Towards a Prototype 0.1.0
This PR entails multiple in-sync forward moving developments.
It should be a major step towards (although not yet completing) a prototype with limited, but partially stable functionality for further testing.

• Supports a single backend: duckdb.
• Fixes Add tests for functions #28
  • Add/fix tests for mimic_iv_omop, gibleed_omop, synthea27nj_omop
• Fixes OMOP Extraction of measurement, observation, specimen #60
  • Only if all units for a feature are the same; otherwise raises Error
  • Allow to choose value columns "to keep" or just "encode" with a 1 if present.
• Fixes OMOP extraction of era and other interval style tables #61
  • allow to keep either a) only the event start date, b) only the event end date, or c) an event-interval; the interval-table gets populated across the "stretch" of this event-interval, with all intervals intersecting with the event-interval getting populated with this event
  • Allow to choose value columns "to keep" or just "encode" with a 1 if present.
  • drug_exposure
  • condition_occurrence
  • procedure_occurrence
  • device_exposure
  • drug_era
  • dose_era
  • condition_era
  • episode
• Use download slightly adapted/fixed from ehrapy.data._dataloader.py in ehrdata.dt.dataloader since Feature/physionet2012 dataset #64 for omop demo datasets
• Fixes Consistent logging #65: logging instead of print and rich.print
• Fixes Case insensitive handling of column names #66: Able to handle column names in tables to be of different capitalization; internally puts all column names to lowercase.
• Fixes Create option to extract "presence-indicating" variable #67: can extract either the value of a column in a target data table, or a binary presence-indicating variable; for person P, Variable V, timestep T, then edata.r[P,V,T]=1 if P had V recorded within T, else 0. Implemented via data_field_to_keep=[<optionally more data fields>, "is_present"].