mirror of
https://github.com/tiennm99/DocsGPT.git
synced 2026-10-03 09:12:55 +00:00
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.
128 lines
4.3 KiB
YAML
128 lines
4.3 KiB
YAML
name: docsgpt-oss
|
|
services:
|
|
frontend:
|
|
build:
|
|
context: ../frontend
|
|
# Vite dev server with hot reload over the bind mount below. The default
|
|
# target (what Docker Hub publishes) is a static build behind nginx.
|
|
target: dev
|
|
volumes:
|
|
- ../frontend/src:/app/src
|
|
environment:
|
|
# Every VITE_* the app reads. A bare name is passed through only when it is
|
|
# set in the shell or the --env-file, so an unset one does not reach the
|
|
# container as an empty string and override the image's own default.
|
|
- VITE_API_HOST=http://localhost:7091
|
|
- VITE_API_STREAMING=${VITE_API_STREAMING:-true}
|
|
- VITE_BASE_URL
|
|
- VITE_GOOGLE_CLIENT_ID
|
|
- VITE_GOOGLE_PICKER_API_KEY
|
|
- VITE_SHARE_POINT_CLIENT_ID
|
|
- VITE_CONFLUENCE_CLIENT_ID
|
|
- VITE_NOTIFICATION_TEXT
|
|
- VITE_NOTIFICATION_LINK
|
|
- VITE_ENABLE_VOICE_INPUT
|
|
- VITE_DISABLE_SOURCE_FE
|
|
- VITE_USE_V
|
|
ports:
|
|
- "5173:5173"
|
|
depends_on:
|
|
- backend
|
|
|
|
backend:
|
|
user: root
|
|
build:
|
|
context: ..
|
|
dockerfile: docsgpt/Dockerfile
|
|
args:
|
|
# Optional extras to bake in (comma-separated): docling, milvus. The
|
|
# docling extra brings the layout-model parser/OCR backend and its
|
|
# models. INSTALL_DOCLING=true is the older spelling of EXTRAS=docling.
|
|
EXTRAS: ${EXTRAS:-}
|
|
INSTALL_DOCLING: ${INSTALL_DOCLING:-false}
|
|
# Bake the tesseract binary behind OCR_ENABLED=true (~35 MB): set
|
|
# INSTALL_TESSERACT=true in ../.env or the shell (setup.sh does this
|
|
# when OCR is enabled). Off by default, like docling. Upgrading a
|
|
# deployment that already runs OCR_ENABLED=true with tesseract: add
|
|
# INSTALL_TESSERACT=true to ../.env before rebuilding, or scanned
|
|
# pages fail with an install hint.
|
|
INSTALL_TESSERACT: ${INSTALL_TESSERACT:-false}
|
|
env_file:
|
|
- ../.env
|
|
environment:
|
|
# Override URLs to use docker service names
|
|
- CELERY_BROKER_URL=redis://redis:6379/0
|
|
- CELERY_RESULT_BACKEND=redis://redis:6379/1
|
|
- CACHE_REDIS_URL=redis://redis:6379/2
|
|
- POSTGRES_URI=postgresql://docsgpt:docsgpt@postgres:5432/docsgpt
|
|
ports:
|
|
- "7091:7091"
|
|
volumes:
|
|
- ../docsgpt/indexes:/app/indexes
|
|
- ../docsgpt/inputs:/app/inputs
|
|
- ../docsgpt/vectors:/app/vectors
|
|
depends_on:
|
|
redis:
|
|
condition: service_started
|
|
postgres:
|
|
condition: service_healthy
|
|
|
|
worker:
|
|
user: root
|
|
build:
|
|
context: ..
|
|
dockerfile: docsgpt/Dockerfile
|
|
args:
|
|
EXTRAS: ${EXTRAS:-}
|
|
INSTALL_DOCLING: ${INSTALL_DOCLING:-false}
|
|
INSTALL_TESSERACT: ${INSTALL_TESSERACT:-false}
|
|
# Consumes the default queue AND the dedicated `parsing` (read_document /
|
|
# parse_document) and `embeddings` (query embedding) queues. Without `parsing`
|
|
# the read_document await never resolves; without `embeddings` every search
|
|
# fails after EMBEDDINGS_DELEGATE_TIMEOUT, because EMBEDDINGS_DELEGATE_TO_WORKER
|
|
# is on by default. For heavy/OCR parsing run a separate worker with `-Q parsing`;
|
|
# to keep query latency off the ingest pool, another with `-Q embeddings`.
|
|
command: celery -A docsgpt.app.celery worker -l INFO -B -Q docsgpt,parsing,embeddings
|
|
env_file:
|
|
- ../.env
|
|
environment:
|
|
# Override URLs to use docker service names
|
|
- CELERY_BROKER_URL=redis://redis:6379/0
|
|
- CELERY_RESULT_BACKEND=redis://redis:6379/1
|
|
- API_URL=http://backend:7091
|
|
- CACHE_REDIS_URL=redis://redis:6379/2
|
|
- POSTGRES_URI=postgresql://docsgpt:docsgpt@postgres:5432/docsgpt
|
|
volumes:
|
|
- ../docsgpt/indexes:/app/indexes
|
|
- ../docsgpt/inputs:/app/inputs
|
|
- ../docsgpt/vectors:/app/vectors
|
|
depends_on:
|
|
redis:
|
|
condition: service_started
|
|
postgres:
|
|
condition: service_healthy
|
|
|
|
redis:
|
|
image: redis:6-alpine
|
|
ports:
|
|
- 6379:6379
|
|
|
|
postgres:
|
|
image: postgres:16-alpine
|
|
environment:
|
|
- POSTGRES_USER=docsgpt
|
|
- POSTGRES_PASSWORD=docsgpt
|
|
- POSTGRES_DB=docsgpt
|
|
ports:
|
|
- "5432:5432"
|
|
volumes:
|
|
- postgres_data:/var/lib/postgresql/data
|
|
healthcheck:
|
|
test: ["CMD-SHELL", "pg_isready -U docsgpt -d docsgpt"]
|
|
interval: 5s
|
|
timeout: 5s
|
|
retries: 10
|
|
|
|
volumes:
|
|
postgres_data:
|