!pip install orion-ml no longer works in Colab

[AungDev]

Problem

Running !pip install orion-ml in Google Colab fails to install the package as expected. This appears to be a recent issue—previously, the package could be installed without errors.

Observed behavior

• The installation command returns errors or fails to resolve dependencies.
• Previous Colab notebooks that depended on orion-ml now fail at the installation step.

Expected behavior

• !pip install orion-ml should successfully install the package in a Colab notebook environment.

Steps to reproduce

1. Open a new Google Colab notebook.
2. Run the following cell:

   !pip install orion-ml

3. The error messages:

Collecting orion-ml
  Downloading orion_ml-0.1.0-py2.py3-none-any.whl.metadata (15 kB)
Collecting baytune<0.3,>=0.2.3 (from orion-ml)
  Downloading baytune-0.2.5-py2.py3-none-any.whl.metadata (10 kB)
Collecting mlblocks<0.4,>=0.3.0 (from orion-ml)
  Downloading mlblocks-0.3.4-py2.py3-none-any.whl.metadata (16 kB)
Collecting mlprimitives<0.3,>=0.2.2 (from orion-ml)
  Downloading mlprimitives-0.2.4-py2.py3-none-any.whl.metadata (25 kB)
Collecting mongoengine<0.17,>=0.16.3 (from orion-ml)
  Downloading mongoengine-0.16.3.tar.gz (149 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 149.0/149.0 kB 10.7 MB/s eta 0:00:00
  error: subprocess-exited-with-error
  
  × python setup.py egg_info 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.
  Preparing metadata (setup.py) ... error
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

note: This is an issue with the package mentioned above, not pip.
hint: See above for details.

[sarahmish]

Hi @AungDev, currently Orion supports python 3.9, 3.10, and 3.11. However, colab has upgraded to python > 3.11. We are working on upgrading our python versions but in the meantime, maybe you can search for options to downgrade colab's python version.

[AungDev]

Hi @sarahmish , Downgrading Colab's Python version to 3.11 using condacolab, works for orion-ml but unfortunately causes conflicts with other libraries (e.g., protobuf).

A temporary workaround I found is to downgrade the entire Colab runtime:
Runtime > Change runtime type > Runtime version > select 2025.07

With this older runtime, installing and using orion-ml works as expected for now. However, please note this is only a short-term solution—once Google phases out the 2025.07 runtime, this workaround will no longer be available.