JupyterLab is a modern interactive development environment (IDE) that allows you to work with code, data, and the Jupyter notebook format. Starting with
v1.5.0, the ArcGIS API for Python can be used in JupyterLab for a truly powerful development experience.
JupyterLab builds upon all of the major components of the classic Jupyter Notebook experience (notebooks, terminal, text editor, file browser, ipywidgets, etc.) to create a flexible and powerful user experience. Click here or here to read more about JupyterLab.
- Install the latest ArcGIS API for Python in a new conda environment. Activate that environment.
jupyterlabin your conda environment via
conda install jupyterlab=2
- Note: you will need a 2.X version of JupyterLab, and at least version 1.8.2 of the ArcGIS API for Python for this to work.
- Versions 1.5 -> 1.8.1 work in JupyterLab 1.X, but have different installation instructions.
- You will need to have
nodejsinstalled on your computer to install widget extensions. Visit get npm, or install in conda via
conda install nodejs.
- Make sure you have the latest compatible version of
nodejsinstalled (14.6 at the time of writing this guide, but can change in the future)
- Make sure you have the latest compatible version of
jupyter labto launch a new jupyterlab instance.
- When jupyter lab is first launched, you will need to enable extensions. Press the jigsaw puzzle piece, and "Enable" to enable instructions.
- Install the
@jupyter-widgets/jupyterlab-managerextension by searching for it and pressing "Install".
- Install the
arcgis-map-ipywidgetextension by searching for it and pressing "Install".
- NOTE: your installed ArcGIS API for Python version must EXACTLY match this installed
arcgis-map-ipywidgetnpm jupyterlab extension, or else it will not work. (ex. 1.8.2 of the Python API installed, 1.8.2 of the
jupyter labextension listwill show you all versions of extensions installed)
- Rebuild Jupyter Lab by pressing "Rebuild":
- If this rebuild fails in the UI, run
jupyter lab buildfrom your OS's command line to rebuild all extensions manually. You also may need to restart your jupyter lab instance.
When prompted, reload the page.
All extensions should be properly loaded for you to use the ArcGIS API for Python with JupyterLab! Let's get started with some basics:
Just like the classic Jupyter Notebook experience, JupyterLab offers a file explorer to open existing notebooks, create new notebooks, and organize your content. JupyterLab's file explorer is on the left pane of the main view.
JupyterLab interacts with the same .ipynb notebook format. The only difference is the user interface, and the additions of some other external extensions.
Using Windows and Tabs
You may notice that Jupyter has a concept of 'windows' and 'tabs', unlike the classic Jupyter Notebook experience. This is a very powerful feature of JupyterLab: you can stack notebooks, place notebooks side by side, organize notebooks by tabs, etc. Simply click and drag any 'tab' as seen below:
Any window can be dragged like this. JupyterLab lets you view and edit file types such as .csv, .json, etc. in new windows, as seen above.
Using Cell Utilities
Similar to the windows and tabs above, JupyterLab allows users to move cells in a notebook by dragging and dropping them. JupyterLab also supports dragging cells from one notebook to another notebook. Simply click the area to the left of the cell you want to move, and drag it wherever:
JupyterLab also lets you select multiple cells by holding the Shift key. You can move these cells as mentioned before, or right click and select 'Copy Cells' to copy them. JupyterLab has many options in their right click context menu worth exploring, including the 'Create New View For Output' option. This allows you to take any cell output and duplicate it in a new window, allowing you to stack it, view it side-by-side, etc.
There are other cool things you can do with cells in JupyterLab not described in this guide, keep on exploring!
Integration with the Map Widget
v1.5.0 of the ArcGIS API for Python introduces many new exciting features for the Map Widget, including 2D rotation, 3D mode, 3D renderer support, and more. You can read this guide page for more information about these other features of the new widget: this guide will highlight the map widget's seamless JupyterLab integration.
MapView class's default view behavior is the same as in the classic Jupyter Notebook environment: the widget is displayed it in a cell's output area. However, you may notice a new UI button that is only visible in a JupyterLab environment:
By pressing this button, your widget will move from your cell's output area into a new window. This window can be tabbed, split, etc. like all other windows. If you want to restore the widget to the notebook it originated from, press this button:
Try this for yourself! If you're reading this guide in a JupyterLab setting, run the below cell:
from arcgis.gis import GIS #Create a map widget like you have done many times before gis = GIS() map1 = gis.map() map1
WebMaps and WebScenes
WebMap and WebScene Objects include this same UI button: run the two below cells to try for yourself!
from arcgis.mapping import WebMap webmap_item = gis.content.get("ab42b088573d4253a22a8b38ee698ccd") webmap = WebMap(webmap_item) webmap
from arcgis.mapping import WebScene webscene_item = gis.content.get("421433baeb8d487b903d4a89df79149b") webscene = WebScene(webscene_item) webscene
Controlling New Windows Programatically
Although you just drove this cool new window functionality with your mouse, you still are using Python: what sort of Python API would this be if we didn't provide you the option to control this functionality programmatically?!
Each instances of a
MapView widget has a
tab_mode property and a
toggle_window_view() method. Setting
tab_mode will change how the map widget goes into a new window. Try running the below cells to experiment with this behavior:
#First, draw a map map2 = gis.map() map2
#Set the tab mode to one of the below options, run the cell, then press the 'new window' button #map2.tab_mode = "split-bottom" #map2.tab_mode = "split-right" #map2.tab_mode = "auto" map2.tab_mode = "tab-after"
After a map is displayed, you can call the
toggle_window_view() method to move the map to a new window. You can then call this method while the widget is already in a new window to restore the widget to the original notebook. (i.e., each call to this method is the functional equivalent of pressing the button). Try running the below cells:
You can also programatically specify the
title (text displayed on the tab) and
tab_mode in each method call:
#First, draw a map map3 = gis.map(mode="3D") map3
#Run this cell to move it in a new window map3.toggle_window_view(title="My 3D Map", tab_mode="split-top")
#Then, run this cell to move it back to the notebook map3.toggle_window_view()
import time #Run this cell to cycle through all tab_modes tab_modes = ['auto', 'split-top', 'split-bottom', 'split-left', 'split-right', 'tab-before', 'tab-after'] for tab_mode in tab_modes: #to new window map3.toggle_window_view(title=tab_mode, tab_mode=tab_mode) time.sleep(4) #to original notebook map3.toggle_window_view() time.sleep(4)
This functionality can be very powerful in many ways: Say you have a list of 3 or more WebScene items, and you want to display them all in different windows, docked and visible in 1 cell call. You would write something like this:
#Run this cell, it works! from arcgis.widgets import MapView tab_modes = ['split-top', 'split-left', 'split-right'] i=0 for webscene_id in ['31874da8a16d45bfbc1273422f772270', '91b46c2b162c48dba264b2190e1dbcff', '46c47340708f446ba7f112f139e8ae5e']: webscene_item = gis.content.get(webscene_id) map4 = MapView(gis=gis, item=webscene_item, mode="3D") map4.toggle_window_view(title=webscene_item.title, tab_mode=tab_modes[i]) i+=1
See the API reference for more information on programatic control of the Map Widget.
JupyterLab implements many features you can leverage right now to make the most of your experience with the ArcGIS API for Python. Try it out, and please leave us a comment, suggest a feature, or report a bug! We look forward to seeing what you will create in JupyterLab!