fixing 3 bugs & giving extra info for scoring

[SunnyBingoMe]

• granularity decision was only 1-unit, e.g. 1min, 1h, 1..., but not for 5min etc. and causing missing of non-matching items
• pd.drop_duplicate was dropping values instead of index
• the first TS item was missing due to _process_long_term_data using wrong range/index
• now providing residuals etc. for scoring