editing trial results to align with target trial output

[morganle-48]

No description provided.

[morganle-48]

I don't think this will be the final form of the output - perhaps you can think of a nicer way to store these values in the final table @yiwen-h to make them easier to plot eventually? I am however quite happy I got to the point of taking yearly snapshots and pulling that all together into a snapshot results data frame at the end. NOTE: I've left the original results data frame as is for now, if you like what I've done we can consider dropping it.

[yiwen-h]

Works well and I see the aim!

Opportunities for improvement (I'll make issues for them):

• We could move away from naming all the column names individually in process_model_results and process_snapshot_results - too much opportunity for errors as we develop the model and change the things we record in results. Consider just looping through all columns in the df or passing in a dict?
• snapshot_results_df seems to basically repeat everything in results_df - not DRY. let's just store results_df with snapshot_time as well?
• instead of death_from_{modality} we could just record treatment_modality_at_death or something similar (then we just have one column instead of 6.