chore: S3_MODEL_PATH as variable

.env.sample

@@ -0,0 +1,5 @@
+HUGGING_FACE_TOKEN="xxxxxxxxxxxxxxx"
+S3_ACCESS_KEY_ID="xxxxxxxxxxxxxxx"
+S3_SECRET_ACCESS_KEY="xxxxxxxxxxxxxxx"
+S3_ENDPOINT_URL="https://s3.gra.io.cloud.ovh.net/"
+S3_MODEL_PATH="distilgpt2"

.gitignore

@@ -15,3 +15,5 @@ build/
 coverage.json
 diff_amendments.csv
 grouped_matching.csv
+test.sh
+.env

docker-compose.yaml

@@ -0,0 +1,14 @@
+services:
+  llm-inference-server:
+    build:
+      context: .
+      dockerfile: dockerfile
+    env_file:
+      - .env
+    volumes:
+      - my_docker_volume:/app/models
+    ports:
+      - "8000:8000"
+
+volumes:
+  my_docker_volume:

dockerfile

@@ -25,19 +25,26 @@ WORKDIR /app
 # Copy only requirements
 COPY pyproject.toml poetry.lock* /app/
 
-# Install project dependencies
+# Install project dependencies (including dev dependencies)
 RUN poetry config virtualenvs.create false \
-    && poetry install --no-dev --no-interaction --no-ansi \
+    && poetry install --no-interaction --no-ansi \
     && rm -rf ${POETRY_CACHE_DIR}
 
 # Start a new stage for a smaller final image
 FROM python:3.11-slim-buster
 
-# Copy installed packages and application code from builder
-COPY --from=builder /usr/local/lib/python3.11/site-packages /usr/local/lib/python3.11/site-packages
-COPY --from=builder /usr/local/bin /usr/local/bin
+# Copy the entire Python environment and application code from builder
+COPY --from=builder /usr/local /usr/local
 COPY --from=builder /app /app
 
+# Set environment variables
+ENV PYTHONUNBUFFERED=1 \
+    POETRY_HOME="/opt/poetry" \
+    POETRY_VENV="/opt/poetry-venv" \
+    POETRY_CACHE_DIR="/opt/.cache" \
+    PYTHONPATH="/app" \
+    PATH="${PATH}:/opt/poetry/bin"
+
 # Set working directory
 WORKDIR /app
 
@@ -54,5 +61,5 @@ ARG INFERENCE_SERVER=llm_inference.s3_inference_server
 # Set an environment variable using the argument
 ENV INFERENCE_SERVER=${INFERENCE_SERVER}
 
-# Run the application
-CMD ["sh", "-c", "poetry run python -m ${INFERENCE_SERVER}"]
+# Run the application directly with Python (without Poetry)
+CMD python -m ${INFERENCE_SERVER}

llm_inference/s3_inference_server.py

@@ -13,7 +13,7 @@
 
 # S3 configuration
 BUCKET_NAME = "vllm-cache"
-S3_MODEL_PATH = "Meta-Llama-3.1-8B"
+S3_MODEL_PATH = os.getenv("S3_MODEL_PATH")
 local_model_dir = f"models/{S3_MODEL_PATH}"
 s3_access_key_id = os.getenv("S3_ACCESS_KEY_ID")
 s3_secret_access_key = os.getenv("S3_SECRET_ACCESS_KEY")