Cleaned picks.csv formatting with the csv library

[lennijusten]

I described issue #9 a while ago and here is the PR with the fix. The fix is a two-part solution.

First, I open the picks.csv file in the write w mode with the csv.writer from the csv library and write the header row. The write w mode will overwrite any existing file with the same filename.

I then open picks.csv in the append a mode and write the results to picks.csv by looping through each batch. The results are converted from arrays to lists with the tolist() method which removes the excess white spaces. I then append the results to the csv with the writerow() method by passing the results in list format. This automatically delimits the csv columns into fname, itp, tp_prob, its, ts_prob and adresses any issues like overfull lines automatically (see issue #6).

I have tried and tested the method and it works as hoped. When reading the csv files with pandas, one should still note that the pandas.read_csv() method will default to reading all entries as strings. This is described fully in a stackoverflow post. An easy fix is using the literal_eval method from ast.

>>> from ast import literal_eval
>>> literal_eval('[1.23, 2.34]')
[1.23, 2.34]

Best,
Lenni

[lennijusten]

Here is an excel copy of an example output.

picks.xlsx