search

tox-dev / pipdeptree

A command line utility to display dependency tree of the installed Python packages
https://pypi.python.org/pypi/pipdeptree
MIT License
2.79k stars 150 forks source link

version detection problem when conflicting dependencies #112

Closed EricDeveaud closed 5 months ago

EricDeveaud commented 5 years ago

hello, just to let you know that when Possibly conflicting dependencies found the version detection breaks see:

tars-900:scratch/gensoft_tests > pipdeptree --user    
Warning!!! Possibly conflicting dependencies found:
* h5py==2.8.0
 - numpy [required: >=1.7, installed: ?]
* biom-format==2.1.5
 - numpy [required: >=1.3.0, installed: ?]
 - scipy [required: >=0.13.0, installed: ?]
------------------------------------------------------------------------
biom-format==2.1.5
  - click [required: Any, installed: 7.0]
  - future [required: >=0.14.3, installed: 0.16.0]
  - numpy [required: >=1.3.0, installed: ?]
  - pyqi [required: Any, installed: 0.3.2]
  - scipy [required: >=0.13.0, installed: ?]
h5py==2.8.0
  - numpy [required: >=1.7, installed: ?]
  - six [required: Any, installed: 1.12.0]

in this scenario nor installed numpy neither scipy version are detected.

NB: using version 0.13.1

regards

Eric

naiquevin commented 5 years ago

Thanks for reporting this and sorry about the delay. I will have a look at it soon.

xiacunshun commented 6 months ago

@EricDeveaud Please test with the newest code to see if this problem exists. We have refactored the code and I can not reproduce it.

kemzeb commented 5 months ago

I haven't been able to reproduce this when using pipdeptree==0.13.1, biom-format==2.1.5, and pip==18.1 (using this pip version since pidpeptree would crash with modern pip and it was released close to when the user reported this issue):

$ pipdeptree --version
0.13.1
$ pipdeptree --user
biom-format==2.1.5
  - click [required: Any, installed: 8.1.7]
  - future [required: >=0.14.3, installed: 1.0.0]
  - numpy [required: >=1.3.0, installed: 1.24.4]
  - scipy [required: >=0.13.0, installed: 1.10.1]
    - numpy [required: >=1.19.5,<1.27.0, installed: 1.24.4]
. . .

Comparing this with the reporter's output, to me it looks like either numpy and scipy aren't installed or an older pip was being used that had issues when looking for installed packages. Going to close this since I can't reproduce and no other users have had similar issues 👍