bindings: the libraries in process, for Python

crates/bindings/python/catena/__init__.py

@@ -0,0 +1,35 @@
+"""The catena libraries, in process.
+
+Every function here is the same code the command-line tool runs, and hands
+back the same record shapes it writes to stdout — dictionaries rather than
+lines of JSON.
+
+    >>> import catena
+    >>> chunks = catena.chunk_markdown(markdown, target=512)
+    >>> chunks[0]["page"]
+    41
+"""
+
+from ._catena import (
+    __version__,
+    chunk_markdown,
+    count_tokens,
+    fuse_hits,
+    identify,
+    pack_cues,
+    parse_subtitles,
+    scrub,
+    split_collection,
+)
+
+__all__ = [
+    "__version__",
+    "chunk_markdown",
+    "count_tokens",
+    "fuse_hits",
+    "identify",
+    "pack_cues",
+    "parse_subtitles",
+    "scrub",
+    "split_collection",
+]

crates/bindings/python/catena/__init__.pyi

@@ -0,0 +1,32 @@
+from typing import Any, Iterable, Mapping, Sequence
+
+__version__: str
+
+def chunk_markdown(
+    markdown: str,
+    *,
+    target: int | None = ...,
+    overlap: int | None = ...,
+    min_tokens: int | None = ...,
+    hard_max: int | None = ...,
+) -> list[dict[str, Any]]: ...
+def scrub(markdown: str, *, pages: bool = ...) -> str: ...
+def count_tokens(text: str) -> int: ...
+def parse_subtitles(
+    content: str, *, keep_repeats: bool = ...
+) -> list[dict[str, Any]]: ...
+def pack_cues(
+    cues: Iterable[Mapping[str, Any]],
+    *,
+    seconds: float | None = ...,
+    max_tokens: int | None = ...,
+) -> list[dict[str, Any]]: ...
+def split_collection(markdown: str) -> list[dict[str, Any]]: ...
+def fuse_hits(
+    hits: Iterable[Mapping[str, Any]],
+    *,
+    k: float | None = ...,
+    weights: Mapping[str, float] | None = ...,
+    limit: int | None = ...,
+) -> list[dict[str, Any]]: ...
+def identify(name: str, data: bytes) -> dict[str, Any]: ...