Advanced Anonymizer

Create working anonymous project copies.

About

For double blinded conferences you often want to provide support material, that allows replication of your claims, but does not expose your identity. Projects like Anonymous 4 Open Science clear a project from a provided list of identifiers, but produce no actual git copies that could be afterwards cloned and executed.
This little tool is a standalone, offline script to support full replication. It likewise substitutes a list of provided keywords, but you can select the substitute, and you obtain the resulting sources, not just a browsable website reflecting the content.
Likewise, this anonymizer treats file and directory names, and applies the same subsititions. This way you can esure your code is still executable, even after substitution, e.g. if you anonymize package names.

Usage

Create a file substitutions.txt, next to the anonymize.sh script
Provide binary keyword substitute tuples, alphanumeric and separated by a single whitespace.
Words may begin with a . character. Example:

`Maximilian anonymousresearcher`
`mcgill someuniversity`
`.ca .country`
...

Note: These are NOT regex substition rules.

Call the anonymizer, provide exatly one argument: the location of your folder to anonymize.
Example:

./anonymize ~/Code/MySuperSecretProject

How it works

Your original sources are not mofied. The script creates a replacement folder, with prefix Anonymized...
The anonymizer removes all traces of any dot files (.git, .DS_Store, ...), to prevent identification via metadata details, e.g. git commits.
The anonymizer seraches for all provided keywords, Case Sensitive and replaces all occurrences by the specified subsititue. Note that all occurrences will be substituted, so if your key is short, you may accidentally replace substrings of other words.
Substitution considers:
- File content
- File/Directory names
The anonymizer searches once more for all provided keywords, Case Insensitive and prints a warning for each remaining occurrence.

Known Issues

Overly short key strings may substitute more than you want, as the replacement also triggers on substring. E.g. of you provide a rule a x then all as in your text are replaced.
Umlauts in substitution rules are not supported.

Author