starry: importing starry but get an error on attribute 'geweke'

>>> import starry
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/starry/__init__.py", line 17, in <module>
    from . import kepler, linalg, maps
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/starry/kepler.py", line 4, in <module>
    from .maps import MapBase, RVBase, ReflectedBase
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/starry/maps.py", line 4, in <module>
    from ._core import OpsYlm, OpsLD, OpsReflected, OpsRV, math
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/starry/_core/__init__.py", line 2, in <module>
    from . import core, ops, utils, math
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/starry/_core/core.py", line 31, in <module>
    import exoplanet
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/exoplanet/__init__.py", line 12, in <module>
    from . import distributions, interp, orbits
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/exoplanet/distributions/__init__.py", line 16, in <module>
    from . import eccentricity
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/exoplanet/distributions/eccentricity.py", line 7, in <module>
    import pymc3 as pm
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/pymc3/__init__.py", line 64, in <module>
    from .stats import *
  File "/Users/wang.12220/opt/anaconda3/lib/python3.6/site-packages/pymc3/stats/__init__.py", line 47, in <module>
    geweke = map_args(az.geweke)
AttributeError: module 'arviz' has no attribute 'geweke'

About this issue

  • Original URL
  • State: closed
  • Created 3 years ago
  • Comments: 20 (9 by maintainers)

Most upvoted comments

This is a PyMC3 issue (old versions of PyMC are not compatible with new versions of ArviZ), not a starry issue! You’ll need to update PyMC3 or downgrade arviz (this is probably the easier approach). @rodluger can probably suggest which would be best. To downgrade ArviZ, use:

python -m pip uninstall arviz
python -m pip install "arviz<0.11"