Skip to main content

Hi all,

 

I try to find an easy solution for the following situation:

I am triggering an Python script in ML Workbench via Action Flow.

This Script is using pycelonis and extracts an OLAP table out of an Analysis.

 

The OLAP table is containing exactly the information, columns etc. how we need the data. My first idea was to create an .xlsx file directly within ML-Workbench from dataframe and get the .xlsx file that is then stored within Jupyter File Browser via another Action Flow module. So far I haven't found a solution to access the files stored on ML-Workbench.

 

Using Celonis Query Data -> CSV Advanced etc. is a total pain because in both modules I have to add all the columns again and do the mapping and all this stuff although the structure itself is already available on Analysis OLAP table level. Another issue is that I wanted to do this within one Action Flow and the Action Flow is not waiting till Python script execution is completed. Current workaround is the use of Sleep module which is also no manageable solution in my opinion.

 

What other ways do you think can be used in order to get the dataframe information from python as .xlsx file so that I can upload it afterwards to Sharepoint by using the Sharepoint module.

 

Thank you very much in advance.

Use python to upload the result xlsx into some AF-supported repository...? But, yes, is kind of messy, and probably you can run into some infosec concerns.


Hi Guillermo,

thanks for your answer.

 

I know that it is possilbe upload files using python but this is something we want to avoid.

First of all due to security concerns and we would like to make use of the Action Flow modules but it seems like getting any kind of files out of ML-Workbench back to Action Flows was not considered during implementation.

 

I also tried to do it via "Download Link" you can get in ML-Workbench by doing a right click on a file in combination with http get file module but this is also not working. I only get an file.html with the information that my browser cannot handle javascript or something like this. I have also seen that when I use the Download link manually I also get first time an html file that I have to open and where I have to agree that this file should be checked with an Anti Virus software etc.

Do you know if it is possible to deactivate this kind of "security step" in ML-Workbench to directly download the file? Maybe than http module would work?

 

Or I was thinking about pushing the file from ML-Workbench into the EMS integrated SFTP server but I am not that expert when it comes to this logic in python.

 

At the end I am very frustrated about the small amount of features you can use in Celonis module / http module when it comes to simple data extraction out of analyses.

 


Hi Guillermo,

thanks for your answer.

 

I know that it is possilbe upload files using python but this is something we want to avoid.

First of all due to security concerns and we would like to make use of the Action Flow modules but it seems like getting any kind of files out of ML-Workbench back to Action Flows was not considered during implementation.

 

I also tried to do it via "Download Link" you can get in ML-Workbench by doing a right click on a file in combination with http get file module but this is also not working. I only get an file.html with the information that my browser cannot handle javascript or something like this. I have also seen that when I use the Download link manually I also get first time an html file that I have to open and where I have to agree that this file should be checked with an Anti Virus software etc.

Do you know if it is possible to deactivate this kind of "security step" in ML-Workbench to directly download the file? Maybe than http module would work?

 

Or I was thinking about pushing the file from ML-Workbench into the EMS integrated SFTP server but I am not that expert when it comes to this logic in python.

 

At the end I am very frustrated about the small amount of features you can use in Celonis module / http module when it comes to simple data extraction out of analyses.

 

Hmmm.... well, for me the ML Workbench looks like an afterthought.

And of course AF was from a company they bought 2-3 years ago (Integromat? now is Make) and is understandable there is still things to iron out.

 

No idea about EMS SFTP, I guess that is not even active in my environments.

 

On the "warning security step" I am pretty sure accepting will set a cookie some place. So if you can find that cookie and put it in your http request, it SHOULD (I stress "should" 😛 ) get you the file. Maybe with 2 https request modules: one to "accept" and retrieve the cookie

 

All said, try to find more info on the EMS SFTP....

 

Sorry!


Reply