change_notes.md

• 20180112 @conchoecia change notes

  • Changed the parallelized version of the program to use tqdm instead of progressbar. This shows the progress of parallelized processes.
  • Added a heuristic to find the optimum number of files per process. This should speed up the analyses drastically if there are many processes on the computer being used.

• 20171223 @conchoecia change notes

  • enable the user to input multiple reference files and multiple directories to search in
  • I did some preliminary testing of bottlenecks in the code and found that it only took 0.23 seconds to save a list of 68k files, but 1166 seconds to do the fast5 io and processing. This program would greatly benefit from parallelization.
  • Removed the extra conditional from : fast5_read_id in read_ids and filepath not in filepaths
    • Going to use a set to save the final filepaths if printing to output. Will only print output at the very end to ensure there are no duplicates. I will also use a set if printing to stdout to avoid this problem. This will also speed up the processing time.
  • Made more feedback for the user as well as a timed progressbar

• 20171221 @conchoecia change notes

  • In extract_read_ids(), removed the os.path.abspath() command since FullPaths action takes care of this.
  • made a _get_file_extension(args.reference) since that code is recycled a few times
  • removed abs_path = os.path.abspath(os.path.realpath(args.fast5_dir)) since FUllPaths figures this out
  • Cleaned up the way that read ids are parsed from the fast5 files and dealt with inconsistencies in file naming.
  • Made timestamps and useful progress feedback for the user print to stderr.
  • Made the program work with both py2 and py3
  • refactored the code and made it more readable.
  • made all the messages say fast5seek since that makes more sense than fast5_in_ref (at least to me!)

• 20180115 @mbhall88 change notes

  • changed name of repository to fast5seek
  • changed all references of fast5_in_ref to fast5seek