Select to view content in your preferred language

ArcGIS Python API in Fabric?

300
4
03-03-2025 09:41 PM
SimonGIS
Regular Contributor

Is it possible to install the ArcGIS Python API into Fabric?

I am needing to work with getting a geometry from a WFS service, and I believe the ArcGIS Python API might support this.  From there I would pass it into spatiotemporal join tool in ArcGIS Geoanalytics for Fabric.

0 Kudos
4 Replies
SBattersby
Esri Contributor

I found this reference that might help: https://www.proserveit.com/blog/installing-the-arcgis-python-module-in-a-fabric-notebook

 

I have been successful in installing the API in a notebook using the instructions in the web page listed above:

 

%pip install ujson
%pip install six
%pip install requests
%pip install cryptography
%pip install cachetools
%pip install lxml
%pip install ntlm_auth
%pip install requests_oauthlib
%pip install requests_toolbelt
%pip install requests_ntlm
%pip install geomet
%pip install puremagic
%pip install dask[dataframe]
%pip install keyring
%pip install pylerc
%pip install pyshp
%pip install arcgis --no-deps

 

From there I can add another cell and import and look at the version, which suggests to me that it should work:

import arcgis
print(arcgis.__version__)
 
I don't know if this will get you far enough along, but hopefully it helps at least a bit... I'll admit my ArcGIS Python API knowledge is a bit slim. 🙂
 
-Sarah.
0 Kudos
SimonGIS
Regular Contributor

Thanks Sarah, will try this and report back.

MarkWestergaard
Occasional Contributor

Take this with a huge grain of salt since I have not done much testing, but I believe you could get away with just the %conda magic command to install the arcgis library and dependencies in a Fabric notebook (PySpark Runtime 1.3):

%conda install -c esri arcgis

If this works, you can probably create a new environment in your Fabric workspace with arcgis included so you don't have to run this every time. I will post an update as I do additional testing.

SBattersby
Esri Contributor

Your suggestion of just using %conda install -c esri arcgis worked for me.  Definitely seems like a cleaner pathway to get things up and running

0 Kudos