mirror of
https://github.com/tiennm99/DocsGPT.git
synced 2026-10-05 18:14:21 +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.
65 lines
1.8 KiB
Python
65 lines
1.8 KiB
Python
"""Secret redaction for reflected ``stack_logs`` data.
|
|
|
|
``stacks`` is built by reflecting every public attribute of runtime
|
|
objects (the ``llm`` component carries the deployment provider
|
|
``api_key`` and the caller's ``user_api_key``). The unified-logs endpoint
|
|
returns ``stacks`` to the client, so credentials must be scrubbed — at
|
|
write time for new rows, and at read time for rows written before
|
|
redaction existed.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
REDACTED = "[REDACTED]"
|
|
|
|
# Substrings marking a key as a credential. Compound token-count fields
|
|
# (``prompt_tokens`` / ``generated_tokens`` / ``token_budget`` / ...) are
|
|
# intentionally not matched: secret token forms are spelled out
|
|
# (``access_token`` etc.) and a bare ``token`` is handled separately.
|
|
_SECRET_SUBSTRINGS = (
|
|
"api_key",
|
|
"apikey",
|
|
"api_token",
|
|
"access_token",
|
|
"refresh_token",
|
|
"auth_token",
|
|
"id_token",
|
|
"session_token",
|
|
"secret",
|
|
"password",
|
|
"passwd",
|
|
"passphrase",
|
|
"private_key",
|
|
"credential",
|
|
"authorization",
|
|
"bearer",
|
|
)
|
|
|
|
|
|
def is_secret_key(key: str) -> bool:
|
|
"""True when ``key`` names a credential that must not be persisted/returned."""
|
|
k = key.lower()
|
|
if k == "token":
|
|
return True
|
|
return any(s in k for s in _SECRET_SUBSTRINGS)
|
|
|
|
|
|
def redact_secrets(obj):
|
|
"""Recursively replace secret-keyed values with ``[REDACTED]``.
|
|
|
|
Walks dicts and lists; leaf scalars pass through unchanged. ``None``
|
|
returns ``None`` so callers can redact an optional payload directly.
|
|
"""
|
|
if isinstance(obj, dict):
|
|
return {
|
|
k: (
|
|
REDACTED
|
|
if isinstance(k, str) and is_secret_key(k)
|
|
else redact_secrets(v)
|
|
)
|
|
for k, v in obj.items()
|
|
}
|
|
if isinstance(obj, list):
|
|
return [redact_secrets(v) for v in obj]
|
|
return obj
|