Blind files

Generates a mapping from file names to blind but memorable file names. This script assumes that you have a directory that contains files and / or subdirectories with samples from an experiment. The names of these files and directories reveal which group the samples belong to, but the contents of the files do not.

The script will move these files to a new directory, renaming them so that the new file names do not reveal which group the samples belong to. It will also generate a mapping file to indicate how the new files map to the original files.

Installing

Using pipx

pipx install blind_files

Running

This script takes an input dir, and generates a directory containing a script, blind.sh, that can be used to blind the files in the input dir. It also generates a mapping csv, mapping.csv, that can be used after the user has done the analysis to see how the original names map to blinded names.

The script has two modes of operation:

Using a delimiter

In the first mode of operation, you can specify a delimiter to use such that all the text before the delimiter in each file name will be replaced. For example:

blind-files \
   --mode delimiter \
   --delimiter _foo \
   --input-dir input_dir \
   --output-dir output_dir \
   --mapping-dir mapping_dir

In this case, if input_dir contains the following files:

sample_1_foo.txt
sample_1_foo-bar.csv
sample_2_foo.txt
hello.txt

Then after running mapping_dir/blind.sh, output_dir will contain

golf_elbow_foo.txt
golf_elbow_foo-bar.csv
co-producer_reputation_foo.txt
hello.txt

In mapping_dir you will also find a file mapping.csv with the contents:

original,blinded
sample_1,golf_elbow
sample_2,co-producer_reputation

Limitations

This will only replace names at the top level of the input directory. If you have a more complex nested directory structure, where the identifer names may be buried in the directory tree, use identifier list approach described below.

Using a list of identifiers

In the second mode of operation, you can specify list of identifiers that should be blinded whenever they are encountered in the input directory tree. For example, if identifiers.txt contains the following:

group_a_1
group_b_1

then running

blind-files \
   --mode identifiers \
   --identifiers identifiers.txt \
   --input-dir input_dir \
   --output-dir output_dir \
   --mapping-dir mapping_dir

In this case, if input_dir contains the following files:

group_a_1/group_a_1/foo.txt
group_b_1/group_b_1/foo.txt
hello.txt

Then after running mapping_dir/blind.sh, output_dir will contain

head_bottle/head_bottle/foo.txt
eponym_curtain/eponym_curtain/foo.txt
hello.txt

In mapping_dir you will also find a file mapping.csv with the contents:

original,blinded
group_a_1,head_bottle
group_b_1,eponym_curtain

Limitations

No identifier can be a substring of any other identifier. For example, it is not allowed to have identifiers sample_1 and sample_11. However, sample_01 and sample_11 would be fine.

General limitations

• This script should work on any platform, but has only been tested on Mac OS.
• This script should handle symlinks by simply moving the symlink, without following it, but this behavior has not been tested.

Credits

This package was created with Cookiecutter.

nounlist from here.