MLOps with Azure ML

Proposal: split train.py into train.py and train_aml.py (#219)

This change splits train.py into two files.

The new train.py is standalone, and has no references to AzureML. It defines three functions, split_data to split a dataframe into test/train data, and train_model which takes the test/train data and a parameter object and trains the model, and get_model_metrics, which evaluates metrics about the model. The script can be run locally, in which case it loads a dataset from a file.

The second file, train_aml.py contains reasonably general AzureML logic. It reads data from a dataset, then calls the split_data function from train.py. It loads input parameters from a config file and logs them, then calls train_model from train.py. It then uploads the model and logs any metrics returned by get_model_metrics.

The hope with these changes is to demonstrate a simple interface for integrating an existing ML script with MLOpsPython, as well as providing an example for how the core ML functionality can be invoked in multiple ways for development purposes.

Co-authored-by: Bryan J Smith <[email protected]>

MLOps with Azure ML

Getting Started fixes for the Variable Group section (#221)

Some fixes based on Eugene's feedback.
- Ordered the descriptions to match the table
- Added ACI description
- Added location description and link
- Fixed ACI acronym description

Additional changes
- Added a clarification that the variables with Azure resource names don't have to exist yet
- Removed 'diabetes' from the name of ACI and AKS deployments and switched to generic 'mlops' to match the other resource naming convention

MLOps with Azure ML

Getting started refactor (#216)

To address feedback about some confusion when following the guide (especially to do with service connections) I have refactored the document structure.

- Optional tasks / extra info has been moved down to either 'Further Exploration' (directly for getting started) or 'Next Steps' (specifically for integrating ML code.)
- Grammar fixes and simplified wording in parts
- Removed manual word wrapping to let autoformat do its thing
- Added a table of contents
- Update ARM service connection screenshot. Fixes #215
- Replace some outdated guide images. Fixes #210

MLOps with Azure ML

125675

Dataset versioning (#218)

MLOps with Azure ML

Remove BuildId parameter (#214)

Remove the need to add the BuildId parameter and tagging logic to the train/evaluate scripts. The tags are instead specified when the experiment is submitted. The register script pulls the BuildId from the tag.

MLOps with Azure ML

Sudivate/environment (#217)

reusable aml enviroment

MLOps with Azure ML

Rename pipeline YAMLs for clarity (#212)

* Rename pipeline and pipeline template YAMLs to new convention

* Rename references to pipelines and pipeline templates

* Sort pipeline entries in code_description.md

MLOps with Azure ML

Pipeline variables comments and clean up (#211)

* adding comments for variables

* update default model name

MLOps with Azure ML

124318

Configuration for environment customization (#206)

MLOps with Azure ML

fixing pipeline condition (#209)

Fixing regression introduced in #207 - don't continue pipeline on lint/unit test errors.