Refactor process_parquet

[cbutsko]

The existing process_parquet functions needs to be refactored:

• remove redundant columns used as index so that the function is compatible with the Phase II extractions
• replace indexing functions that rely on month value with more generic ones
• add the handling of the labels from the new legend

[jdegerickx]

@cbutsko, in branch point-extractions, I just revised handling of crop types in process_parquet using legend from artifactory and assuming the dataset has phase II labels, no longer Phase I labels

[cbutsko]

In this PR, the original process_parquet got refactored #124