fix: use wheel files for installation in trtllm build

[nv-anants]

Overview:

Use wheel files to install dynamo and nixl packages.

Using --find-links brings in newer version of nixl from pypi repo. from pip docs - https://pip.pypa.io/en/latest/cli/pip_install/#finding-packages

pip looks for packages in a number of places: on PyPI (or the index given as --index-url, if not disabled via --no-index), in the local filesystem, and in any additional repositories specified via --find-links or --extra-index-url. There is no priority in the locations that are searched. Rather they are all checked, and the “best” match for the requirements (in terms of version number - see the specification for details) is selected.

Tested in build,

#102 32.55 Installed 11 packages in 3ms
#102 32.55  + ai-dynamo==0.4.0 (from file:///workspace/wheelhouse/ai_dynamo-0.4.0-py3-none-any.whl)
#102 32.55  + ai-dynamo-runtime==0.4.0 (from file:///workspace/wheelhouse/ai_dynamo_runtime-0.4.0-cp312-cp312-manylinux_2_28_x86_64.whl)
#102 32.55  + circus==0.19.0
#102 32.55  - click==8.2.1
#102 32.55  + click==8.1.8
#102 32.55  + nixl==0.4.1 (from file:///workspace/wheelhouse/nixl-0.4.1-cp312-cp312-linux_x86_64.whl)
#102 32.55  - pydantic==2.11.7
#102 32.55  + pydantic==2.10.6

[coderabbitai]

Walkthrough

The Dockerfile for the TensorRT LLM container was updated to change how Python packages are installed. Instead of using package names with the --find-links option, the installation now specifies exact wheel files in the local wheelhouse directory for ai-dynamo and nixl.

Changes

Cohort / File(s)	Change Summary
Dockerfile Package Installation container/Dockerfile.tensorrt_llm	Modified the Python package installation step to directly reference specific wheel files for ai-dynamo and nixl using explicit file patterns, replacing the previous use of package names and --find-links.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~7 minutes

Possibly related PRs

• chore: fix install #2191: Adjusts pip install commands in Dockerfiles for ai-dynamo, focusing on installation syntax and source, similar to this PR.
• feat: add runtime image for trtllm container build #1796: Introduced the runtime stage in the Dockerfile and initial wheel installation, directly related to the changes in this PR.
• build: fixes to enable vLLM slim runtime image #1058: Alters Dockerfile logic to install Python wheels explicitly for NIXL, conceptually similar to the explicit wheel installation approach in this PR.

Poem

In Docker’s den, the wheels now roll,
With patterns clear, they reach their goal.
No more searching, links to find—
Just named files, precisely lined.
Rabbits cheer, the build runs fast,
Explicit wheels install at last!
🐇✨

Note

🔌 MCP (Model Context Protocol) integration is now available in Early Access!

Pro users can now connect to remote MCP servers under the Integrations page to get reviews and chat conversations that understand additional development context.

---

📜 Recent review details

Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between f7e468c and 2019709.

📒 Files selected for processing (1)

• container/Dockerfile.tensorrt_llm (1 hunks)

🧰 Additional context used

🧠 Learnings (3)

📓 Common learnings

Learnt from: dmitry-tokarev-nv
PR: ai-dynamo/dynamo#2300
File: pyproject.toml:64-66
Timestamp: 2025-08-05T22:51:59.230Z
Learning: The ai-dynamo/dynamo project does not ship ARM64 wheels, so platform markers to restrict dependencies to x86_64 are not needed in pyproject.toml dependencies.

Learnt from: ptarasiewiczNV
PR: ai-dynamo/dynamo#2027
File: container/deps/vllm/install_vllm.sh:0-0
Timestamp: 2025-07-22T10:22:28.972Z
Learning: The `--torch-backend=auto` flag works with vLLM installations via uv pip install, even though it's not a standard pip option. This flag is processed by vLLM's build system during installation to automatically match PyTorch distribution with container CUDA versions.

Learnt from: zaristei
PR: ai-dynamo/dynamo#2020
File: container/deps/vllm/install_vllm.sh:115-118
Timestamp: 2025-07-21T00:10:56.947Z
Learning: Graceful fallback for PyTorch wheel installation is broken on ARM architecture, so immediate exit on pinned version failure is preferred over fallback mechanisms in container/deps/vllm/install_vllm.sh for ARM64.

📚 Learning: 2025-08-05T22:51:59.230Z

Learnt from: dmitry-tokarev-nv
PR: ai-dynamo/dynamo#2300
File: pyproject.toml:64-66
Timestamp: 2025-08-05T22:51:59.230Z
Learning: The ai-dynamo/dynamo project does not ship ARM64 wheels, so platform markers to restrict dependencies to x86_64 are not needed in pyproject.toml dependencies.

Applied to files:

• container/Dockerfile.tensorrt_llm

📚 Learning: 2025-07-21T00:10:56.947Z

Learnt from: zaristei
PR: ai-dynamo/dynamo#2020
File: container/deps/vllm/install_vllm.sh:115-118
Timestamp: 2025-07-21T00:10:56.947Z
Learning: Graceful fallback for PyTorch wheel installation is broken on ARM architecture, so immediate exit on pinned version failure is preferred over fallback mechanisms in container/deps/vllm/install_vllm.sh for ARM64.

Applied to files:

• container/Dockerfile.tensorrt_llm

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)

• GitHub Check: Build and Test - dynamo

🔇 Additional comments (1)

container/Dockerfile.tensorrt_llm (1)

489-489: Ensure offline-only, deterministic pip install and explicit failure on missing wheels

Update at container/Dockerfile.tensorrt_llm:489:

-    uv pip install /workspace/wheelhouse/ai_dynamo_runtime*cp312*.whl /workspace/wheelhouse/ai_dynamo*any.whl /workspace/wheelhouse/nixl*.whl
+    # Install only from local wheelhouse and fail if any wheel is missing
+    uv pip install --no-index --find-links /workspace/wheelhouse \
+        /workspace/wheelhouse/ai_dynamo_runtime*cp312*.whl \
+        /workspace/wheelhouse/ai_dynamo*any.whl \
+        /workspace/wheelhouse/nixl*.whl

Follow-up checks:

• Verify that the actual NIXL wheel filename in /workspace/wheelhouse matches the nixl*.whl glob.
• Confirm the ARM64 build produces a cp312 wheel for ai_dynamo_runtime in the wheelhouse (so it errors out rather than falling back to PyPI).

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

🪧 Tips

Chat

There are 3 ways to chat with CodeRabbit:

• Review comments: Directly reply to a review comment made by CodeRabbit. Example:
  • I pushed a fix in commit <commit_id>, please review it.
  • Explain this complex logic.
  • Open a follow-up GitHub issue for this discussion.
• Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
  • @coderabbitai explain this code block.
• PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
  • @coderabbitai gather interesting stats about this repository and render them as a table. Additionally, render a pie chart showing the language distribution in the codebase.
  • @coderabbitai read src/utils.ts and explain its main purpose.
  • @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.

Support

Need help? Create a ticket on our support page for assistance with any issues or questions.

CodeRabbit Commands (Invoked using PR comments)

• @coderabbitai pause to pause the reviews on a PR.
• @coderabbitai resume to resume the paused reviews.
• @coderabbitai review to trigger an incremental review. This is useful when automatic reviews are disabled for the repository.
• @coderabbitai full review to do a full review from scratch and review all the files again.
• @coderabbitai summary to regenerate the summary of the PR.
• @coderabbitai generate sequence diagram to generate a sequence diagram of the changes in this PR.
• @coderabbitai resolve resolve all the CodeRabbit review comments.
• @coderabbitai configuration to show the current CodeRabbit configuration for the repository.
• @coderabbitai help to get help.

Other keywords and placeholders

• Add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.
• Add @coderabbitai summary to generate the high-level summary at a specific location in the PR description.
• Add @coderabbitai anywhere in the PR title to generate the title automatically.

Documentation and Community

• Visit our Documentation for detailed information on how to use CodeRabbit.
• Join our Discord Community to get help, request features, and share feedback.
• Follow us on X/Twitter for updates and announcements.

[nv-tusharma]

We probably need to apply the same changes across all Dockerfiles.

[nv-tusharma]

Actually, is there a reason why we don't have the nixl dependency for tensorrt_llm in the pyproject.toml?
https://github.com/ai-dynamo/dynamo/blob/main/pyproject.toml#L62-L66

[nv-anants]

We probably need to apply the same changes across all Dockerfiles.

Others dockerfiles have different way of install, seems working for now. Long term -- need to standardize this part

Actually, is there a reason why we don't have the nixl dependency for tensorrt_llm in the pyproject.toml?

checking with Tanmay

[dmitry-tokarev-nv]

Do we have to build NIXL from source for every backend?
Are we using some build flags that make it more performant?