search

holoviz-topics / EarthML

Tools for working with machine learning in earth science
https://earthml.holoviz.org
BSD 3-Clause "New" or "Revised" License
94 stars 21 forks source link

Adding appveyor config #81

Closed jsignell closed 5 years ago

jsignell commented 5 years ago

I set up appveyor so that we can get windows builds and see what is missing:

conda env update -f environment.yml -n earthml
Solving environment: ...working... failed
ResolvePackageNotFound: 
  - intake-xarray
  - scipy
  - datashader
  - numba[version='>=0.37.0']
  - datashader
  - scikit-image
  - scipy
  - keras
  - scipy
  - keras
  - theano
  - keras
  - keras-preprocessing=1.0.2
  - dask-ml
  - dask-glm
  - scikit-learn[version='>=0.18']
  - dask-ml
  - dask-glm
  - scipy[version='>=0.18.1']
  - dask-ml
  - dask-searchcv
  - dask-ml
  - numba
  - scikit-learn
  - cartopy
  - numpy=1.10
  - cartopy
  - scipy
  - geopandas
  - gdal==1.11.4
  - geopandas
  - pysal
  - scipy
  - geoviews==1.6.0a2
  - datashader
  - numba[version='>=0.37.0']
  - geoviews==1.6.0a2
  - datashader
  - scikit-image
  - scipy
  - geoviews==1.6.0a2
  - geopandas
  - pysal
  - geoviews==1.6.0a2
  - geoviews-core=1.6.0a2
  - cartopy[version='>=0.14.2']
jsignell commented 5 years ago

EDIT: All the packages are available now for windows, and appveyor is only failing because of this bug: https://github.com/appveyor/ci/issues/2822

jsignell commented 5 years ago

I am seeing a real error from dask_ml on windows:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<timed eval> in <module>

C:\Miniconda36-x64\envs\earthml\lib\site-packages\dask_ml\cluster\spectral.py in fit(self, X, y)
    192                 km = KMeans(
    193                     n_clusters=n_clusters,
--> 194                     random_state=draw_seed(rng, 2 ** 32 - 1, dtype="uint"),
    195                 )
    196             elif self.assign_labels == "sklearn-kmeans":

C:\Miniconda36-x64\envs\earthml\lib\site-packages\dask_ml\_utils.py in draw_seed(random_state, low, high, size, dtype, chunks)
     15         kwargs["chunks"] = chunks
     16 
---> 17     seed = random_state.randint(low, high, **kwargs)
     18     if dtype is not None and isinstance(seed, (da.Array, np.ndarray)):
     19         seed = seed.astype(dtype)

mtrand.pyx in mtrand.RandomState.randint()

ValueError: high is out of bounds for int32
ValueError: high is out of bounds for int32
jsignell commented 5 years ago

Commented there https://github.com/dask/dask-ml/issues/230