I’m completely new to Azure ML, and to be honest, I’m absolutely lost.

To begin with, I’ve been following this tutorial: Create production machine learning pipelines. However, it doesn’t quite work for my code.

My question is: Is there any way in which I can deploy a python code that’s all in a loop in Azure ML?

Context: What are the problems I’m facing?

1. The aforementioned tutorial is structured in such a way that it doesn’t work for my code. For one, it expects a train/test format, which doesn’t work for me, because my code works as an unsupervised manual gridsearch, in which I calculate several parameters by a constant process of combination. The final output are several CSV files with all the possible combinations of parameters.

2. As mentioned before, my whole code is a loop, from when I begin to prepare the data, to the very end when various CSV files are generated. Thus the Azure ML structure (1. Data Preparation Component, 2. Training Component, and 3. Output component) doesn’t work for me. I understand these steps are important so that Azure ML understand what’s going on, but since everything in my code is in a loop I’ve got no idea how to log the metrics.

3. My code uses several customised functions, but I haven’t been able to find any documentation as to how I can call them. I’m not sure if I can just create a “scripts” folder in the “components” directory and that might work.

So far, I haven’t been able to run my code in Azure ML, as I’ve got no idea how to make it work within an Azure ML function. I don’t know if I have to rewrite my code so that it’s adapted to Azure ML, but I do need it to be in a loop, but given how Azure ML works, I’m not sure if that’s a possibility.