mirror of
https://github.com/tiennm99/DocsGPT.git
synced 2026-10-03 20:12:55 +00:00
Alex
574f96341e
The backend import package is now docsgpt, the name it will carry on PyPI; application was far too generic to install into anyone's site-packages. git mv plus a mechanical rewrite of every import, dotted string and path reference: 734 Python files, the compose files, Dockerfile, workflows, docs, setup scripts, devcontainer, k8s manifests, vscode config, pytest and coverage config, .gitignore. Behaviour is unchanged. Kept for one release: - A top-level application package whose meta-path finder resolves application.x.y to the already-imported docsgpt.x.y object, so old imports and entry points (celery -A application.app.celery, uvicorn application.asgi:asgi_app) keep working with a FutureWarning. - Celery registers every application.* task name as an alias of its docsgpt.* task on start-up, so messages queued by the previous release still run. The redbeat key prefix moves to redbeat:docsgpt:v2: so schedule entries the previous release wrote are left unread instead of firing twice. The backend image builds from the repository root (docker build -f docsgpt/Dockerfile .) so it can ship the alias package; a root .dockerignore allow-lists docsgpt/ and application/ and keeps caches, local data, .env files, the sample index files and the Dockerfile out. Compose and the image workflows point at the new context.
58 lines
2.1 KiB
Python
58 lines
2.1 KiB
Python
"""String-keyed registry for chunking strategies.
|
|
|
|
Mirrors ``RetrieverCreator``: features register new strategies (``recursive``,
|
|
``markdown``, ``parent_child``, ...) without touching the dispatch site. The
|
|
classic strategy is registered under ``classic_chunk`` by ``chunking.py``.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from typing import Type
|
|
|
|
|
|
class ChunkerCreator:
|
|
chunkers: dict[str, Type] = {}
|
|
_strategies_loaded: bool = False
|
|
|
|
@classmethod
|
|
def _ensure_builtin(cls) -> None:
|
|
"""Register built-in chunkers if they are not registered yet.
|
|
|
|
Self-bootstraps so ``create_chunker`` works regardless of import order:
|
|
``docsgpt.parser.chunking`` registers ``classic_chunk`` and
|
|
``docsgpt.parser.chunking_strategies`` registers ``recursive`` /
|
|
``markdown`` / ``parent_child``.
|
|
"""
|
|
if not cls.chunkers:
|
|
import docsgpt.parser.chunking # noqa: F401 (registers classic_chunk)
|
|
if not cls._strategies_loaded:
|
|
cls._strategies_loaded = True
|
|
import docsgpt.parser.chunking_strategies # noqa: F401
|
|
|
|
@classmethod
|
|
def create_chunker(cls, strategy: str, *args, **kwargs):
|
|
"""Instantiate the chunker registered under ``strategy``.
|
|
|
|
Args:
|
|
strategy: Registry key (e.g. ``classic_chunk``).
|
|
*args: Positional args forwarded to the chunker constructor.
|
|
**kwargs: Keyword args forwarded to the chunker constructor.
|
|
|
|
Returns:
|
|
A chunker instance exposing ``chunk(documents) -> List[Document]``.
|
|
|
|
Raises:
|
|
ValueError: If no chunker is registered for ``strategy``.
|
|
"""
|
|
cls._ensure_builtin()
|
|
key = (strategy or "classic_chunk").lower()
|
|
chunker_class = cls.chunkers.get(key)
|
|
if not chunker_class:
|
|
raise ValueError(f"No chunker class found for strategy {strategy}")
|
|
return chunker_class(*args, **kwargs)
|
|
|
|
@classmethod
|
|
def register(cls, key: str, chunker_class: Type) -> None:
|
|
"""Register ``chunker_class`` under ``key`` (idempotent)."""
|
|
cls.chunkers[key] = chunker_class
|