Docker containerization with baked HuggingFace model cache

.dockerignore

@@ -0,0 +1,10 @@
+venv/
+__pycache__/
+*.pyc
+.env
+qdrant_storage/
+bm25_index.pkl
+data/
+chroma_db/
+*.pkl
+.git/

Dockerfile

@@ -0,0 +1,29 @@
+FROM python:3.11-slim
+
+WORKDIR /app
+
+# Cài system deps cho underthesea + EasyOCR
+RUN apt-get update && apt-get install -y \
+    gcc g++ git libgl1 libglib2.0-0 \
+    && rm -rf /var/lib/apt/lists/*
+
+# Cài Python deps trước (cache layer)
+COPY requirements.txt .
+RUN pip install --no-cache-dir --trusted-host pypi.org --trusted-host files.pythonhosted.org --timeout 300 -r requirements.txt
+
+# Copy source code
+COPY src/ ./src/
+COPY static/ ./static/
+COPY api.py .
+
+# Copy HuggingFace model cache
+COPY hf_models/ /root/.cache/huggingface/hub/
+ENV TRANSFORMERS_OFFLINE=1
+ENV HF_DATASETS_OFFLINE=1
+
+# Qdrant storage và BM25 index mount từ volume
+VOLUME ["/app/qdrant_storage", "/app/data"]
+
+EXPOSE 8000
+
+CMD ["uvicorn", "api:app", "--host", "0.0.0.0", "--port", "8000"]

docker-compose.yml

@@ -0,0 +1,16 @@
+services:
+  app:
+    build: .
+    ports:
+      - "8000:8000"
+    volumes:
+      - ./qdrant_storage:/app/qdrant_storage
+      - ./bm25_index.pkl:/app/bm25_index.pkl
+      - ./data:/app/data
+      - hf_cache:/root/.cache/huggingface
+    env_file:
+      - .env
+    restart: unless-stopped
+
+volumes:
+  hf_cache:

requirements.txt

@@ -1,20 +1,15 @@
-# LLM & Embedding
-anthropic>=0.40.0
-google-generativeai>=0.8.0
-openai>=1.50.0
-sentence-transformers>=3.0.0
-
-# Vector store
-chromadb>=0.5.0
-
-# Document parsing
-pypdf>=4.0.0
-python-docx>=1.1.0
-
-# Data
-pandas>=2.0.0
-numpy>=1.26.0
-
-# Utils
-python-dotenv>=1.0.0
-tqdm>=4.66.0
+--extra-index-url https://download.pytorch.org/whl/cpu
+torch==2.5.1+cpu
+fastapi==0.136.3
+uvicorn==0.49.0
+python-dotenv==1.2.2
+google-genai==2.8.0
+groq==1.4.0
+qdrant-client==1.18.0
+rank-bm25==0.2.2
+sentence-transformers==2.7.0
+underthesea==9.5.0
+pypdf==6.13.0
+easyocr==1.7.2
+tqdm==4.68.1
+pydantic==2.13.4