Importing sklearn in Rhino IronPython

Jack_Perry 2017-06-01 05:36:59 UTC #1

Hi,

I've been working on a program in Rhino which requires the sklearn module which you can import for python applications.

However, IronPython is not compatible with this module. I was wondering if there is a way to bypass IronPython and use my local version of Python instead of the one inbuilt to Rhino.

I understand there might not be a solution to this issue.

Thanks,

Jack.

stevebaer 2017-06-01 14:21:51 UTC #2

You can't replace the python engine inside of Rhino with another implementation. The two are incompatible. You can always treat your installed python like another application and have your script in Rhino execute python as a child process.

AndersDeleuran 2017-06-01 14:55:42 UTC #3

Another option would be to implement one of the .NET machine learning libraries instead. My colleague Mateusz has gone quite far with this (including the Owl Grasshopper plugin). I once did a quick test implementing AForge in IronPython/GHPython based on some of his VB code.

Mat_S 2017-06-01 15:59:01 UTC #4

You can call you external python,

github.com

RaghavSood/FBMessageScraper/blob/master/dumper.py#L97


		except KeyError:
			pass #no more messages
	else:
		print "Error in retrieval. Retrying after " + str(error_timeout) + "s"
		print "Data Dump:"
		print json_data
		time.sleep(error_timeout)
		continue
	outfile.write(messages_data)
	outfile.close()	
	command = "python -mjson.tool " + directory + str(offset) + "-" + str(limit+offset) + ".json > " + pretty_directory + str(offset) + "-" + str(limit+offset) + ".pretty.json"
	os.system(command)
	offset = offset + limit
	time.sleep(general_timeout) 


finalfile = open(directory + "complete.json", 'wb')
finalfile.write(json.dumps(messages))
finalfile.close()
command = "python -mjson.tool " + directory + "complete.json > " + pretty_directory + "complete.pretty.json"
os.system(command)

Tom_Holt 2017-06-03 11:10:57 UTC #5

You can call accord.net for kmeans and import from clr.