Update Dockerfile with extensions support

.github/workflows/publish-docker.yml

@@ -0,0 +1,51 @@
+name: Docker Build + Publish
+
+on:
+  # For now, just manually trigger
+  # push:
+  #   branches:
+  #     - main
+  # pull_request:
+  #   branches:
+  #     - main
+  workflow_dispatch:
+
+jobs:
+    build-docker-image:
+
+        runs-on: aws-avx2-192G-4-a10g-96G
+        timeout-minutes: 240
+
+        steps:
+
+          - name: Set up Docker Buildx
+            id: buildx
+            uses: docker/setup-buildx-action@v3
+
+          - name: Login to Github Packages
+            uses: docker/login-action@v3
+            with:
+              registry: ghcr.io
+              username: ${{ github.actor }}
+              password: ${{ secrets.GITHUB_TOKEN }}
+
+          - name: Checkout code
+            uses: actions/checkout@v3
+            with:
+              fetch-depth: 1
+              submodules: recursive
+
+          - name: Get version tag
+            id: extract_tag
+            run: echo "tag=$(date +%Y%m%d)" >> $GITHUB_OUTPUT
+
+          - name: Current Version Name
+            run: echo ${{ steps.extract_tag.outputs.tag }}
+
+          - name: nm-vllm latest
+            uses: docker/build-push-action@v5
+            with:
+              context: .
+              target: vllm-openai
+              push: true
+              tags: ghcr.io/neuralmagic/nm-vllm-openai:${{ steps.extract_tag.outputs.tag }},ghcr.io/neuralmagic/nm-vllm-openai:latest

Dockerfile

@@ -108,6 +108,10 @@ COPY requirements.txt requirements.txt
 RUN --mount=type=cache,target=/root/.cache/pip \
     pip install -r requirements.txt
 
+# UPSTREAM SYNC: Install sparsity extras
+RUN --mount=type=cache,target=/root/.cache/pip \
+    pip install nm-magic-wand
+
 # Install flash attention (from pre-built wheel)
 RUN --mount=type=bind,from=flash-attn-builder,src=/usr/src/flash-attention-v2,target=/usr/src/flash-attention-v2 \
     pip install /usr/src/flash-attention-v2/*.whl --no-cache-dir