csv2ann truncates paths containing dots, lacks write_dir, tests and docs #595

Found while reviewing #591.

csv2ann in wfdb/io/convert/csv.py has some problems:

• The record name is derived with file_name.split(".")[0] (

  wfdb-python/src/wfdb/io/convert/csv.py

  Line 627 in b8914e1

  file_name = file_name.split(".")[0]

  ), which breaks on paths containing dots: ./x.csv becomes "" and data/v1.2/x.csv becomes data/v1. os.path.splitext would avoid this. (csv_to_wfdb also uses the base name only, so the expected output location should be the same for both.)
• There is no write_dir parameter, unlike csv_to_wfdb and wrann, so output paths depend on the input path and the current directory.
• It has no tests.
• It is not included in the API docs (docs/convert.rst only lists csv_to_wfdb).