Add Transform component to the pipeline

[deep-diver]

This PR includes the following changes:

• tfx-pipeline.ipynb: modified to demonstrate how Transform can be integrated into the pipeline
• preprocessing.py: handle low-resolution data preprocessing
• preprocessing_full_res.py: handle full-resolution data preprocessing
• model.py and model_full_res.py: replaced parse_tfr and preprocess steps from the previous version with transformed dataset from Transform
• local_pipeline.py: Transform component is added to the local pipeline
• pipeline.py: Transform component is added to the Vertex pipeline. Also, there is a condition to run Transform component in Dataflow
• configs.py: replace data source path to the re-generated TFRecords for the low-resolution dataset

Note: In order to use Transform, we need StatisticsGen and SchemaGen, so they are added to the pipeline as well.

Note: This code changes are verified to work on local and Vertex environment.

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[sayakpaul]

@deep-diver the low-res PR is a dependency of this PR, right? Let's please hold off until it gets resolved.

[deep-diver]

@sayakpaul

this is up for the review as well!

[deep-diver]

@sayakpaul

addressed your comments! thanks for the great comments. I really like them 👍🏼

[deep-diver]

@sayakpaul

fixed the broken parts, and it is verified to be working properly.