Importing scPRINT v2.0.0 fails with "No such file or directory"

[lazappi]

Describe the bug

Importing scprint fails with a "No such file or directory" error

Python 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import scprint
→ connected lamindb: anonymous/main
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.10/dist-packages/scprint/__init__.py", line 3, in <module>
    from .model.model import scPrint
  File "/usr/local/lib/python3.10/dist-packages/scprint/model/__init__.py", line 1, in <module>
    from .model import *
  File "/usr/local/lib/python3.10/dist-packages/scprint/model/model.py", line 24, in <module>
    from . import decoders, encoders, fsq, loss, utils
  File "/usr/local/lib/python3.10/dist-packages/scprint/model/utils.py", line 404, in <module>
    class WeightedMasker:
  File "/usr/local/lib/python3.10/dist-packages/scprint/model/utils.py", line 408, in WeightedMasker
    TFs: list[str] = utils.fileToList(FILEDIR + "/../../data/main/TFs.txt"),
  File "/usr/local/lib/python3.10/dist-packages/scprint/utils/utils.py", line 53, in fileToList
    with open(filename) as f:
FileNotFoundError: [Errno 2] No such file or directory: '/usr/local/lib/python3.10/dist-packages/scprint/model/../../data/main/TFs.txt'

It seems like the data/ directory containing TFs.txt isn't installed with the package:

root@41d356636499:/workspace# ls /usr/local/lib/python3.10/dist-packages/scprint/
VERSION  __init__.py  __main__.py  __pycache__  base.py  cli.py  model  tasks  tokenizers  trainer  utils

Desktop (please complete the following information):

• OS: Linux (Docker)
• Version: 2.0.0

[jkobject]

Hello, Wow sorry, I just saw this issue, I don't know why I don't get notified. Can you install it from git directly? git pull ... as it is in the github. I think I need to add it to pypi manually otherwise

[jkobject]

I will make a new release with a solution for this.

[jkobject]

it should be solved now with version 2.1.0

[lazappi]

I just tried this with v2.1.0 and get the same error. Again, looking at the installation directory there is no data/ folder, only what is in scprint/. Is there something different you need to do to install the data files?

This is what I did to install it in a fresh conda environment:

mamba create -n scprint python==3.10
mamba activate scprint
pip install scprint
pip list | grep scprint
# scprint                   2.1.0

And then in Python:

Python 3.10.0 | packaged by conda-forge | (default, Nov 20 2021, 02:24:10) [GCC 9.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import scprint
✓ wrote new records from public sources.yaml to /home/luke/.lamin/bionty/versions/sources_local.yaml!

if you see this message repeatedly, run: import bionty; bionty.base.reset_sources()
! schema module 'wetlab' is not installed → no access to its labels & registries (resolve via `pip install wetlab`)
→ connected lamindb: laminlabs/lamindata
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint/__init__.py", line 6, in <module>
    from .model.model import scPrint
  File "/home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint/model/__init__.py", line 1, in <module>
    from .model import *
  File "/home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint/model/model.py", line 24, in <module>
    from . import decoders, encoders, fsq, loss, utils
  File "/home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint/model/utils.py", line 404, in <module>
    class WeightedMasker:
  File "/home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint/model/utils.py", line 408, in WeightedMasker
    TFs: list[str] = utils.fileToList(FILEDIR + "/../../data/main/TFs.txt"),
  File "/home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint/utils/utils.py", line 53, in fileToList
    with open(filename) as f:
FileNotFoundError: [Errno 2] No such file or directory: '/home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint/model/../../data/main/TFs.txt'

This is what is in the installation directory:

$ ls /home/luke/miniforge3/envs/scprint/lib/python3.10/site-packages/scprint
__init__.py  __main__.py  __pycache__  base.py  cli.py  model  tasks  tokenizers  trainer  utils  VERSION

[jkobject]

Sh*t, sorry. I am back at it then. Please use the git (git clone .. or pip install git+....) version in the meantime.

[jkobject]

Hello, were you able to make it work with git clone / pip install git+?

[jkobject]

this has been debugged and tested on 2.1.1

[jkobject]

@lazappi let me know if there is anything else I can do to make scprint work on the open problems benchmark (I was wondering what the "foundation model" benchmark was) and if it could be added to the denoising task as well.

[lazappi]

I haven't had a chance to check with 2.1.1 or from GitHub yet. If you wanted to open a PR updating the version on one of the tasks (maybe batch integration) that would be helpful.

The foundation model task is still a work in progress but it will combine results from other tasks to give an overview of foundation models. I think we added scPRINT to denoising as well but we haven't done a run of that benchmark yet to update the results.