Merge remote-tracking branch 'upstream/main' into dependabot/uv/langsmith-0.8.18

Signed-off-by: Sébastien Han <seb@redhat.com>

Author: leseb

.github/CODEOWNERS

@@ -7,4 +7,3 @@
 # OpenAPI SDK generation and publishing
 /client-sdks/openapi/ @ashwinb @leseb @bbrowning
 /.github/workflows/openapi-generator-validation.yml @ashwinb @leseb @bbrowning
-/.github/workflows/publish-openapi-sdk.yml @ashwinb @leseb @bbrowning

.github/actions/setup-vllm-gpu/action.yml

@@ -96,9 +96,9 @@ runs:
 
         echo "=== Installing vLLM with CUDA support ==="
         if [ "$VLLM_VERSION" = "latest" ]; then
-          uv pip install --no-config vllm --torch-backend=auto
+          uv pip install --no-config vllm "prometheus-fastapi-instrumentator==8.0.2" --torch-backend=auto
         else
-          uv pip install --no-config "vllm==$VLLM_VERSION" --torch-backend=auto
+          uv pip install --no-config "vllm==$VLLM_VERSION" "prometheus-fastapi-instrumentator==8.0.2" --torch-backend=auto
         fi
 
         echo "=== Verifying installation ==="
@@ -107,31 +107,6 @@ runs:
         python -c "import torch; print(f'CUDA device count: {torch.cuda.device_count()}')"
         vllm --version
 
-    - name: Patch vLLM metrics route discovery
-      shell: bash
-      run: |
-        source /tmp/vllm-env/bin/activate
-
-        python - <<'PY'
-        from __future__ import annotations
-
-        from pathlib import Path
-        import prometheus_fastapi_instrumentator.routing as routing
-
-        routing_path = Path(routing.__file__)
-        source = routing_path.read_text()
-        old = "route_name = route.path"
-        new = (
-            "route_name = getattr(route, 'path', None)\n"
-            "            if route_name is None:\n"
-            "                continue"
-        )
-        if old not in source:
-            raise SystemExit(f"Expected route.path assignment not found in {routing_path}")
-        routing_path.write_text(source.replace(old, new, 1))
-        print(f"Patched {routing_path}")
-        PY
-
     - name: Resolve model identifier
       shell: bash
       env:

.github/workflows/README.md

@@ -28,7 +28,6 @@ OGX uses GitHub Actions for Continuous Integration (CI). Below is a table detail
 | Pre-commit | [pre-commit.yml](pre-commit.yml) | Run pre-commit checks |
 | Prepare release | [prepare-release.yml](prepare-release.yml) | Prepare release |
 | Test OGX Build | [providers-build.yml](providers-build.yml) | Test ogx build and list-deps |
-| Publish OpenAPI SDK to PyPI | [publish-openapi-sdk.yml](publish-openapi-sdk.yml) | Publish ogx-open-client to PyPI |
 | Build, test, and publish packages | [pypi.yml](pypi.yml) | Build, test, and publish packages |
 | Integration Tests (Record) | [record-integration-tests.yml](record-integration-tests.yml) | Auto-record missing test recordings for PR |
 | vLLM GPU Recording | [record-vllm-gpu-tests.yml](record-vllm-gpu-tests.yml) | GPU recording for gpt-oss:20b (${{ inputs.suite }} suite) |

.github/workflows/build-distributions.yml

@@ -154,7 +154,7 @@ jobs:
           output: 'trivy-container-${{ matrix.distro }}.sarif'
 
       - name: Upload container scan results to GitHub Security
-        uses: github/codeql-action/upload-sarif@7211b7c8077ea37d8641b6271f6a365a22a5fbfa # v3
+        uses: github/codeql-action/upload-sarif@8aad20d150bbac5944a9f9d289da16a4b0d87c1e # v3
         if: always()
         with:
           sarif_file: 'trivy-container-${{ matrix.distro }}.sarif'

.github/workflows/file-processors-tests.yml

@@ -56,7 +56,7 @@ jobs:
         run: |
           uv run --no-sync ogx stack run \
             --providers "file_processors=inline::docling,files=inline::localfs" \
-            --port 8321 &
+            --port 8321 --insecure &
           # Wait for server to be ready
           for i in $(seq 1 60); do
             if curl -sf http://localhost:8321/v1/health > /dev/null 2>&1; then

.github/workflows/integration-auth-tests.yml

@@ -156,7 +156,7 @@ jobs:
 
           # avoid line breaks in the server log, especially because we grep it below.
           export OGX_LOG_WIDTH=200
-          nohup uv run ogx stack run "$run_dir/config.yaml" > server.log 2>&1 &
+          nohup uv run ogx stack run "$run_dir/config.yaml" --insecure > server.log 2>&1 &
 
       - name: Wait for OGX server to be ready
         run: |

.github/workflows/integration-responses-conversations-auth-tests.yml

@@ -299,7 +299,7 @@ jobs:
           # avoid line breaks in the server log, especially because we grep it below.
           export OGX_LOG_WIDTH=200
           # OPENAI_API_KEY is set via env at step level
-          nohup uv run ogx stack run "$run_dir/config.yaml" > server.log 2>&1 &
+          nohup uv run ogx stack run "$run_dir/config.yaml" --insecure > server.log 2>&1 &
 
       - name: Wait for OGX server to be ready
         run: |

.github/workflows/openapi-generator-validation.yml

@@ -187,7 +187,7 @@ jobs:
 
       - name: Generate Python SDK
         working-directory: client-sdks/openapi
-        run: make sdk OPEN=1
+        run: make sdk OPEN=0
 
       - name: Validate generated SDK
         working-directory: client-sdks/openapi
@@ -217,14 +217,14 @@ jobs:
 
       - name: Install generated SDK
         run: |
-          echo "Reinstalling OpenAPI-generated SDK (ogx_open_client)..."
-          uv pip uninstall ogx-open-client || true
+          echo "Reinstalling OpenAPI-generated SDK (ogx_client)..."
+          uv pip uninstall ogx-client || true
 
           # Install SDK using uv pip
           uv pip install -e client-sdks/openapi/sdks/python
 
           echo "Verifying installation..."
-          uv run python -c "import ogx_open_client; print(f'Installed: {ogx_open_client.__name__}')"
+          uv run python -c "import ogx_client; print(f'Installed: {ogx_client.__name__}')"
 
       - name: Setup Ollama (for integration tests)
         if: runner.os == 'Linux'
@@ -242,7 +242,7 @@ jobs:
             || echo "::warning::Some integration tests failed - this may indicate SDK compatibility issues"
 
           # Show which SDK is actually being used
-          uv run python -c "import ogx_open_client; import inspect; print(f'SDK location: {inspect.getfile(ogx_open_client)}')"
+          uv run python -c "import ogx_client; import inspect; print(f'SDK location: {inspect.getfile(ogx_client)}')"
 
       - name: Summary
         if: runner.os == 'Linux'
@@ -255,5 +255,5 @@ jobs:
             echo "✅ Integration tests executed (check logs for results)"
             echo ""
             echo "**Platform**: ${{ matrix.os }}"
-            echo "**Package**: ogx_open_client (OpenAPI-generated)"
+            echo "**Package**: ogx_client (OpenAPI-generated)"
           } >> "$GITHUB_STEP_SUMMARY"

.github/workflows/openresponses-conformance.yml

@@ -90,7 +90,7 @@ jobs:
         run: |
           mkdir -p /tmp/ogx-conformance
           export OGX_LOG_WIDTH=200
-          nohup uv run ogx stack run ci-tests --port "$OGX_PORT" \
+          nohup uv run ogx stack run ci-tests --port "$OGX_PORT" --insecure \
             > /tmp/ogx-conformance/server.log 2>&1 &
           echo "Server PID: $!"