Transcript Python & ModelBuilder

Python & ModelBuilder
Overview
• Python/ModelBuilder Concepts
– The Geoprocessor
– Checking some environment variables
– Providing feedback from your model/script
– Starting a script with a model
– Bringing a completed script into ArcToolbox
• Exercises
Why model or write scripts?
•
•
•
•
•
•
Automate workflows
Test hypotheses
Reproducibility
Extend Functionality
Simplification
Quality Control
The Geoprocessing Object
•
•
•
•
•
•
Also known as the "geoprocessor"
Has many methods
Manages layers and selections
Provides access to all of ArcGIS's tools
Provides status of ArcGIS licensing
Handles errors and warning messages
The Geoprocessing Object
gp.Clip_analysis("Roads_2007", "Counties", "Mercer_Roads.shp")
Check Data Existence
• gp.exists() returns a boolean depending on
whether or not the feature class exists
• Check most user-submitted data for existence
– if gp.exists(data):
# continue working
else:
# warn user about missing data
• ArcToolbox will check, but scripts run from the
command line will not
Defining a Workspace
• When the geoprocessor has a workspace
defined, all processing results will be stored in
that folder, geodatabase, or SDE gdb
• gp.workspace = "C:\\Student\\"
• Once defined, we do not need to specify
absolute paths to datasets or outputs located
in that workspace
Describe Method
• gp.describe returns an object with numerous
properties based on the type of data passed
– Feature Classes (shapefiles, geodatabase layers)
– Feature Datasets
– Workspaces
– Raster Data
• gp.describe provides you with GIS-specific
information about your data
Using the Describe Method
• Use Describe to ensure that data passed to
your model fits certain criteria
– Prevent line or point features being used as Clip
Features
– Add an area and perimeter field to shapefiles if
not present
– Convert rasters to 8-bit unsigned if not already
– Use Python to get a list of .shp files, then process
only the polygon shapefiles
Check License Availability
• Models and scripts might be dependent on
the level of ArcGIS available
• The gp.CheckProduct("level") returns
"Available" or "NotLicensed" depending on
the availability of the product level passed
• The gp.SetProduct("level") tells the
geoprocessor to use that specific level
Handling Input
• ArcGIS passes information to your Python
script through sys.argv
• For example, a clip script would need to listen
for three inputs
– Input Feature / sys.argv[1]
– Clip Feature / sys.argv[2]
– Output Feature / sys.argv[3]
• Gets tricky when we ask for multiple inputs
Providing Feedback
• Geoprocessor methods to provide feedback
– gp.AddMessage(message string)
– gp.AddWarning(message string)
– gp.AddError(message string)
• Adds messages to the message list
• Message list is simply that – stores all the
errors, warnings and messages generated by
the script
Providing Feedback
• gp.GetMessage() and gp.MessageCount
• GetMessage(x) returns message number x in the
list
• print gp.GetMessage(gp.MessageCount - 1)
– prints the last message generated
• Print messages at the end of script or when an
error is caught
• Print before and after a lengthy processing step
Defining a Function
• A Python Function allows you to write a block
of code that accepts arguments and returns
results
• Functions allow you to write something once
and reuse it as needed
• Helps debugging
• Keeps code clear and concise
• Easier to program
Using Functions
• We use the def statement to start a block of
code that will only be run when our function is
called
• We want our functions to return values we
can use in our geoprocessing
• The last statement in our function should be
return.
Basic Function
• A simple function that adds 1 to value passed
• def addone(x):
return x + 1
• print addone(2)
>>> 3
• var = addone(3)
print var
>>> 4
Multiple Clips
• Like Exercise #2 from last week, we will
provide two lists to a script
– List of Input Features
– List of Clip Features
• We also need to tell the script where to place
the output, using gp.workspace()
Start in ModelBuilder
Export to Python
Our First Function
• We need a custom function that will take a
string of text, split it into pieces and return
those pieces as a list
• SplitMultiInputs will clean up a line of code
and return it as a list
Adding SplitMultiFeatures
• Add the function and split the input string and
the clippers string passed by the Toolbox to
our script
Two Loops
• We will need another loop for each of the
inputs
• The "inputs" loop contains the "clippers" loop
• The loops are structured like so:
for input in inputs: (runs m times)
for clip in clippers: (runs n times)
• The script will then perform mn clips
Completed Script
Back to the Toolbox
• Once our script is complete, we can test it in
IDLE and on the command line
• What we'd really want is to run our script like
any other tool in the Toolbox
• We can add the script back to the Toolbox
• We need to specify how to talk to the script
Add Your Script
• Open ArcToolbox
• Right click on the
Toolbox you want your
script to be located
• Add > Script…
Add Your Script
• Our model takes two
input parameters
– Input Feature Classes
– Clipping Feature Classes
– Output Workspace
• Note that parameters
can be set for multiple
values
• MultiValue = Yes
County Clipper
• We’ll take a small ModelBuilder model that
clips one input using the statewide Counties
layer
• Using Python, we’ll enhance the model to clip
the input 21 times for each of the separate
counties
• Add functionality to clip multiple inputs by
each county
Setting Up Our Model
• We need a County layer as our clip feature,
and a user-defined layer as our input feature
• Make Feature Layer
• Select By Attributes
• Clip (of course)
• User-defined Output Workspace
Basic Model to Python Script
loop just a portion of the model
Exported Python Script
Adding County Tuple
• We need a static list of County names to pass
as part of the Select by Attributes “where”
clause
• Put all the County names into a tuple
• Use a for loop to loop over the Select by
Attributes and Clip steps with different
Counties each time
– for county in counties
Looping Our Model
loop just a portion of the model
Looping Our Model
• For each county in the model
– prepare a SQL statement to select County
– run the selection to create our clipping features
– use selected features to clip
– create 21 output feature classes
– each output layer named input_county
SQL Statement
• SQL statements use single and double quotes
to delimit field and value names
• Python also uses quotes to delimit strings
• Keep track of which quotes are being used
• Escape quotes using backslash (\') or (\")
• SQL: "COUNTY" = 'MERCER'
Python: "\"COUNTY\" = 'MERCER'"
Looping Over SQL Expressions
• Each time the loop is run, we want to change
the SQL statement to select the current
county in the loop
• We can concatenate the parts of the SQL
statement with the county name
• Example:
Changing the Output Feature
• Each output feature name needs to be unique
• Append the current County name
• Wetlands.shp becomes
– Wetlands_Atlantic.shp
– Wetlands_Bergen.shp
–…
• Some string methods are needed
Changing the Output Feature
Finalizing Our Loop
Input and Output
County Layer
Our For Loop
Modified SQL Statement
Create Output Path
Geoprocessing Tasks
USGS Quad Boundaries Clipped
Remember Multiple Inputs
• How will we modify the
script to accept multiple
inputs?
• We'll need another for
loop in the script
• Set the Input parameter
to accept multiple
values
• MultiValue = Yes
Clip To Counties Tool
• If the parameters are
set correctly, the dialog
box will look similar to
the one at right
• Our CountyClipper now
ready to handle
multiple layers, clipping
each by county