Add DevContainer configuration file

[yvan-sraka]

DevContainer is a great tool for providing a one-click development environment for new contributors. It works seamlessly through the GitHub interface (with GitHub Codespaces) or locally with tools like VSCode or JetBrains’ built-in plugins. Maybe, DevContainer can help achieve a basic level of reproducibility between the CI and local development environments, e.g., the devcontainer CLI could be integrated into GitHub Actions?

[yvan-sraka]

EDIT: actually, the issue was that this version of torch is not compatible with Python 3.13 (so I updated pyproject.toml), but using Python 3.12 allowed it to resolve the correct version :) uv actually provides better error messages and helped me with this one!

I believe this PR provides a simple way to demonstrate that, on a typical x86_64 Linux machine with the correct NVIDIA system dependencies installed, running pip install -e ".[test]" in a fresh virtual environment fails with the following error:

Collecting vllm (from outlines==0.1.15.dev1+g7c7115b)
  Downloading vllm-0.5.1.tar.gz (790 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 790.6/790.6 kB 37.0 MB/s eta 0:00:00
  Installing build dependencies ... error
  error: subprocess-exited-with-error
  
  × pip subprocess to install build dependencies did not run successfully.
  │ exit code: 1
  ╰─> [10 lines of output]
      Collecting cmake>=3.21
        Using cached cmake-3.31.4-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.5 kB)
      Collecting ninja
        Using cached ninja-1.11.1.3-py3-none-manylinux_2_12_x86_64.manylinux2010_x86_64.whl.metadata (5.3 kB)
      Collecting packaging
        Using cached packaging-24.2-py3-none-any.whl.metadata (3.2 kB)
      Collecting setuptools>=49.4.0
        Using cached setuptools-75.8.0-py3-none-any.whl.metadata (6.7 kB)
      ERROR: Could not find a version that satisfies the requirement torch==2.3.0 (from versions: 2.5.0, 2.5.1)
      ERROR: No matching distribution found for torch==2.3.0
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.

[notice] A new release of pip is available: 24.3.1 -> 25.0
[notice] To update, run: pip install --upgrade pip
error: subprocess-exited-with-error

× pip subprocess to install build dependencies did not run successfully.
│ exit code: 1
╰─> See above for output.

note: This error originates from a subprocess, and is likely not a problem with pip.

I have encountered this issue multiple times across different setups. AFAIU, the problem seems comes from wrong constraints in the pyproject.toml. I've tried using different versions of vllm, but I haven't been able to resolve the dependency issues with any of them...

[yvan-sraka]

So, as advised in the Contribute guide, running conda env create -f environment.yml installs the following versions in the container:

Successfully installed Pygments-2.19.1 accelerate-1.3.0 airportsdata-20241001 anyio-4.8.0 beartype-0.15.0 chardet-5.2.0 click-8.1.8 cloudpickle-3.1.1 cmake-3.31.4 coverage-7.6.10 cramjam-2.9.1 diff-cover-9.2.2 diskcache-5.6.3 distro-1.9.0 exllamav2-0.2.7 fastapi-0.115.8 fastparquet-2024.11.0 genson-1.3.0 h11-0.14.0 httpcore-1.0.7 httptools-0.6.4 httpx-0.28.1 interegular-0.3.3 jax-0.5.0 jaxlib-0.5.0 jiter-0.8.2 lark-1.2.2 llama-cpp-python-0.3.7 lm-format-enforcer-0.10.1 markdown-it-py-3.0.0 mdurl-0.1.2 ml_dtypes-0.5.1 mpmath-1.3.0 msgpack-1.1.0 nest_asyncio-1.6.0 networkx-3.4.2 ninja-1.11.1.3 numpy-1.26.4 nvidia-cublas-cu12-12.1.3.1 nvidia-cuda-cupti-cu12-12.1.105 nvidia-cuda-nvrtc-cu12-12.1.105 nvidia-cuda-runtime-cu12-12.1.105 nvidia-cudnn-cu12-8.9.2.26 nvidia-cufft-cu12-11.0.2.54 nvidia-curand-cu12-10.3.2.106 nvidia-cusolver-cu12-11.4.5.107 nvidia-cusparse-cu12-12.1.0.106 nvidia-ml-py-12.570.86 nvidia-nccl-cu12-2.20.5 nvidia-nvjitlink-cu12-12.4.127 nvidia-nvtx-cu12-12.1.105 openai-1.60.2 opt_einsum-3.4.0 outlines-0.1.15.dev1+g32d3700 outlines_core-0.1.26 pillow-11.1.0 prometheus-fastapi-instrumentator-7.0.2 prometheus_client-0.21.1 protobuf-5.29.3 psutil-6.1.1 py-cpuinfo-9.0.0 pycountry-24.6.1 pytest-benchmark-5.1.0 pytest-cov-6.0.0 pytest-mock-3.14.0 python-dotenv-1.0.1 ray-2.41.0 responses-0.25.6 rich-13.9.4 sentencepiece-0.2.0 sniffio-1.3.1 starlette-0.45.3 sympy-1.13.1 tiktoken-0.8.0 torch-2.3.0 torchvision-0.18.0 triton-2.3.0 uvicorn-0.34.0 uvloop-0.21.0 vllm-0.5.1 vllm-flash-attn-2.5.9 watchfiles-1.0.4 websockets-14.2 xformers-0.0.26.post1

done
#
# To activate this environment, use
#
#     $ conda activate dottxt-ai
#
# To deactivate an active environment, use
#
#     $ conda deactivate

However, when I run pytest, it fails... Below are a few log traces, I would really appreciate any help!

• pytest_log_docker.txt
• pytest_log_nix.txt (this is another attempt to create a working test environment Reproducible Python virtual environment using uv and nix #1387)

[rlouf]

Great! Should this be documented somewhere?

[yvan-sraka]

Great! Should this be documented somewhere?

We could add a few lines in the Contribute guide! Do we merge that PR and I open a new issue for the failing tests?

[rlouf]

Yes, let's do that