Use uv pip install instead of pip install (#921)

## What does this PR do? 

See issue: #747 -- `uv` is just plain better. This PR does the bare
minimum of replacing `pip install` by `uv pip install` and ensuring `uv`
exists in the environment.

## Test Plan 

First: create new conda, `uv pip install -e .` on `llama-stack` -- all
is good.
Next: run `llama stack build --template together` followed by `llama
stack run together` -- all good
Next: run `llama stack build --template together --image-name yoyo`
followed by `llama stack run together --image-name yoyo` -- all good
Next: fresh conda and `uv pip install -e .` and `llama stack build
--template together --image-type venv` -- all good.

Docker: `llama stack build --template together --image-type container`
works!

.github/workflows/pre-commit.yml

@@ -11,10 +11,10 @@ jobs:
 
     steps:
       - name: Checkout code
-        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938 # v4.2.0
+        uses: actions/checkout@v4
 
       - name: Set up Python
-        uses: actions/setup-python@f677139bbe7f9c59b41e40162b753c062f5d49a3 # v5.2.0
+        uses: actions/setup-python@v5
         with:
           python-version: '3.11'
           cache: pip
@@ -22,4 +22,4 @@ jobs:
             **/requirements*.txt
             .pre-commit-config.yaml
 
-      - uses: pre-commit/action@2c7b3805fd2a0fd8c1884dcaebf91fc102a13ecd #v3.0.1
+      - uses: pre-commit/action@v3.0.1

.pre-commit-config.yaml

@@ -5,10 +5,8 @@ default_language_version:
 
 repos:
 -   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: 6306a48f7dae5861702d573c9c247e4e9498e867
+    rev: v5.0.0  # Latest stable version
     hooks:
-    -   id: trailing-whitespace
-    -   id: check-ast
     -   id: check-merge-conflict
     -   id: check-added-large-files
         args: ['--maxkb=1000']
@@ -28,23 +26,32 @@ repos:
           - --license-filepath
           - docs/license_header.txt
 
--   repo: https://github.com/pycqa/flake8
+-   repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: 34cbf8ef3950f43d09b85e2e45c15ae5717dc37b
+    rev: v0.9.4
     hooks:
-    -   id: flake8
+    -   id: ruff
-        additional_dependencies:
+        args: [
-          - flake8-bugbear == 22.4.25
+            --fix,
-          - pep8-naming == 0.12.1
+            --exit-non-zero-on-fix
-          - torchfix
+        ]
-        args: ['--config=.flake8']
+    -   id: ruff-format
 
--   repo: https://github.com/omnilib/ufmt
+-   repo: https://github.com/adamchainz/blacken-docs
-    rev: v2.7.0
+    rev: 1.19.0
     hooks:
-    -   id: ufmt
+    -   id: blacken-docs
         additional_dependencies:
-          - black == 24.4.2
+        - black==24.3.0
-          - usort == 1.0.8
+
+# -   repo: https://github.com/pre-commit/mirrors-mypy
+#     rev: v1.14.0
+#     hooks:
+#     -   id: mypy
+#         additional_dependencies:
+#           - types-requests
+#           - types-setuptools
+#           - pydantic
+#         args: [--ignore-missing-imports]
 
 # - repo: https://github.com/jsh9/pydoclint
 #   rev: d88180a8632bb1602a4d81344085cf320f288c5a
@@ -71,3 +78,7 @@ repos:
 #         require_serial: true
 #         files: ^llama_stack/templates/.*$
 #         stages: [manual]
+
+ci:
+    autofix_commit_msg: 🎨 [pre-commit.ci] Auto format from pre-commit.com hooks
+    autoupdate_commit_msg: ⬆ [pre-commit.ci] pre-commit autoupdate

llama_stack/cli/stack/_build.py

@@ -183,9 +183,9 @@ def run_stack_build_command(
             build_config.distribution_spec.providers
         )
         normal_deps += SERVER_DEPENDENCIES
-        print(f"pip install {' '.join(normal_deps)}")
+        print(f"uv pip install {' '.join(normal_deps)}")
         for special_dep in special_deps:
-            print(f"pip install {special_dep}")
+            print(f"uv pip install {special_dep}")
         return
 
     _run_stack_build_command_from_build_config(

llama_stack/distribution/build.py

@@ -95,11 +95,11 @@ def print_pip_install_help(providers: Dict[str, List[Provider]]):
     normal_deps, special_deps = get_provider_dependencies(providers)
 
     cprint(
-        f"Please install needed dependencies using the following commands:\n\npip install {' '.join(normal_deps)}",
+        f"Please install needed dependencies using the following commands:\n\nuv pip install {' '.join(normal_deps)}",
         "yellow",
     )
     for special_dep in special_deps:
-        cprint(f"pip install {special_dep}", "yellow")
+        cprint(f"uv pip install {special_dep}", "yellow")
     print()
 
 

llama_stack/distribution/build_conda_env.sh

@@ -78,10 +78,12 @@ ensure_conda_env_python310() {
   eval "$(conda shell.bash hook)"
   conda deactivate && conda activate "${env_name}"
 
+  $CONDA_PREFIX/bin/pip install uv
+
   if [ -n "$TEST_PYPI_VERSION" ]; then
     # these packages are damaged in test-pypi, so install them first
-    $CONDA_PREFIX/bin/pip install fastapi libcst
+    uv pip install fastapi libcst
-    $CONDA_PREFIX/bin/pip install --extra-index-url https://test.pypi.org/simple/ \
+    uv pip install --extra-index-url https://test.pypi.org/simple/ \
       llama-models==$TEST_PYPI_VERSION \
       llama-stack-client==$TEST_PYPI_VERSION \
       llama-stack==$TEST_PYPI_VERSION \
@@ -90,7 +92,7 @@ ensure_conda_env_python310() {
       IFS='#' read -ra parts <<<"$special_pip_deps"
       for part in "${parts[@]}"; do
         echo "$part"
-        $CONDA_PREFIX/bin/pip install $part
+        uv pip install $part
       done
     fi
   else
@@ -102,7 +104,7 @@ ensure_conda_env_python310() {
       fi
 
       printf "Installing from LLAMA_STACK_DIR: $LLAMA_STACK_DIR\n"
-      $CONDA_PREFIX/bin/pip install --no-cache-dir -e "$LLAMA_STACK_DIR"
+      uv pip install --no-cache-dir -e "$LLAMA_STACK_DIR"
     else
       PYPI_VERSION="${PYPI_VERSION:-}"
       if [ -n "$PYPI_VERSION" ]; then
@@ -110,7 +112,7 @@ ensure_conda_env_python310() {
       else
         SPEC_VERSION="llama-stack"
       fi
-      $CONDA_PREFIX/bin/pip install --no-cache-dir $SPEC_VERSION
+      uv pip install --no-cache-dir $SPEC_VERSION
     fi
 
     if [ -n "$LLAMA_MODELS_DIR" ]; then
@@ -120,18 +122,18 @@ ensure_conda_env_python310() {
       fi
 
       printf "Installing from LLAMA_MODELS_DIR: $LLAMA_MODELS_DIR\n"
-      $CONDA_PREFIX/bin/pip uninstall -y llama-models
+      uv pip uninstall -y llama-models
-      $CONDA_PREFIX/bin/pip install --no-cache-dir -e "$LLAMA_MODELS_DIR"
+      uv pip install --no-cache-dir -e "$LLAMA_MODELS_DIR"
     fi
 
     # Install pip dependencies
     printf "Installing pip dependencies\n"
-    $CONDA_PREFIX/bin/pip install $pip_dependencies
+    uv pip install $pip_dependencies
     if [ -n "$special_pip_deps" ]; then
       IFS='#' read -ra parts <<<"$special_pip_deps"
       for part in "${parts[@]}"; do
         echo "$part"
-        $CONDA_PREFIX/bin/pip install $part
+        uv pip install $part
       done
     fi
   fi

llama_stack/distribution/build_container.sh

@@ -76,6 +76,8 @@ RUN apt-get update && apt-get install -y \
        bubblewrap \
        && rm -rf /var/lib/apt/lists/*
 
+ENV UV_SYSTEM_PYTHON=1
+RUN pip install uv
 EOF
 fi
 
@@ -83,7 +85,7 @@ fi
 # so we can reuse layers.
 if [ -n "$pip_dependencies" ]; then
   add_to_container << EOF
-RUN pip install --no-cache $pip_dependencies
+RUN uv pip install --no-cache $pip_dependencies
 EOF
 fi
 
@@ -91,7 +93,7 @@ if [ -n "$special_pip_deps" ]; then
   IFS='#' read -ra parts <<<"$special_pip_deps"
   for part in "${parts[@]}"; do
     add_to_container <<EOF
-RUN pip install --no-cache $part
+RUN uv pip install --no-cache $part
 EOF
   done
 fi
@@ -109,16 +111,16 @@ if [ -n "$LLAMA_STACK_DIR" ]; then
   # so that changes will be reflected in the container without having to do a
   # rebuild. This is just for development convenience.
   add_to_container << EOF
-RUN pip install --no-cache -e $stack_mount
+RUN uv pip install --no-cache -e $stack_mount
 EOF
 else
   if [ -n "$TEST_PYPI_VERSION" ]; then
     # these packages are damaged in test-pypi, so install them first
     add_to_container << EOF
-RUN pip install fastapi libcst
+RUN uv pip install fastapi libcst
 EOF
     add_to_container << EOF
-RUN pip install --no-cache --extra-index-url https://test.pypi.org/simple/ \
+RUN uv pip install --no-cache --extra-index-url https://test.pypi.org/simple/ \
   llama-models==$TEST_PYPI_VERSION llama-stack-client==$TEST_PYPI_VERSION llama-stack==$TEST_PYPI_VERSION
 
 EOF
@@ -129,7 +131,7 @@ EOF
       SPEC_VERSION="llama-stack"
     fi
     add_to_container << EOF
-RUN pip install --no-cache $SPEC_VERSION
+RUN uv pip install --no-cache $SPEC_VERSION
 EOF
   fi
 fi
@@ -141,8 +143,8 @@ if [ -n "$LLAMA_MODELS_DIR" ]; then
   fi
 
   add_to_container << EOF
-RUN pip uninstall -y llama-models
+RUN uv pip uninstall -y llama-models
-RUN pip install --no-cache $models_mount
+RUN uv pip install --no-cache $models_mount
 
 EOF
 fi

llama_stack/distribution/build_venv.sh

@@ -51,17 +51,18 @@ run() {
   local pip_dependencies="$2"
   local special_pip_deps="$3"
 
+  pip install uv
   if [ -n "$TEST_PYPI_VERSION" ]; then
     # these packages are damaged in test-pypi, so install them first
-    pip install fastapi libcst
+    uv pip install fastapi libcst
-    pip install --extra-index-url https://test.pypi.org/simple/ \
+    uv pip install --extra-index-url https://test.pypi.org/simple/ \
       llama-models==$TEST_PYPI_VERSION llama-stack==$TEST_PYPI_VERSION \
       $pip_dependencies
     if [ -n "$special_pip_deps" ]; then
       IFS='#' read -ra parts <<<"$special_pip_deps"
       for part in "${parts[@]}"; do
         echo "$part"
-        pip install $part
+        uv pip install $part
       done
     fi
   else
@@ -73,9 +74,9 @@ run() {
       fi
 
       printf "Installing from LLAMA_STACK_DIR: $LLAMA_STACK_DIR\n"
-      pip install --no-cache-dir -e "$LLAMA_STACK_DIR"
+      uv pip install --no-cache-dir -e "$LLAMA_STACK_DIR"
     else
-      pip install --no-cache-dir llama-stack
+      uv pip install --no-cache-dir llama-stack
     fi
 
     if [ -n "$LLAMA_MODELS_DIR" ]; then
@@ -85,18 +86,18 @@ run() {
       fi
 
       printf "Installing from LLAMA_MODELS_DIR: $LLAMA_MODELS_DIR\n"
-      pip uninstall -y llama-models
+      uv pip uninstall -y llama-models
-      pip install --no-cache-dir -e "$LLAMA_MODELS_DIR"
+      uv pip install --no-cache-dir -e "$LLAMA_MODELS_DIR"
     fi
 
     # Install pip dependencies
     printf "Installing pip dependencies\n"
-    pip install $pip_dependencies
+    uv pip install $pip_dependencies
     if [ -n "$special_pip_deps" ]; then
       IFS='#' read -ra parts <<<"$special_pip_deps"
       for part in "${parts[@]}"; do
         echo "$part"
-        pip install $part
+        uv pip install $part
       done
     fi
   fi

llama_stack/distribution/client.py

@@ -68,9 +68,9 @@ def create_api_client_class(protocol) -> Type:
                 return_type = None
             else:
                 return_type = extract_non_async_iterator_type(sig.return_annotation)
-                assert (
+                assert return_type, (
-                    return_type
+                    f"Could not extract return type for {sig.return_annotation}"
-                ), f"Could not extract return type for {sig.return_annotation}"
+                )
 
             async with httpx.AsyncClient() as client:
                 params = self.httpx_request_params(method_name, *args, **kwargs)
@@ -87,9 +87,9 @@ def create_api_client_class(protocol) -> Type:
             webmethod, sig = self.routes[method_name]
 
             return_type = extract_async_iterator_type(sig.return_annotation)
-            assert (
+            assert return_type, (
-                return_type
+                f"Could not extract return type for {sig.return_annotation}"
-            ), f"Could not extract return type for {sig.return_annotation}"
+            )
 
             async with httpx.AsyncClient() as client:
                 params = self.httpx_request_params(method_name, *args, **kwargs)

llama_stack/scripts/install_packages.sh

@@ -11,5 +11,5 @@ VERSION="$1"
 set -euo pipefail
 set -x
 
-pip install -U --extra-index-url https://test.pypi.org/simple \
+uv pip install -U --extra-index-url https://test.pypi.org/simple \
   llama-stack==$VERSION llama-models==$VERSION llama-stack-client==$VERSION

llama_stack/templates/meta-reference-gpu/doc_template.md

@@ -71,7 +71,7 @@ docker run \
 
 ### Via Conda
 
-Make sure you have done `pip install llama-stack` and have the Llama Stack CLI available.
+Make sure you have done `uv pip install llama-stack` and have the Llama Stack CLI available.
 
 ```bash
 llama stack build --template {{ name }} --image-type conda

llama_stack/templates/meta-reference-quantized-gpu/doc_template.md

@@ -73,7 +73,7 @@ docker run \
 
 ### Via Conda
 
-Make sure you have done `pip install llama-stack` and have the Llama Stack CLI available.
+Make sure you have done `uv pip install llama-stack` and have the Llama Stack CLI available.
 
 ```bash
 llama stack build --template {{ name }} --image-type conda

llama_stack/templates/ollama/doc_template.md

@@ -93,7 +93,7 @@ docker run \
 
 ### Via Conda
 
-Make sure you have done `pip install llama-stack` and have the Llama Stack CLI available.
+Make sure you have done `uv pip install llama-stack` and have the Llama Stack CLI available.
 
 ```bash
 export LLAMA_STACK_PORT=5001

llama_stack/templates/remote-vllm/doc_template.md

@@ -119,7 +119,7 @@ docker run \
 
 ### Via Conda
 
-Make sure you have done `pip install llama-stack` and have the Llama Stack CLI available.
+Make sure you have done `uv pip install llama-stack` and have the Llama Stack CLI available.
 
 ```bash
 export INFERENCE_PORT=8000

llama_stack/templates/template.py

@@ -58,9 +58,9 @@ class RunConfigSettings(BaseModel):
                     )
 
                 config_class = provider_registry[api][provider_type].config_class
-                assert (
+                assert config_class is not None, (
-                    config_class is not None
+                    f"No config class for provider type: {provider_type} for API: {api_str}"
-                ), f"No config class for provider type: {provider_type} for API: {api_str}"
+                )
 
                 config_class = instantiate_class_type(config_class)
                 if hasattr(config_class, "sample_run_config"):

llama_stack/templates/tgi/doc_template.md

@@ -111,7 +111,7 @@ docker run \
 
 ### Via Conda
 
-Make sure you have done `pip install llama-stack` and have the Llama Stack CLI available.
+Make sure you have done `uv pip install llama-stack` and have the Llama Stack CLI available.
 
 ```bash
 llama stack build --template {{ name }} --image-type conda

pyproject.toml

@@ -59,3 +59,8 @@ install-wheel-from-presigned = "llama_stack.cli.scripts.run:install_wheel_from_p
 
 [tool.setuptools]
 packages = ["llama_stack"]
+
+[[tool.uv.index]]
+name = "pytorch-cpu"
+url = "https://download.pytorch.org/whl/cpu"
+explicit = true