765977_en-US

liliconcom

JMP 18 Python Scripting with JMP Workflow Builder

I have a customer where most of his data are parquet files in an s3 bucket. He was super impressed with the JMP 18 Python integration webinar video that goes over loading parquet files into JMP. He wants to combine that python script to an existing JMP workflow that has everything he wants minus that first import step. What would be the best/easiest way to go about this?

 

Re: JMP 18 Python Scripting with JMP Workflow Builder

I don't know a lot about the workflow builder, and I think at present it only handles JSL scripts.  However, the JSL Python interface remains a supported interface.  I would start by building the functionality desired using the pure Python integration but defining the work necessary as Python functions within the file.  Something like my_workflow.py having Python functions to do the bulk of the work.  Then in the same directory, create the code for the workflow in JSL (as long as the files are in the same directory, JMP's Python knows the location and is able to use the Python import mechanism on a local file.)  Then there are two options to import the Python code.

Names Default to Here(1);
Python Submit File("my_workflow.py");
...
Python Execute( {inputs, ... }, {outputs, ...}, "\[
# call my python function(s) here
]\");
...

Or

Names Default to Here(1);
Python Execute( {inputs, ... }, {outputs, ...}, "\[
import jmp
import my_workflow

# run my python function(s)
]\");

..

Re: JMP 18 Python Scripting with JMP Workflow Builder

Recall you mentioning this earlier, sounded like you figured it out but just in case... i'm interpreting the idea as after Python has loaded the Parquet files then it runs the JMP workflow, e.g.,

jmp.run_jsl('''
	include("Parquet Post Python Workflow.jsl");
''');