blahah / transrate

Understand your transcriptome assembly
http://hibberdlab.com/transrate
Other
100 stars 34 forks source link

Create a tidy directory structure to contain interim files #148

Closed blahah closed 9 years ago

blahah commented 9 years ago

Currently, passing in multiple assemblies with the same filename, but in different directories will break transrate. This is because the filename (without the path) is used as the identifier when creating interim files.

Also, the working directory fills up with lots of files. This is ugly and makes debugging and interpretation harder for users.

The solution should be to create a neat directory structure. This could use a transrate_results subdirectory in the working directory. This directory would contain a single directory per assembly, with the unique part of each assembly's path, with / replaced by _.

So for example, if the user passed in:

transrate --assembly /assemblies/rice/oases/contigs.fa,/assemblies/maize/trinity.fa

the following structure would be created:

transrate_results
  - rice_oases
    - contigs.csv
    - ...etc.
  - maize_trinity
    - contigs.csv
    - ...etc.
rozovr commented 9 years ago

one thing that was odd to me was that the '-o' input flag applies only to CSV files and not all the outputs produced

blahah commented 9 years ago

noted - I think we will eliminate the -o argument once this new structure is used, as it will not really be needed (it's easy to just rename the transrate_results if necessary)

rozovr commented 9 years ago

Why not use it to specify the output path (or prefix) for everything? You can keep the working directory as the default

blahah commented 9 years ago

that seems a reasonable idea

blahah commented 9 years ago

@rozovr thanks, I implemented your idea and it will be in the next release