feat: restructure source code for python packaging (#3201)

Signed-off-by: Anant Sharma <anants@nvidia.com>

.devcontainer/post-create.sh

@@ -77,9 +77,10 @@ cargo build --locked --profile dev --features dynamo-llm/block-manager
 #     echo "maturin is already installed"
 # fi
 
+# install ai-dynamo-runtime
 (cd $DYNAMO_HOME/lib/bindings/python && retry maturin develop)
 
-# installs overall python packages, grabs binaries from .build/target/debug
+# install ai-dynamo
 cd $DYNAMO_HOME && retry env DYNAMO_BIN_PATH=$CARGO_TARGET_DIR/debug uv pip install -e .
 
 { set +x; } 2>/dev/null
@@ -96,23 +97,6 @@ if [ -n "${PYTHONPATH:-}" ]; then
         echo "# PYTHONPATH modified from /workspace to use DYNAMO_HOME" >> ~/.bashrc
         show_and_run echo "export PYTHONPATH=\"$PYTHONPATH\"" >> ~/.bashrc
     fi
-else
-    # PYTHONPATH not set, extract from README.md or use backup
-    PYTHONPATH_LINE=$(grep "^export PYTHONPATH=" $DYNAMO_HOME/README.md | head -n1)
-    if [ -n "$PYTHONPATH_LINE" ]; then
-        # Remove the ${PYTHONPATH}: prefix if it exists, then replace $(pwd) with the actual path
-        MODIFIED_LINE=$(echo "$PYTHONPATH_LINE" | sed 's/\${PYTHONPATH}://g' | sed "s|\$(pwd)|$DYNAMO_HOME|g")
-        eval "$MODIFIED_LINE"
-        # Also add to .bashrc for persistence (with expanded path)
-        if ! grep -q "export PYTHONPATH=" ~/.bashrc; then
-            # MODIFIED_LINE already has $DYNAMO_HOME expanded to /home/ubuntu/dynamo
-            echo "# PYTHONPATH is derived from the README.md" >> ~/.bashrc
-            show_and_run echo "$MODIFIED_LINE" >> ~/.bashrc
-        fi
-    else
-        # Back-up version. Make sure to sync this with the README.md's PYTHONPATH. This is the version from 2025-08-19
-        show_and_run export PYTHONPATH=$DYNAMO_HOME/components/frontend/src:$DYNAMO_HOME/components/planner/src:$DYNAMO_HOME/components/backends/vllm/src:$DYNAMO_HOME/components/backends/sglang/src:$DYNAMO_HOME/components/backends/trtllm/src:$DYNAMO_HOME/components/backends/llama_cpp/src:$DYNAMO_HOME/components/backends/mocker/src
-    fi
 fi
 
 if ! grep -q "export GPG_TTY=" ~/.bashrc; then

.github/filters.yaml

@@ -29,18 +29,21 @@ vllm: &vllm
   - 'container/deps/requirements.vllm.txt'
   - 'container/deps/vllm/**'
   - 'components/backends/vllm/**'
+  - 'components/src/dynamo/vllm/**'
   - 'tests/serve/test_vllm.py'
 
 sglang: &sglang
   - 'container/Dockerfile.sglang'
   - 'container/Dockerfile.sglang-wideep'
   - 'components/backends/sglang/**'
+  - 'components/src/dynamo/sglang/**'
   - 'container/build.sh'
   - 'tests/serve/test_sglang.py'
 
 trtllm: &trtllm
   - 'container/Dockerfile.trtllm'
   - 'components/backends/trtllm/**'
+  - 'components/src/dynamo/trtllm/**'
   - 'container/build.sh'
   - 'container/build_trtllm_wheel.sh'
   - 'container/deps/**'

README.md

@@ -314,14 +314,9 @@ maturin develop --uv
 
 ```
 cd $PROJECT_ROOT
-uv pip install .
-# For development, use
-export PYTHONPATH="${PYTHONPATH}:$(pwd)/components/frontend/src:$(pwd)/components/planner/src:$(pwd)/components/backends/vllm/src:$(pwd)/components/backends/sglang/src:$(pwd)/components/backends/trtllm/src:$(pwd)/components/backends/llama_cpp/src:$(pwd)/components/backends/mocker/src"
+uv pip install -e .
 ```
 
-> [!Note]
-> Editable (`-e`) does not work because the `dynamo` package is split over multiple directories, one per backend.
-
 You should now be able to run `python -m dynamo.frontend`.
 
 Remember that nats and etcd must be running (see earlier).

components/README.md

@@ -31,7 +31,7 @@ Each engine provides launch scripts for different deployment patterns in their r
 
 ## Core Components
 
-### [Backends](backends/)
+### [Backends](src/dynamo/)
 
 The backends directory contains inference engine integrations and implementations, with a key focus on:
 
@@ -40,7 +40,7 @@ The backends directory contains inference engine integrations and implementation
 - **TensorRT-LLM** - TensorRT-LLM integration with disaggregated serving capabilities
 
 
-### [Frontend](frontend/)
+### [Frontend](src/dynamo/frontend/)
 
 The frontend component provides the HTTP API layer and request processing:
 
@@ -49,7 +49,7 @@ The frontend component provides the HTTP API layer and request processing:
 - **Router** - Routes requests to appropriate workers based on load and KV cache state
 - **Auto-discovery** - Automatically discovers and registers available workers
 
-### [Planner](planner/)
+### [Planner](src/dynamo/planner/)
 
 The planner component monitors system state and dynamically adjusts worker allocation:
 

components/backends/sglang/README.md

@@ -116,11 +116,9 @@ uv pip install maturin
 cd $DYNAMO_HOME/lib/bindings/python
 maturin develop --uv
 cd $DYNAMO_HOME
-uv pip install .
-export PYTHONPATH="${PYTHONPATH}:$(pwd)/components/backends/sglang/src"
-# install target sglang version (you can choose any version)
-# we include the prerelease flag in order to install flashinfer rc versions
-uv pip install --prerelease=allow sglang[all]==0.4.9.post6
+# installs sglang supported version along with dynamo
+# include the prerelease flag to install flashinfer rc versions
+uv pip install --prerelease=allow -e .[sglang]
 ```
 
 </details>

components/backends/trtllm/README.md

@@ -61,8 +61,8 @@ git checkout $(git describe --tags $(git rev-list --tags --max-count=1))
 
 ### Large Scale P/D and WideEP Features
 
-| Feature            | TensorRT-LLM | Notes                                                                 |
-|--------------------|--------------|-----------------------------------------------------------------------|
+| Feature            | TensorRT-LLM | Notes                                                           |
+|--------------------|--------------|-----------------------------------------------------------------|
 | **WideEP**         | ✅           |                                                                 |
 | **DP Rank Routing**| ✅           |                                                                 |
 | **GB200 Support**  | ✅           |                                                                 |

components/backends/llama_cpp/README.md → components/src/dynamo/llama_cpp/README.md

@@ -13,10 +13,10 @@ python -m dynamo.llama_cpp --model-path /data/models/Qwen3-0.6B-Q8_0.gguf [args]
 
 ## Request Migration
 
-You can enable [request migration](../../../docs/architecture/request_migration.md) to handle worker failures gracefully. Use the `--migration-limit` flag to specify how many times a request can be migrated to another worker:
+You can enable [request migration](/docs/architecture/request_migration.md) to handle worker failures gracefully. Use the `--migration-limit` flag to specify how many times a request can be migrated to another worker:
 
 ```bash
 python3 -m dynamo.llama_cpp ... --migration-limit=3
 ```
 
-This allows a request to be migrated up to 3 times before failing. See the [Request Migration Architecture](../../../docs/architecture/request_migration.md) documentation for details on how this works.
+This allows a request to be migrated up to 3 times before failing. See the [Request Migration Architecture](/docs/architecture/request_migration.md) documentation for details on how this works.