Open kalefranz opened 7 years ago
From @asmeurer on August 25, 2015 20:5
Yes, it is. Note that there have been some improvements to it in recent versions of conda-build
.
From @arsenovic on August 25, 2015 20:13
so, the usage is to:
git clone xyz
conda develop /path/to/xyz
?
From @Hornswoggles on June 23, 2016 5:33
Bump. I also don't understand how this is supposed to be used. I've read through the documentation. Can you give an example? The documentation as written is insufficient.
Ok, so I ran the command and it created the conda.pth fils in site-packages. However python can't find the module. How is this .pth supposed to work? Can't find anything in the docs.
I had a lot of problems when trying to install a development version of a package into a virtual environment and using it inside spyder. Here are the steps that worked for me (Windows 10-64bit): Note: All commands are run inside anaconda-prompt. Do not activate the virtual environment as you will not be able to use conda-build commands.
Install conda-build which includes conda-develop command.
$ conda install conda-build
(For Spyder users only) Clone the git repository you want to build into the site-packages folder of the virtual environment you are using. Otherwise only a link is created and spyder does not see the package. Alternatively you can add a path to the git repo directory inSpyder>Tools>PYTHONPATH manager
Run conda develop and specify the path of git repo and a virtual env to install it into
$ conda-develop <path to git repo> -n <env name>
This should allow you to use the development version of a library when using spyder from a conda virtual environment.
Bump. Docs are really insufficient for this command. Keep getting NoPackagesFound
error.
Got around the first error because conda-build
was not installed properly. Installing that gave me:
InstallError: Error: the following specs depend on 'conda' and can only be installed
into the root environment: conda-build
Dealt with that by sudo conda install conda-build
from root dir.
Now I am getting
$ conda-develop path/to/packake -n myPython2env
FileNotFoundError: [Errno 2] No such file or directory: '/home/user/anaconda3/envs/myPython2env/lib/python3.5/site-packages/conda.pth'
Odd thing is that it is looking for python3.5 in my python2.7 env.
$ conda-develop path/to/packake -n myPython35env
FileNotFoundError: [Errno 2] No such file or directory: '/home/user/anaconda3/envs/myPython35env/lib/python3.5/site-packages/conda.pth'
Works fine.
Thanks for the reminder. This function has not been maintained in some time. Ideally, it would create environments for you, with your code set up with python setup.py develop
or pip install -e .
. Unfortunately, time has not permitted me to make this what I want it to be.
Instead, I recommend creating whatever environment you want, activating it, and then running python setup.py develop
or pip install -e .
so that your package is installed in develop mode in that environment.
@msarahan do you have any recommendations for how to create a conda environment based on the runtime dependencies declared in a meta.yaml file? Ideally, the file would go through Jinja templating.
I've noticed that that conda build test environments look very similar to what I want, except that they try and install the built package. For conda develop, we would want a test environment excluding the package under test.
Nothing that exists as a nice bundled command right now. If you'd like to add a PR to change conda-develop, it would be most welcome.
It should be readily achievable with something like:
from conda_build import api
from conda_build.environ import create_env
# this gets tricky for recipes with multiple outputs, but is simple for recipes with only one. To simplify, pick only the first metadata object.
metadata = api.render(path_to_recipe)[0][0]
# collect dependencies from run and test
deps = metadata.get_value('requirements/run') + metadata.get_value('test/requires')
create_env(desired_location, deps, env='run', config=metadata.config, subdir=metadata.config.subdir)
Putting that behind a nice CLI is more than I have time for right now, but if you have questions, I'll be happy to answer them.
I ran the 'conda-develop
It seems like even from within my new environment, if I try to import the development version, it will find the installed version from the default environment. Is there an easy way around this?
Seems to me that it would be better to have a --develop
flag for conda-build. Then each time you ran conda-build it would compile all your non python stuff as well via build.sh.
It would install all the built binaries to bin, lib etc. It could use a temporary symlink to handle anything that the build script copies to the site-packages directory.
Obviously there are more details to work out though.
Beside the built binaries, there are also the activate.d / deactivate.d scripts to consider. I mainly (or rather, only) use these to set environment variables. Perhaps something to keep in mind for conda/conda#6820?
Why this would be a nice feature to have, is because using setup.py develop
doesn't take care of dependencies. Currently, I have to do:
conda build mypkg
conda create -n myenv --use-local mypkg # will pull in deps
source activate myenv
conda remove mypkg
python setup.py develop
cd ~/.conda/envs/myenv
mkdir -p etc/conda/activate.d
mkdir -p etc/conda/deactivate.d
# Populate scripts in activate.d and deactivate.d
Non-python binaries have to be built twice, i.e. once when I build the package, and again in my develop environment.
Of course, I'm not complaining, as any of us here could take the time to make this work, if we had it. Just making the case for something better.
That's more or less my workflow as well.
One way to deal with the built binaries, activate scripts, etc is to move these to one subpackage, and the Python source to another. Setting up a Python dev environment would then amount to installing only the former subpackage, while using conda develop
to do a "symlinked install" of the latter.
This will undoubtedly work (haven't tested it though). But I don't like it very much because it splits the original package up into multiple outputs, while their contents clearly belongs together.
Also, everybody who wants to replicate this kind of setup has to fiddle with multiple outputs
sections in their recipe. I'd much rather have a more evolved conda develop
that would handle this situation automatically.
However, this begs the question: would the only use of such a command be Python development, or does it need to target other languages as well?
If the answer is no, then it would be straightforward to design IMHO ("install dependencies and build & copy everything except things in site-packages
-- that should be symlinked").
If the answer is yes, I wouldn't know where to begin.
Anyway, I think we're straying a bit offtopic here. Perhaps time for a separate issue?
Is someone working on a fix to this? I'm asking because, if not, I'd love to take a crack at this. I've been spending a lot of free time recently trying to get the new windows terminal to build, and this seems a lot more fun/useful.
No, I don't think anyone is actively working on this right now. If you have questions on anything, let us know.
I've been spending a lot of free time recently trying to get the new windows terminal to build
This also sounds like great fun to me.
No, I don't think anyone is actively working on this right now. If you have questions on anything, let us know.
Awesome! I'll try to wrap my head around it first until I have some specific questions.
I see now that this is quite a ball of yarn 🤔
Just take it in small steps. conda-debug
already lays a decent foundation. Try to think of it as conda-debug
but with an extra layer to connect some external source into the created environment. This is hard to generalize, but easy enough for python stuff. You'd typically just create an env that combined build, host, run, and test deps from the recipe, and then run pip install -e <path to source code>
What conda-debug
already does is to create the build and host envs, or the run/test env, not all rolled into one. I hope it won't be too hard to have a way to combine them.
After that, the tricky thing to get right is the user workflow. Where does the environment live on the user's hard drive by default? How can people tell their IDE's to use that environment in an easy way? Perhaps the IDEs could recognize conda recipes in a standard way and offer to use conda-develop when hacking on a project?
so is conda develop
ready to be used or should we still be using pip install -e
?
Currently, I have to do:
conda build mypkg conda create -n myenv --use-local mypkg # will pull in deps source activate myenv conda remove mypkg
This procedure doesn't seem to work anymore in conda 4.7 because the remove
step will automatically remove (dangling?) dependencies too.
Apparently there is conda install --only-deps
, which should be able to replace the install/uninstall routine to get dependencies.
Yes, --only-deps
is the recommended approach. It has been around since conda 4.4, I think, but has become especially important with the changes in conda 4.7, as you mention.
@brando90 pip install -e
is still recommended. Conda develop has not seen any development lately.
I'd love to clear up some of the conda docs, and this is definitely one I think could benefit. I'm happy to go and do so, but so I understand before I do, what are the limitations of conda develop that make it not suitable for use currently?
@msarahan have you got a summary of the limitations?
I haven't looked at conda-develop in so long that I can't even begin to say what it does, let alone its shortcomings. Given my recommendations above regarding conda-debug and pip install, you may want to try to replicate that with conda-develop. I don't think you can, but that will show you its shortcomings.
I think a user would expect that it would put your environment in the state that it would be if the filepath in its current state had conda build run, the artifact uploaded, and they had installed that artifact - is that runtime agnostic? For compiled code I guess there would have to be some logic to force recompilation on file change I guess, but that wouldn't to be in a first pass.
Would you be happy for me to make a documentation update to reflect that it's deprecated functionality? I understand the difficulties, but I must have been on the conda develop doc page 20 times over the years things thinking it would do what pip -e
does!
how do I see the packages that I've installed in development mode?
I ran conda develop .
and it seems it installed it:
(automl-meta-learning) brandomiranda~/automl-meta-learning/automl ❯ conda develop .
path exists, skipping /Users/brandomiranda/automl-meta-learning/automl
completed operation for: /Users/brandomiranda/automl-meta-learning/automl
however, I get errors when I run my scripts:
(automl-meta-learning) brandomiranda~/automl-meta-learning/automl/automl/meta_optimizers ❯ python differentiable_SGD.py
Traceback (most recent call last):
File "differentiable_SGD.py", line 8, in <module>
from automl.utils.torch_utils import helloworld
ModuleNotFoundError: No module named 'automl.utils'
but when I do conda list I don't see anything I recognize:
(automl-meta-learning) brandomiranda~/automl-meta-learning/automl/automl/meta_optimizers ❯ conda list
# packages in environment at /Users/brandomiranda/miniconda3/envs/automl-meta-learning:
#
# Name Version Build Channel
appnope 0.1.0 py37_0
asn1crypto 1.3.0 py37_0
astroid 2.3.3 py37_0
attrs 19.3.0 py_0
backcall 0.1.0 py37_0
beautifulsoup4 4.8.2 py37_0
blas 1.0 mkl
bleach 3.1.0 py37_0
bzip2 1.0.8 h1de35cc_0
ca-certificates 2019.11.27 0
certifi 2019.11.28 py37_0
cffi 1.13.2 py37hb5b8e2f_0
chardet 3.0.4 py37_1003
conda 4.8.1 py37_0
conda-build 3.18.11 py37_0
conda-package-handling 1.6.0 py37h1de35cc_0
cryptography 2.8 py37ha12b0ac_0
cycler 0.10.0 py37_0
dbus 1.13.12 h90a0687_0
decorator 4.4.1 py_0
defusedxml 0.6.0 py_0
entrypoints 0.3 py37_0
expat 2.2.6 h0a44026_0
filelock 3.0.12 py_0
freetype 2.9.1 hb4e5f40_0
gettext 0.19.8.1 h15daf44_3
glib 2.63.1 hd977a24_0
glob2 0.7 py_0
icu 58.2 h4b95b61_1
idna 2.8 py37_0
importlib_metadata 1.3.0 py37_0
intel-openmp 2019.4 233
ipykernel 5.1.3 py37h39e3cac_1
ipython 7.11.1 py37h39e3cac_0
ipython_genutils 0.2.0 py37_0
ipywidgets 7.5.1 py_0
isort 4.3.21 py37_0
jedi 0.15.2 py37_0
jinja2 2.10.3 py_0
jpeg 9b he5867d9_2
jsonschema 3.2.0 py37_0
jupyter 1.0.0 py37_7
jupyter_client 5.3.4 py37_0
jupyter_console 6.0.0 py37_0
jupyter_core 4.6.1 py37_0
kiwisolver 1.1.0 py37h0a44026_0
lazy-object-proxy 1.4.3 py37h1de35cc_0
libarchive 3.3.3 h786848e_5
libcxx 4.0.1 hcfea43d_1
libcxxabi 4.0.1 hcfea43d_1
libedit 3.1.20181209 hb402a30_0
libffi 3.2.1 h475c297_4
libgfortran 3.0.1 h93005f0_2
libiconv 1.15 hdd342a3_7
liblief 0.9.0 h2a1bed3_2
libpng 1.6.37 ha441bb4_0
libsodium 1.0.16 h3efe00b_0
libtiff 4.1.0 hcb84e12_0
libxml2 2.9.9 hf6e021a_1
lz4-c 1.8.1.2 h1de35cc_0
lzo 2.10 h362108e_2
markupsafe 1.1.1 py37h1de35cc_0
matplotlib 3.1.1 py37h54f8f79_0
mccabe 0.6.1 py37_1
mistune 0.8.4 py37h1de35cc_0
mkl 2019.4 233
mkl-service 2.3.0 py37hfbe908c_0
mkl_fft 1.0.15 py37h5e564d8_0
mkl_random 1.1.0 py37ha771720_0
more-itertools 8.0.2 py_0
nbconvert 5.6.1 py37_0
nbformat 4.4.0 py37_0
ncurses 6.1 h0a44026_1
ninja 1.9.0 py37h04f5b5a_0
notebook 6.0.2 py37_0
numpy 1.18.1 py37h7241aed_0
numpy-base 1.18.1 py37h6575580_0
olefile 0.46 py37_0
openssl 1.1.1d h1de35cc_3
pandoc 2.2.3.2 0
pandocfilters 1.4.2 py37_1
parso 0.5.2 py_0
pcre 8.43 h0a44026_0
pexpect 4.7.0 py37_0
pickleshare 0.7.5 py37_0
pillow 7.0.0 py37h4655f20_0
pip 19.3.1 py37_0
pkginfo 1.5.0.1 py37_0
prometheus_client 0.7.1 py_0
prompt_toolkit 2.0.10 py_0
psutil 5.6.7 py37h1de35cc_0
ptyprocess 0.6.0 py37_0
py-lief 0.9.0 py37h1413db1_2
pycosat 0.6.3 py37h1de35cc_0
pycparser 2.19 py37_0
pygments 2.5.2 py_0
pylint 2.4.4 py37_0
pyopenssl 19.1.0 py37_0
pyparsing 2.4.6 py_0
pyqt 5.9.2 py37h655552a_2
pyrsistent 0.15.6 py37h1de35cc_0
pysocks 1.7.1 py37_0
python 3.7.6 h359304d_2
python-dateutil 2.8.1 py_0
python-graphviz 0.13.2 pypi_0 pypi
python-libarchive-c 2.8 py37_13
pytorch 1.4.0 py3.7_0 pytorch
pytz 2019.3 py_0
pyyaml 5.2 py37h1de35cc_0
pyzmq 18.1.0 py37h0a44026_0
qt 5.9.7 h468cd18_1
qtconsole 4.6.0 py_1
readline 7.0 h1de35cc_5
requests 2.22.0 py37_1
ripgrep 11.0.2 he32d670_0
ruamel_yaml 0.15.87 py37h1de35cc_0
send2trash 1.5.0 py37_0
setuptools 44.0.0 py37_0
sip 4.19.8 py37h0a44026_0
six 1.13.0 py37_0
soupsieve 1.9.5 py37_0
sqlite 3.30.1 ha441bb4_0
terminado 0.8.3 py37_0
testpath 0.4.4 py_0
tk 8.6.8 ha441bb4_0
torchvision 0.5.0 py37_cpu pytorch
torchviz 0.0.1 pypi_0 pypi
tornado 6.0.3 py37h1de35cc_0
tqdm 4.41.1 py_0
traitlets 4.3.3 py37_0
urllib3 1.25.7 py37_0
wcwidth 0.1.7 py37_0
webencodings 0.5.1 py37_1
wheel 0.33.6 py37_0
widgetsnbextension 3.5.1 py37_0
wrapt 1.11.2 py37h1de35cc_0
xz 5.2.4 h1de35cc_4
yaml 0.1.7 hc338f04_2
zeromq 4.3.1 h0a44026_3
zipp 0.6.0 py_0
zlib 1.2.11 h1de35cc_3
zstd 1.3.7 h5bba6e5_0
anyone know whats going on or at least where the packages I've installed in development mode would show up?
how do I see the packages that I've installed in development mode?
I ran
conda develop .
and it seems it installed it:(automl-meta-learning) brandomiranda~/automl-meta-learning/automl ❯ conda develop . path exists, skipping /Users/brandomiranda/automl-meta-learning/automl completed operation for: /Users/brandomiranda/automl-meta-learning/automl
however, I get errors when I run my scripts:
(automl-meta-learning) brandomiranda~/automl-meta-learning/automl/automl/meta_optimizers ❯ python differentiable_SGD.py Traceback (most recent call last): File "differentiable_SGD.py", line 8, in <module> from automl.utils.torch_utils import helloworld ModuleNotFoundError: No module named 'automl.utils'
but when I do conda list I don't see anything I recognize:
(automl-meta-learning) brandomiranda~/automl-meta-learning/automl/automl/meta_optimizers ❯ conda list # packages in environment at /Users/brandomiranda/miniconda3/envs/automl-meta-learning: # # Name Version Build Channel appnope 0.1.0 py37_0 asn1crypto 1.3.0 py37_0 astroid 2.3.3 py37_0 attrs 19.3.0 py_0 backcall 0.1.0 py37_0 beautifulsoup4 4.8.2 py37_0 blas 1.0 mkl bleach 3.1.0 py37_0 bzip2 1.0.8 h1de35cc_0 ca-certificates 2019.11.27 0 certifi 2019.11.28 py37_0 cffi 1.13.2 py37hb5b8e2f_0 chardet 3.0.4 py37_1003 conda 4.8.1 py37_0 conda-build 3.18.11 py37_0 conda-package-handling 1.6.0 py37h1de35cc_0 cryptography 2.8 py37ha12b0ac_0 cycler 0.10.0 py37_0 dbus 1.13.12 h90a0687_0 decorator 4.4.1 py_0 defusedxml 0.6.0 py_0 entrypoints 0.3 py37_0 expat 2.2.6 h0a44026_0 filelock 3.0.12 py_0 freetype 2.9.1 hb4e5f40_0 gettext 0.19.8.1 h15daf44_3 glib 2.63.1 hd977a24_0 glob2 0.7 py_0 icu 58.2 h4b95b61_1 idna 2.8 py37_0 importlib_metadata 1.3.0 py37_0 intel-openmp 2019.4 233 ipykernel 5.1.3 py37h39e3cac_1 ipython 7.11.1 py37h39e3cac_0 ipython_genutils 0.2.0 py37_0 ipywidgets 7.5.1 py_0 isort 4.3.21 py37_0 jedi 0.15.2 py37_0 jinja2 2.10.3 py_0 jpeg 9b he5867d9_2 jsonschema 3.2.0 py37_0 jupyter 1.0.0 py37_7 jupyter_client 5.3.4 py37_0 jupyter_console 6.0.0 py37_0 jupyter_core 4.6.1 py37_0 kiwisolver 1.1.0 py37h0a44026_0 lazy-object-proxy 1.4.3 py37h1de35cc_0 libarchive 3.3.3 h786848e_5 libcxx 4.0.1 hcfea43d_1 libcxxabi 4.0.1 hcfea43d_1 libedit 3.1.20181209 hb402a30_0 libffi 3.2.1 h475c297_4 libgfortran 3.0.1 h93005f0_2 libiconv 1.15 hdd342a3_7 liblief 0.9.0 h2a1bed3_2 libpng 1.6.37 ha441bb4_0 libsodium 1.0.16 h3efe00b_0 libtiff 4.1.0 hcb84e12_0 libxml2 2.9.9 hf6e021a_1 lz4-c 1.8.1.2 h1de35cc_0 lzo 2.10 h362108e_2 markupsafe 1.1.1 py37h1de35cc_0 matplotlib 3.1.1 py37h54f8f79_0 mccabe 0.6.1 py37_1 mistune 0.8.4 py37h1de35cc_0 mkl 2019.4 233 mkl-service 2.3.0 py37hfbe908c_0 mkl_fft 1.0.15 py37h5e564d8_0 mkl_random 1.1.0 py37ha771720_0 more-itertools 8.0.2 py_0 nbconvert 5.6.1 py37_0 nbformat 4.4.0 py37_0 ncurses 6.1 h0a44026_1 ninja 1.9.0 py37h04f5b5a_0 notebook 6.0.2 py37_0 numpy 1.18.1 py37h7241aed_0 numpy-base 1.18.1 py37h6575580_0 olefile 0.46 py37_0 openssl 1.1.1d h1de35cc_3 pandoc 2.2.3.2 0 pandocfilters 1.4.2 py37_1 parso 0.5.2 py_0 pcre 8.43 h0a44026_0 pexpect 4.7.0 py37_0 pickleshare 0.7.5 py37_0 pillow 7.0.0 py37h4655f20_0 pip 19.3.1 py37_0 pkginfo 1.5.0.1 py37_0 prometheus_client 0.7.1 py_0 prompt_toolkit 2.0.10 py_0 psutil 5.6.7 py37h1de35cc_0 ptyprocess 0.6.0 py37_0 py-lief 0.9.0 py37h1413db1_2 pycosat 0.6.3 py37h1de35cc_0 pycparser 2.19 py37_0 pygments 2.5.2 py_0 pylint 2.4.4 py37_0 pyopenssl 19.1.0 py37_0 pyparsing 2.4.6 py_0 pyqt 5.9.2 py37h655552a_2 pyrsistent 0.15.6 py37h1de35cc_0 pysocks 1.7.1 py37_0 python 3.7.6 h359304d_2 python-dateutil 2.8.1 py_0 python-graphviz 0.13.2 pypi_0 pypi python-libarchive-c 2.8 py37_13 pytorch 1.4.0 py3.7_0 pytorch pytz 2019.3 py_0 pyyaml 5.2 py37h1de35cc_0 pyzmq 18.1.0 py37h0a44026_0 qt 5.9.7 h468cd18_1 qtconsole 4.6.0 py_1 readline 7.0 h1de35cc_5 requests 2.22.0 py37_1 ripgrep 11.0.2 he32d670_0 ruamel_yaml 0.15.87 py37h1de35cc_0 send2trash 1.5.0 py37_0 setuptools 44.0.0 py37_0 sip 4.19.8 py37h0a44026_0 six 1.13.0 py37_0 soupsieve 1.9.5 py37_0 sqlite 3.30.1 ha441bb4_0 terminado 0.8.3 py37_0 testpath 0.4.4 py_0 tk 8.6.8 ha441bb4_0 torchvision 0.5.0 py37_cpu pytorch torchviz 0.0.1 pypi_0 pypi tornado 6.0.3 py37h1de35cc_0 tqdm 4.41.1 py_0 traitlets 4.3.3 py37_0 urllib3 1.25.7 py37_0 wcwidth 0.1.7 py37_0 webencodings 0.5.1 py37_1 wheel 0.33.6 py37_0 widgetsnbextension 3.5.1 py37_0 wrapt 1.11.2 py37h1de35cc_0 xz 5.2.4 h1de35cc_4 yaml 0.1.7 hc338f04_2 zeromq 4.3.1 h0a44026_3 zipp 0.6.0 py_0 zlib 1.2.11 h1de35cc_3 zstd 1.3.7 h5bba6e5_0
anyone know whats going on or at least where the packages I've installed in development mode would show up?
I came from the manual of pytest
, which says the following three are equivalent:
pip install -e .
python setup.py develop
conda develop
As a data scientist, I generally use conda
as my management tool. Now, if I want to be a package developer, can someone tell me what practice should I follow?
I came from the manual of
pytest
, which says the following three are equivalent:pip install -e . python setup.py develop conda develop
As a data scientist, I generally use
conda
as my management tool. Now, if I want to be a package developer, can someone tell me what practice should I follow?
what are you trying to do?
I suppose I have the same questions as @WenjieZ has:
Now my questions:
@floschl All good questions. I would add how to combine the requirements in the setup.py
and the requirements in the environment.yml
so to have them in a single place?
AFAIK pip install -e .
will install the libraries in the setup.py, and not the one in the conda environment file.
Yes true, pip install -e . will install the package according to setup.py. Is there an equivalent command for conda to install a conda package locally without creating a channel?
From @asmeurer on August 25, 2015 20:5
Yes, it is. Note that there have been some improvements to it in recent versions of
conda-build
.
is there an additional command we need to run before conda develop .
starts working? I am getting a weird error:
❯ conda develop .
Traceback (most recent call last):
File "/Users/brando/anaconda3/bin/conda-develop", line 11, in <module>
sys.exit(main())
File "/Users/brando/anaconda3/lib/python3.7/site-packages/conda_build/cli/main_develop.py", line 72, in main
return execute(sys.argv[1:])
File "/Users/brando/anaconda3/lib/python3.7/site-packages/conda_build/cli/main_develop.py", line 68, in execute
build_ext=args.build_ext, clean=args.clean, uninstall=args.uninstall)
File "/Users/brando/anaconda3/lib/python3.7/site-packages/conda_build/api.py", line 301, in develop
return execute(recipe_dir, prefix, no_pth_file, build_ext, clean, uninstall)
File "/Users/brando/anaconda3/lib/python3.7/site-packages/conda_build/develop.py", line 175, in execute
write_to_conda_pth(sp_dir, pkg_path)
File "/Users/brando/anaconda3/lib/python3.7/site-packages/conda_build/develop.py", line 50, in write_to_conda_pth
with open(c_file, 'a') as f:
FileNotFoundError: [Errno 2] No such file or directory: '/Users/brando/anaconda3/envs/automl-meta-learning/lib/python3.7/site-packages/conda.pth'
I noticed that if you installed a package which requires a change to your python version, you will end up losing the references to packages installed before.
@brando90
pip install -e
is still recommended. Conda develop has not seen any development lately.
@msarahan thanks! Just curious, why isn't conda develop .
simply removed since it seems you said it's just a remake of something that isn't needed?
update:
btw pip install -e .
does not work for me:
xref: proposal to deprecate and remove conda develop
#4251
Hi there, thank you for your contribution!
This issue has been automatically marked as stale because it has not had recent activity. It will be closed automatically if no further activity occurs.
If you would like this issue to remain open please:
NOTE: If this issue was closed prematurely, please leave a comment.
Thanks!
Yes the issue at hand still exists: that conda develop
only has the CLI reference and no docs explaining how it's used https://docs.conda.io/projects/conda-build/en/stable/resources/commands/conda-develop.html
Draft of a modern Python-standards "no setup.py required" develop replacement https://github.com/conda/conda-build/pull/5380
From @arsenovic on August 24, 2015 20:11
im not sure how to use this command. is this supposed to operate like
python setup.py develop
?Copied from original issue: conda/conda#1546