-
-
Notifications
You must be signed in to change notification settings - Fork 26.4k
Closed
Description
Not sure if this is addressed somewhere, but with the current migration to pytest, it would be nice if pytest was able to correctly collect unit tests from the scikit-learn root directory. Currently,
cd scikit-learn
py.test -sv .
fails (partial traceback below) because it believes that benchmarks are unit tests a and consequently fails to import some packages.
py.test -sv sklearn/
works correctly. There should be a way to modify the test discovery in pytest settings somewhere I imagine..
(partial traceback below)
py.test -sv .
==================================================================== test session starts ====================================================================
platform linux -- Python 3.6.1, pytest-3.0.7, py-1.4.33, pluggy-0.4.0 -- /home/rth/.miniconda3/envs/sklearn-env/bin/python
cachedir: .cache
rootdir: /home/rth/src/scikit-learn, inifile: setup.cfg
plugins: xdist-1.16.0
collecting 0 itemsPartial import of sklearn during the build process.
collecting 0 items / 17 errorsinput data sparsity: 0.050000
true coef sparsity: 0.000100
test data sparsity: 0.027400
model sparsity: 0.000024
r^2 on test data (dense model) : 0.232730
r^2 on test data (sparse model) : 0.232730
collecting 0 items / 17 errors================================================================================
# Text vectorizers benchmark
================================================================================
Using a subset of the 20 newsrgoups dataset (11314 documents).
This benchmarks runs in ~20 min ...
^C^C
========================================================================== ERRORS ===========================================================================
______________________________________________________________ ERROR collecting test_speed.py _______________________________________________________________
test_speed.py:17: in <module>
ipython.magic("%timeit row_norms(Xcsr)")
E AttributeError: 'NoneType' object has no attribute 'magic'
______________________________________________________________ ERROR collecting test_speed.py _______________________________________________________________
test_speed.py:17: in <module>
ipython.magic("%timeit row_norms(Xcsr)")
E AttributeError: 'NoneType' object has no attribute 'magic'
___________________________________________________ ERROR collecting benchmarks/bench_isolation_forest.py ___________________________________________________
benchmarks/bench_isolation_forest.py:10: in <module>
import matplotlib.pyplot as plt
E ModuleNotFoundError: No module named 'matplotlib'
[and so on]
Metadata
Metadata
Assignees
Labels
No labels