ArcGIS Pro 2.2.0 and Python Package Manager

MitchHolley1 · ‎06-27-2018

I updated to Pro 2.2.0 today and noticed the default python environment is now read-only. The previously downloaded packages I had were wiped out with the new update. I need to get the package, spyder, back into my environment. So, I cloned the arcgispro-py3 environment in order to make the install.

Notice that the window says, "Restart ArcGIS Pro for your environment changes to take effect."

So, I clicked 'Exit'.

After the program restarted, I no longer have a cloned environment to select.

Can someone please help?

deactivated-user-ChrisBeaudett · ‎10-21-2018

I have Pro 2.2.3 and I'm seeing the same behavior when attempting to clone an existing environment: the new env just disappears. What's even more frustrating is that it's essentially impossible to get both the arcgis SDK for Python and ArcPy configured in a standalone conda environment w/ Python 3.6+ outside of Pro. So if you want arcgis + arcpy + any custom modules, you can't do it. And, Spyder is also apparently no longer installed w/ Pro. At least, I can't find it. (I'd prefer to use PyCharm anyway, but I'll take any IDE that will let me code + debug in a custom Pro env).

If anyone has done this (custom env with arcgis + arcpy) as of Pro 2.2.3 with Python 3.6+, please let me know your secret sauce.

DanPatterson_Retired · ‎10-22-2018

I use Pro 2.2.3 with python 3.6 and Spyder

/blogs/dan_patterson/2018/07/01/arcgis-pro-your-conda-environments

/blogs/dan_patterson/2018/01/28/spyder

The caveat... The package manager no longer does anything, you have to do a conda install.

Which means you have to have admin rights of some kind to your computer.

If properly installed, Use file explorer and check for the existence of

C:\...Your Install Folder ...\bin\Python\envs\arcgispro-py3\Lib\site-packages\spyder

Plus some other spyder related folders.

If you want to make shortcuts for your desktop, it can be done..

Spyder isn't the only package that is distributed within Pro. If you want the 2.2 tactic reverted...

https://community.esri.com/ideas/15429

otherwise you will have to continue to fill up your system with clones, if you can get them to work

JoshuaBixby · ‎10-22-2018

I have gotten cloning to work, somewhat. For reasons I don't understand, I have to let the cloning dialog sit for 15-20 minutes while it works very slowly in the background to create the new cloned environment. If I close out of the dialog early, the environment is only partially created and doesn't work. Since I have a respectable machine with SSD drives, I can't explain why the cloning is so slow.

deactivated-user-ChrisBeaudett · ‎10-22-2018

Thanks Joshua. I've let the dialog run to completion as well, but still no cloned environment after a restart.

ErikLash1 · ‎10-22-2018

Cloning doesn't actually "clone" your installation. It clones the packages in your installation. To accomplish the clone it downloads all of the packages from online repos as if you were creating a new installation for the first time. So when you don't let if finish out it won't be a complete environment.

Conda everything. ArcGIS Package Manager is not yet up to the task of managing Python environments. It'll get there some day I'm sure but it's not there yet.

Dan Patterson is on the mark. I've set up .bat files which allow me to automate 'cloning' my setup to each of my advanced users machines including Spyder, Anaconda Navigator, and Jupyter Lab while preserving the default ArcGISPro environment in case of a failure or revert or new clone for integration with R etc, requirement.

Or, if you want simplicity and to keep your hard drive from filling up with utterly useless clones, Conda straight into arcgispro-py3. Add Spyder and upgrade the ESRI packages to 1.5 and off you go...If I didn't have users with unique needs that's what I would chose to do.

DanPatterson_Retired · ‎10-22-2018

And conda installs can be made fun as well

/blogs/dan_patterson/2018/10/08/making-conda-package-installs-easier

NicholeSalinger · ‎03-01-2019

David Martinez‌ Alberto Nieto‌

I finally got approval to install tensorflow but running into an install issue with cloning the environment using the Python Package Manager. Finally just copied and pasted the environment as stated in this thread but running into issues updating and installing packages. Please advise. #tensorflowarcgispro-py3 python package manager‌

DanPatterson_Retired · ‎03-01-2019

did you switch into the environment? did you try conda ? a dry-run suggests that the changes are quite a lot, so make sure that it isn't going to break something in other packages

(arcgispro-py3) C:\arc_pro\bin\Python\envs\arcgispro-py3>conda install tensorflow --dry-run
Fetching package metadata .............
Solving package specifications: .

Package plan for installation in environment C:\arc_pro\bin\Python\envs\arcgispro-py3:

The following NEW packages will be INSTALLED:

    _tflow_select:       2.1.0-gpu
    absl-py:             0.7.0-py36_0
    astor:               0.7.1-py36_0
    cudatoolkit:         9.0-1
    cudnn:               7.3.1-cuda9.0_0
    gast:                0.2.2-py36_0
    grpcio:              1.16.1-py36h351948d_1
    keras-applications:  1.0.6-py36_0
    keras-preprocessing: 1.0.5-py36_0
    libprotobuf:         3.6.1-h7bd577a_0
    markdown:            3.0.1-py36_0
    protobuf:            3.6.1-py36h33f27b4_0
    tensorboard:         1.12.2-py36h33f27b4_0
    tensorflow:          1.12.0-gpu_py36ha5f9131_0
    tensorflow-base:     1.12.0-gpu_py36h6e53903_0
    termcolor:           1.1.0-py36_1
    werkzeug:            0.14.1-py36_0

The following packages will be UPDATED:

    cffi:                1.11.5-py36h74b6da3_1     --> 1.12.1-py36h7a1dbc1_0
    ipython:             7.2.0-py36h39e3cac_0      --> 7.3.0-py36h39e3cac_0
    jedi:                0.13.2-py36_0             --> 0.13.3-py36_0
    keyring:             17.1.1-py36_0             --> 18.0.0-py36_0
    markupsafe:          1.1.0-py36he774522_0      --> 1.1.1-py36he774522_0
    openpyxl:            2.5.14-py_0               --> 2.6.0-py36_0
    openssl:             1.1.1a-he774522_0         --> 1.1.1b-he774522_0
    parso:               0.3.2-py36_0              --> 0.3.4-py36_0
    pip:                 19.0.1-py36_0             --> 19.0.3-py36_0
    pluggy:              0.8.1-py36_0              --> 0.9.0-py36_0
    prompt_toolkit:      2.0.8-py_0                --> 2.0.9-py36_0
    py:                  1.7.0-py36_0              --> 1.8.0-py36_0
    pytest:              4.2.0-py36_0              --> 4.3.0-py36_0
    python-dateutil:     2.7.5-py36_0              --> 2.8.0-py36_0
    pyzmq:               17.1.2-py36ha925a31_2     --> 18.0.0-py36ha925a31_0
    setuptools:          40.7.3-py36_0             --> 40.8.0-py36_0
    wheel:               0.32.3-py36_0             --> 0.33.1-py36_0
    win_inet_pton:       1.0.1-py36_1              --> 1.1.0-py36_0‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍