Install Codeflash Github action

.github/workflows/codeflash.yml

@@ -0,0 +1,36 @@
+name: Codeflash Optimization
+on:
+  pull_request:
+    paths:
+      - 'pydantic_ai_slim/pydantic_ai/**'
+  workflow_dispatch:
+concurrency:
+  # Any new push to the PR will cancel the previous run, so that only the latest code is optimized
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+jobs:
+  optimize:
+    name: Optimize new Python code
+    # Don't run codeflash on codeflash-ai[bot] commits, prevent duplicate optimizations
+    if: ${{ github.actor != 'codeflash-ai[bot]' }}
+    runs-on: ubuntu-latest
+    env:
+      CODEFLASH_API_KEY: ${{ secrets.CODEFLASH_API_KEY }}
+    steps:
+      - name: 🛎️ Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - name: 🐍 Setup UV
+        uses: astral-sh/setup-uv@v6
+        with:
+          enable-cache: true
+      - uses: denoland/setup-deno@v2
+        with:
+          deno-version: v2.x
+      - name: 📦 Install Dependencies
+        run: |
+          uv sync --all-extras --all-packages --group lint --group docs
+          uv pip install -U codeflash
+      - name: ⚡️Codeflash Optimization
+        run: uv run codeflash --benchmark

pydantic_ai_slim/pyproject.toml

@@ -103,6 +103,7 @@ dev = [
     "boto3-stubs[bedrock-runtime]",
     "strict-no-cover>=0.1.1",
     "pytest-xdist>=3.6.1",
+    "codeflash>=0.15.6",
 ]
 
 [tool.hatch.metadata]

pyproject.toml

@@ -83,7 +83,11 @@ members = [
 
 [dependency-groups]
 # dev dependencies are defined in `pydantic-ai-slim/pyproject.toml` to allow for minimal testing
-lint = ["mypy>=1.11.2", "pyright>=1.1.390", "ruff>=0.6.9"]
+lint = [
+    "mypy>=1.11.2",
+    "pyright>=1.1.390",
+    "ruff>=0.6.9",
+]
 docs = [
     "pydantic-ai[a2a]",
     "black>=24.10.0",
@@ -287,3 +291,11 @@ check-hidden = true
 # Ignore "formatting" like **L**anguage
 ignore-regex = '\*\*[A-Z]\*\*[a-z]+\b'
 ignore-words-list = 'asend,aci'
+
+[tool.codeflash]
+module-root = "pydantic_ai_slim/pydantic_ai"
+tests-root = "tests"
+test-framework = "pytest"
+ignore-paths = []
+formatter-cmds = ["ruff check --exit-zero --fix $file", "ruff format $file"]
+benchmarks-root = "tests"

tests/test_examples.py

@@ -118,6 +118,7 @@ def test_docs_examples(  # noqa: C901
     allow_model_requests: None,
     env: TestEnv,
     tmp_path_cwd: Path,
+    benchmark: Any,
 ):
     mocker.patch('pydantic_ai.agent.models.infer_model', side_effect=mock_infer_model)
     mocker.patch('pydantic_ai._utils.group_by_temporal', side_effect=mock_group_by_temporal)
@@ -222,9 +223,19 @@ def print(self, *args: Any, **kwargs: Any) -> None:
         test_globals: dict[str, str] = {'__name__': dunder_name}
 
         if eval_example.update_examples:  # pragma: lax no cover
-            eval_example.run_print_update(example, call=call_name, module_globals=test_globals)
+            benchmark(
+                eval_example.run_print_update,
+                example,
+                call=call_name,
+                module_globals=test_globals,
+            )
         else:
-            eval_example.run_print_check(example, call=call_name, module_globals=test_globals)
+            benchmark(
+                eval_example.run_print_check,
+                example,
+                call=call_name,
+                module_globals=test_globals,
+            )
 
 
 def print_callback(s: str) -> str: