Skip to content

astropy.units.core.UnitConversionError: 's / m' and 's / m' are not convertible #587

Closed
#596
Closed
astropy.units.core.UnitConversionError: 's / m' and 's / m' are not convertible#587
#596
Labels
bugIssues describing unexpected behavior or defects. Remember: a bug is a sign of a missing test!priority: very highIssues & PRs of the highest urgency (e.g., broken CI and critical bug fixes)upstream fix requiredIssues & PRs that are blocked due to a problem with a dependency

Description

@StanczakDominik
StanczakDominik
opened on Dec 11, 2018

https://travis-ci.org/PlasmaPy/PlasmaPy/jobs/466396211

After wrestling with system packages (conda, pip and pacman, argh!), I've managed to figure out that this is localized to astropy 3.1. Also, running python setup.py test -d and getting into plasmapy.physics.distribution.Maxwellian_1D:

(Pdb) distFunc
<Quantity 5.91632969e-07 s / m>
(Pdb) p u.s / u.m
Unit("s / m")
(Pdb) p distFunc
<Quantity 5.91632969e-07 s / m>
(Pdb) p distFunc.unit
Unit("s / m")
(Pdb) p distFunc.to(u.s / u.m)
*** astropy.units.core.UnitConversionError: 's / m' and 's / m' are not convertible

Activity

  1. StanczakDominik
    added
    bugIssues describing unexpected behavior or defects. Remember: a bug is a sign of a missing test!
    priority: very highIssues & PRs of the highest urgency (e.g., broken CI and critical bug fixes)
    upstream fix requiredIssues & PRs that are blocked due to a problem with a dependency
    on Dec 11, 2018

  2. StanczakDominik commented on Dec 12, 2018

    @StanczakDominik
    StanczakDominik
    on Dec 12, 2018
    MemberAuthor

    Looks like this is going to be fixed for Astropy 3.1.1 in astropy/astropy#8263 which leaves us to figure out how to block 3.1 in requirements. Basically we want to say Astropy >= 3.0.2 & Astropy != 3.1.0"

  3. manasbedmutha98 commented on Dec 12, 2018

    @manasbedmutha98
    manasbedmutha98
    on Dec 12, 2018
    Contributor

    I could find only this as the package installation file having astropy. Should this be made to Astropy >= 3.0.2 & Astropy != 3.1.0" or I am missing something?

    PlasmaPy/requirements/requirements.txt

    Line 3 in 3e919ff

    astropy (>= 2.0)

  4. ritiek commented on Dec 13, 2018

    @ritiek
    ritiek
    on Dec 13, 2018
    Contributor

    @manasbedmutha98 Yep, modifying that line to a format which requirements.txt accepts would be the way to go. I think replacing it with something like astropy (>= 3.0.2, != 3.1.0) should work.

    Perhaps it would be also be a good idea to add a comment in requirements.txt referencing this issue for why we do not want astropy v3.1.0.

  5. ritiek commented on Dec 13, 2018

    @ritiek
    ritiek
    on Dec 13, 2018
    Contributor

    Wait. This doesn't seems to be working. If I pip install "astropy==3.1.0" and then modify the astropy line in requirements.txt to astropy (>= 3.0.2, != 3.1.0). It doesn't seem to install a different version (anything other than v3.1.0) when running python setup.py install.

    This seems a problem specific to PlasmaPy as I am able to create a dummy test.txt with a line containing astropy (>= 3.0.2, != 3.1.0) and then when installing it with pip install -r test.txt it indeed downgrades to v3.0.5 from v3.1.0.

  6. StanczakDominik
    closed this as completedin #596on Jan 4, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugIssues describing unexpected behavior or defects. Remember: a bug is a sign of a missing test!priority: very highIssues & PRs of the highest urgency (e.g., broken CI and critical bug fixes)upstream fix requiredIssues & PRs that are blocked due to a problem with a dependency

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions