statembed-py 0.1.0__cp313-cp313-win_amd64.whl

statembed_py/__init__.py

@@ -0,0 +1,5 @@
+from .statembed_py import *
+
+__doc__ = statembed_py.__doc__
+if hasattr(statembed_py, "__all__"):
+    __all__ = statembed_py.__all__

statembed_py/__init__.pyi

@@ -0,0 +1,31 @@
+# This file is automatically generated by pyo3_stub_gen
+# ruff: noqa: E501, F401, F403, F405
+
+import builtins
+import typing
+__all__ = [
+    "StaticEmbedding",
+]
+
+@typing.final
+class StaticEmbedding:
+    def __new__(cls, model_dir: builtins.str, normalize: builtins.bool = True) -> StaticEmbedding:
+        r"""
+        Creates a `StaticEmbedding` from a local directory.
+        
+        The directory must contain at least `model.safetensors`. When the
+        `tokenizers` feature is enabled, `tokenizer.json` is also required.
+        
+        # Arguments
+        * `model_dir` - Path to the model directory.
+        * `normalize` - If `True`, output embeddings will be L2-normalized.
+        """
+    def embed_tokens(self, tokens: typing.Sequence[builtins.int]) -> builtins.list[builtins.float]:
+        r"""
+        Generates an embedding for a pre-tokenized sequence of token IDs (use whatever tokenization
+        libary you prefer to generate tokens).
+        
+        The tensor is loaded lazily on first call. Embeddings are mean-pooled
+        and optionally normalized.
+        """
+

statembed_py-0.1.0.dist-info/METADATA

@@ -0,0 +1,82 @@
+Metadata-Version: 2.4
+Name: statembed-py
+Version: 0.1.0
+Classifier: Programming Language :: Rust
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Summary: Fast, lightweight static text embeddings for Python
+License: MIT
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM
+
+# statembed-py
+
+Fast, lightweight static text embeddings.
+
+> _Python bindings for the [`statembed`](../statembed/README.md) Rust crate, built with [PyO3](https://pyo3.rs)._
+
+`statembed-py` loads a static embedding model (a `model.safetensors` file) and produces mean-pooled, optionally L2-normalized embeddings from token IDs. It does not tokenize text itself — pair it with a tokenizer library such as [`tokenizers`](https://pypi.org/project/tokenizers/).
+
+## Installation
+
+```bash
+# with uv
+uv add statembed-py
+# with pip
+pip install statembed-py
+```
+
+### Building from source
+
+Requires [Rust](https://rustup.rs) and [maturin](https://www.maturin.rs):
+
+```bash
+pip install maturin
+maturin develop --release
+```
+
+## Usage
+
+```python
+from functools import lru_cache
+from tokenizers import Tokenizer
+from statembed_py import StaticEmbedding
+
+@lru_cache(maxsize=1)
+def get_embedding_model() -> StaticEmbedding:
+    return StaticEmbedding(model_dir="./my-model")  # must contain model.safetensors
+
+@lru_cache(maxsize=1)
+def get_tokenizer() -> Tokenizer:
+    return Tokenizer.from_file("./my-model/tokenizer.json")
+
+def embed(text: str) -> list[float]:
+    tokens = get_tokenizer().encode(text).ids
+    embedding = get_embedding_model().embed_tokens(tokens)
+    return embedding
+```
+
+## API
+
+### `StaticEmbedding(model_dir, normalize=True)`
+
+Loads a model from a local directory containing `model.safetensors`. The tensor is loaded lazily on the first `embed_tokens` call.
+
+- `model_dir: str` — path to the model directory.
+- `normalize: bool` — if `True` (default), output embeddings are L2-normalized.
+
+### `embed_tokens(tokens: Sequence[int]) -> list[float]`
+
+Mean-pools the embedding rows for the given token IDs into a single fixed-length vector, applying normalization if enabled.
+
+Type stubs (`statembed_py.pyi`) are bundled for editor and type-checker support.
+
+## Development
+
+- `maturin develop` — build and install the extension into the active virtualenv.
+- `cargo run --bin stub_gen` — regenerate `statembed_py.pyi` after changing the PyO3 bindings.
+
+## License
+
+MIT
+

statembed_py-0.1.0.dist-info/RECORD

@@ -0,0 +1,8 @@
+statembed_py/__init__.py,sha256=xUf6OEOLEKMBK0icUjzP-8H2rfQNDBSBu1Ykwpnfp90,131
+statembed_py/__init__.pyi,sha256=1u8nEnH-cgxK_LNZfS-HKsdcolXw3KXwB6gQmSUktmU,1095
+statembed_py/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+statembed_py/statembed_py.cp313-win_amd64.pyd,sha256=CTMXlHgEBZ3YoUY9UAH2jlJYqQ_dE-r-x_r8xm4YLfc,433664
+statembed_py-0.1.0.dist-info/METADATA,sha256=Wux0bJR-D_x6jltW2Tsxh16Qi-uR3sIbJtKX2oNuMnI,2562
+statembed_py-0.1.0.dist-info/WHEEL,sha256=z3sDn4xNPtieBDo9mUKkT1e80gbhCuRsQQi1_g6mdQM,97
+statembed_py-0.1.0.dist-info/sboms/statembed-py.cyclonedx.json,sha256=PHP1D5y2HyKHd4lFnndFQj7iYnMtpAQN7gKnDPV_YdY,154403
+statembed_py-0.1.0.dist-info/RECORD,,

statembed_py-0.1.0.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: maturin (1.14.1)
+Root-Is-Purelib: false
+Tag: cp313-cp313-win_amd64