Switch to UV from PDM

.env.example

@@ -13,4 +13,7 @@ DB_URL=jdbc:h2:../databases/rag
 # If using CAII, fill these in:
 CAII_DOMAIN=
 CAII_INFERENCE_ENDPOINT_NAME=
-CAII_EMBEDDING_ENDPOINT_NAME=
+CAII_EMBEDDING_ENDPOINT_NAME=
+
+# set this to true if you have uv installed on your system, other wise don't include this
+USE_SYSTEM_UV=true

.github/workflows/llm_checks.yml

@@ -14,37 +14,31 @@ jobs:
         with:
           ref: ${{ github.event.inputs.BRANCH }}
 
-      - name: Set up Python
-        uses: pdm-project/setup-pdm@v3
+      - name: Install uv
+        uses: astral-sh/setup-uv@v3
         with:
-          python-version: '3.10'
-          cache: false
-
-      - name: Cache pdm cache folder
-        uses: actions/cache@v4
-        id: python_cache
-        with:
-          path: ./llm-service/.venv
-          key: pdm-venv-${{ hashFiles('./llm-service/pdm.lock') }}
+          enable-cache: true
+          cache-dependency-glob: "./llm-service/uv.lock"
 
       - name: Install Python dependencies
-        if: steps.python_cache.outputs.cache-hit != 'true'
         run: |
-          pdm install
+          uv python install 3.10
+          uv venv --python 3.10
+          uv sync
         working-directory: llm-service
 
       - name: Run ruff
         run: |
-          pdm run ruff check app
+          uv run ruff check app
         working-directory: llm-service
 
       - name: Run mypy
         run: |
-          pdm run mypy app
+          uv run mypy app
         working-directory: llm-service
 
       - name: Test with pytest
         run: |
-          pdm run pytest -sxvvra
+          uv run pytest -sxvvra
         working-directory: llm-service
 

README.md

@@ -86,12 +86,13 @@ the Node service locally, you can do so by following these steps:
 
 #### Python Setup
 
-- Install Python (via [pyenv](https://github.com/pyenv/pyenv), probably) (directly via brew, if you must)
+- Install Python 3.10 (via [pyenv](https://github.com/pyenv/pyenv), probably) (directly via brew, if you must)
 - `cd llm-service`
-- Create a virtual environment (`python -m venv venv; source venv/bin/activate`)
-- Install dependencies (`python -m pip install -r requirements.txt`)
-- `fastapi dev`
-  - ends up running on port 8000
+- Install `uv`. 
+  - We recommend installing via `brew install uv`, but you can also install it directly in your python environment if you prefer.
+- `uv sync` - this creates a `uv` virtual environment in `.venv` and installs the dependencies
+- `uv fastapi dev`
+  - the python-based service ends up running on port 8000
 
 #### Java Setup
 

llm-service/Dockerfile

@@ -1,11 +1,15 @@
 FROM docker-private.infra.cloudera.com/cloudera_base/hardened/cloudera-python:3.10
-RUN pip install pdm
+RUN pip install uv
 COPY ./pyproject.toml /app/
-COPY ./pdm.lock /app/
+COPY ./uv.lock /app/
 WORKDIR /app
-RUN pdm install
+RUN uv sync
 
 COPY ./app /app
-WORKDIR /app
+COPY ./scripts /scripts
+ENV PATH="/app/.venv/bin:$PATH"
+
+# Reset the entrypoint, don't invoke `uv`
+ENTRYPOINT []
 
-CMD [ "pdm", "run", "fastapi", "run", "--host", "0.0.0.0" ]
+CMD [ "fastapi", "run", "--host", "0.0.0.0" ]

llm-service/pyproject.toml

@@ -1,7 +1,7 @@
 [project]
 name = "llm-service"
 version = "0.1.0"
-description = "Default template for PDM package"
+description = "Default template for UV package"
 authors = [
     {name = "Conrado Silva Miranda", email = "[email protected]"},
 ]
@@ -11,7 +11,7 @@ dependencies = [
     "fastapi==0.111.0",
     "pydantic==2.8.2",
     "pydantic-settings==2.3.4",
-    "boto3>=1.35.66",
+    "boto3>=1.35.68",
     "llama-index-embeddings-bedrock==0.2.1",
     "llama-index-llms-bedrock==0.1.13",
     "llama-index-llms-openai==0.1.31",

local-dev.sh

@@ -63,13 +63,15 @@ mkdir -p databases
 docker run --name qdrant_dev --rm -d -p 6333:6333 -p 6334:6334 -v $(pwd)/databases/qdrant_storage:/qdrant/storage:z qdrant/qdrant
 
 cd llm-service
-python3.10 -m venv venv
-source venv/bin/activate
-python -m pip install pdm
-pdm install
-pdm run pytest -sxvvra
+if [ -z "$USE_SYSTEM_UV" ]; then
+  python3.10 -m venv venv
+  source venv/bin/activate
+  python -m pip install uv
+fi
+uv sync
+uv run pytest -sxvvra
 
-pdm run fastapi dev &
+uv run fastapi dev &
 
 # start up the jarva
 cd ../backend

scripts/refresh_project.sh

@@ -46,8 +46,9 @@ cd ui/express
 npm install
 
 cd ../../llm-service
-pip install pdm
-pdm install -v
+pip install uv
+
+uv sync
 
 cd ..
 mkdir -p artifacts

scripts/startup_app.sh

@@ -61,7 +61,7 @@ qdrant/qdrant &
 
 # start Python backend
 cd llm-service
-pdm run fastapi run --host 127.0.0.1 --port 8081 &
+uv run fastapi run --host 127.0.0.1 --port 8081 &
 
 # start up the jarva
 cd ..