Wikipedia +
+BooksCorpus) on synthetic native-writer errors injected over public-domain
+prose (Project Gutenberg). No non-commercial or research-only corpus is in the
+training set. Licensed **Apache-2.0**.
+
+The model file is regenerated from the training pipeline in the main repo
+(`train/` → `train.export --out-dir packages/quire-grammar-model/src/quire_grammar_model/`),
+not edited by hand.

quire_grammar_model-0.2.0/pyproject.toml

@@ -0,0 +1,27 @@
+[build-system]
+requires = ["setuptools>=68"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "quire-grammar-model"
+version = "0.2.0"
+description = "Bundled phase-2 edit-tagger for quire-grammar (data only, no code)."
+readme = "README.md"
+requires-python = ">=3.11"
+license = "Apache-2.0"
+authors = [{ name = "Roger Cooper" }]
+# Data only. The consumer (`quire-grammar[model]`) brings onnxruntime +
+# tokenizers; this package must stay dependency-free.
+dependencies = []
+
+[project.urls]
+Homepage = "https://github.com/jutreuter/quire-grammar"
+
+[tool.setuptools]
+package-dir = { "" = "src" }
+
+[tool.setuptools.packages.find]
+where = ["src"]
+
+[tool.setuptools.package-data]
+quire_grammar_model = ["tagger.onnx", "tokenizer.json", "tag_vocab.json", "py.typed"]

quire_grammar_model-0.2.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

quire_grammar_model-0.2.0/src/quire_grammar_model/__init__.py

@@ -0,0 +1,33 @@
+"""quire-grammar-model — the bundled phase-2 edit-tagger, data only.
+
+Install it with the ``quire-grammar[model]`` extra. There is no code here:
+``quire_grammar.model`` locates ``tagger.onnx`` / ``tokenizer.json`` /
+``tag_vocab.json`` through :func:`path`, and runs them with ``onnxruntime`` +
+``tokenizers`` (also pulled by the extra).
+
+The model is a GECToR-style token tagger fine-tuned from ``distilbert-base-cased``
+(Apache-2.0) on synthetic errors over public-domain prose — hence the
+Apache-2.0 licence on this package.
+"""
+
+from __future__ import annotations
+
+from importlib.resources import files
+from importlib.resources.abc import Traversable
+
+__version__ = "0.2.0"
+
+#: the files that make up a usable model
+FILES: tuple[str, ...] = ("tagger.onnx", "tokenizer.json", "tag_vocab.json")
+
+
+def path() -> Traversable:
+    """The directory that holds :data:`FILES`."""
+    return files(__name__)
+
+
+def available() -> bool:
+    """True if every file in :data:`FILES` is present (a build may ship the
+    package skeleton without the large ``tagger.onnx``)."""
+    root = path()
+    return all((root / name).is_file() for name in FILES)