tdurieux
Download Repository does not apply terms to anonymize, unlike the web interface
Thanks for creating this tool and supporting open science!
When downloading my repository with the "Download Repository" button, some of the instances of the terms identified in "terms to anonymize" are not properly anonymized in the downloaded copy. Which instances fail to anonymize appears somewhat random.
I have this issue as well. I was hoping there would be a way to turn off downloads entirely for a repo in the options.
I dont see how this is possible because it uses the same pipeline. Could you provide me a repo for me to try?
The repo I'd hand you is currently under double-blind review for a publication. I had to go through it and scrub sensitive terms manually. When the review process is finished, I will try and re-create the issue I was having, and provide you with the repo, but no promises. I do not know if the error will occur again.
Here's a repro: I anonymized eringrant/test-anonymous-github at https://anonymous.4open.science/r/test-anonymous-github-B58A. The anonymized terms are:
mercury # XXXX-1
venus # XXXX-2
mars # XXXX-3
jupiter # XXXX-4
saturn # XXXX-5
uranus # XXXX-6
neptune # XXXX-7
pluto # XXXX-8
haumea # XXXX-9
makemake # XXXX-10
eris # XXXX-11
"Download Repository," unzip, and look for non-anonymized terms:
$ grep "mercury\|venus\|mars\|jupiter\|saturn\|uranus\|neptune\|pluto\|haumea\|makemake\|eris" src/*
src/file30.txt:mercury
src/file4.txt:mars
src/file41.txt:saturn
src/file43.txt:mars
src/file50.txt:makemake
These terms are anonymized in the web interface. Downloading the repo again and looking for non-anonymized terms gives a different set of instances:
$ grep "mercury\|venus\|mars\|jupiter\|saturn\|uranus\|neptune\|pluto\|haumea\|makemake\|eris" src/*
src/file11.txt:mercury
src/file13.txt:pluto
src/file19.txt:makemake
src/file24.txt:neptune
src/file26.txt:mercury
src/file28.txt:pluto
src/file31.txt:mercury
src/file49.txt:makemake
src/file58.txt:makemake