Bioconda installation missing a lot of packages

[michoug]

Hi,

I installed your software (v0.4.1) as described

conda create -n rosella -c bioconda rosella
conda activate rosella

When running this command:

rosella bin -i Bam/Test_metabat.txt -r Test_good.fa --output-directory rosella_out/ --threads 28

I got this error for the module threadpoolctl

[2022-04-27T06:47:39Z ERROR bird_tool_utils::command] Error when running flight process. Exitstatus was : ExitStatus(unix_wait_status(256))
[2022-04-27T06:47:39Z ERROR bird_tool_utils::command] The STDERR was: "04/27/2022 08:47:39 AM INFO: Time - 08:47:39 27-04-2022\n04/27/2022 08:47:39 AM INFO: Command - /work/sber/Software/miniconda3/envs/rosella/bin/flight bin --assembly Test_good.fa --input Bam/Test_metabat.txt --kmer_frequencies rosella_out//rosella_kmer_table.tsv --min_contig_size 1500 --min_bin_size 200000 --n_neighbors 200 --output_directory rosella_out// --cores 28\nTraceback (most recent call last):\n  File \"/work/sber/Software/miniconda3/envs/rosella/bin/flight\", line 10, in <module>\n    sys.exit(main())\n  File \"/work/sber/Software/miniconda3/envs/rosella/lib/python3.10/site-packages/flight/flight.py\", line 449, in main\n    args.func(args)\n  File \"/work/sber/Software/miniconda3/envs/rosella/lib/python3.10/site-packages/flight/flight.py\", line 563, in bin\n    rosella = rosella_engine_constructor(args)\n  File \"/work/sber/Software/miniconda3/envs/rosella/lib/python3.10/site-packages/flight/flight.py\", line 534, in rosella_engine_constructor\n    from flight.rosella.rosella import Rosella\n  File \"/work/sber/Software/miniconda3/envs/rosella/lib/python3.10/site-packages/flight/rosella/rosella.py\", line 33, in <module>\n    import threadpoolctl\nModuleNotFoundError: No module named 'threadpoolctl'\n"
thread 'main' panicked at 'Failed to grab stdout from failed flight process', /home/conda/.cargo/registry/src/github.com-1ecc6299db9ec823/bird_tool_utils-0.3.0/src/command.rs:27:14
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

I got the same error for

numba
matplotlib
seaborn
umap
scikit-bio

I had to re-install

pip install flight-genome

which resolved a lot of the issues by installing missing dependencies but not all, the rest of the packages had to be installed with conda
Best
Greg

[rhysnewell]

Hi Greg,

The conda install for Rosella is relatively okay from my understanding, it's the conda & pip versions of flight that are a bit busted. Due to some difficulties with segfaults with pynndescent we've had to resort to only using the GitHub version of flight. I had forgotten to update the documentation to reflect this.

What you need to do is clone this repo: https://github.com/rhysnewell/flight/tree/master
and use pip install . inside the cloned repo. This will install all of the correct dependencies for you. Once you've done that then it should run as expected.

For the moment, until the pynndescent issues are resolved I recommend just letting Aviary handle the installation of Rosella & flight. You can use them without Aviary, but as you've found the installation can be a tad annoying at the moment. Aviary can be found here if you'd prefer to use it: https://github.com/rhysnewell/aviary

Cheers,
Rhys

[rhysnewell]

I've updated the readme page for rosella to better reflect the correct installation procedure, apologies for any confusion