-
Notifications
You must be signed in to change notification settings - Fork 69
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update to Python 3.6 #360
Update to Python 3.6 #360
Conversation
- since that's what we actually use
FYI, I tested the |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Overall idea here looks good, could use a little clean up with the --channel defaults
bits though.
- conda install --yes python=$TRAVIS_PYTHON_VERSION numpy scipy beautiful-soup six scikit-learn==0.18.1 joblib prettytable python-coveralls pyyaml | ||
- if [ ${TRAVIS_PYTHON_VERSION:0:1} == "2" ]; then conda install --yes configparser logutils mock; fi | ||
- if [ ${WITH_PANDAS_AND_SEABORN} == "true" ]; then conda install --yes pandas seaborn; fi | ||
- conda install --yes --channel defaults --channel conda-forge python=$TRAVIS_PYTHON_VERSION numpy scipy beautifulsoup4 six scikit-learn==0.18.1 joblib prettytable python-coveralls ruamel.yaml |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think all the --channel defaults
in here are unnecessary, and I'm not sure why you don't want to install conda-forge versions of configparser
, pandas
, or seaborn
if they exist. conda-forge is basically always on a newer version than the defaults channel, and what's in there will eventually become the defaults channel.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@dan-blanchard it's not unnecessary since if you don't specify that, the scipy and numpy versions don't use mkl
, they use openblas
instead which is less efficient and slower (at least according to what I have read). It's also not always true that conda-forge
is always on a newer version. They didn't have scikit-learn
0.18.1 for at least a couple of months even after defaults
did.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I do see your point that may be we don't need the --channel defaults
bits for the configparser
, seaborn
etc. but I just have them there so that it's explicit for someone reading it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ah, so it's all to make defaults
take precedence. I didn't get that at first, since defaults
is always included in the channel list. Comment withdrawn.
travis.yml
as well as the conda recipe (logutils
does not seem to be used anywhere,pyyaml
should beruamel.yaml
, andbeautiful-soup
has been renamed tobeautifulsoup4
)conda-forge
for some of the packages that are not indefaults
instead of building them and saving them to thedesilinguist
channel.