ilp <subcommand> [<args>]

ilp -i

ilp is a command-line utility which indexes files, and organizes them with tags. It will not touch the files on disk, but instead keep track of them in a database, much like the mlocate utility found on many UNIX or UNIX-like systems. To uniquely identify a file, ilp stores it as a hash of its contents.

<subcommand>

All functionality of ilp is implemented in subcommands, similar to git or other VCS. See SUBCOMMANDS for an overview, or use the command "help".

-i

Start an interactive shell, where you can issue subcommands directly.

Much like a VCS such as git, ilp does all the work with various subcommands. You can either issue them as an argument to the command line utility, or start an interactive shell with -i and issue them directly.

ilp index [-r] <path> [<path>…​]

Add one or more files to the index. The -r switch turns on recursive indexing, in which case <path> can be either a file or directory, and will add all files below a given directory to the index.

ilp forget [-r] <path> [<path>…​]

Counterpart to the index subcommand. Removes files from the index. Correspondingly, -r turns on recursive removing.

ilp tag [-r] <path> <tag> [<tag>…​]

Tag a file with one or more tags. Similar to index and forget, -r will apply the tags recursively to all files below a given directory.

ilp untag [-r] <path> <tag> [<tag>…​]

Counterpart to tag. Remove tags from a file. -r works the same way.

ilp deltag <tag> [<tag>…​]

Delete tags from the database. Note that tags are automatically created when first used, so this command has no counterpart.

ilp list <what>

List items on the index. <what> can be files or tags.

ilp info [-t] <what>

Get info on files or tags. Since filenames and tags can’t meaningfully be differentiated, -t (--tag) needs to be passed if information on a tag is wanted.

ilp search <query>

Search files on the index. See SEARCH QUERIES for how a query works.

ilp clear [-y]

Purge the database. This won’t do anything unless -y (--yes) is passed.

ilp uses a simple DSL for search queries, which is based on basic set operations. It requires a bit of understanding about how tagging in ilp works. If you just want a quick start (it might be all you need, it’s rather intuitive already), see EXAMPLE QUERIES.

Basically, every tag is regarded as a set of all files it is tagging. That might sound counterintuitive at first (because you would probably expect the relation to be the other way round), but makes searching through tagged files incredibly easy, yet powerful - because every operation on a set results in a new set, on which more operations can be applied. Think of it in terms of Venn diagrams.

For example, if you want to have all files which are tagged both "foo" and "bar", you’d simply do a logical and on the two sets, and the resulting set is your wanted result.

This works analogously (and intuitively!) with logical or and not. Even exclusive or (xor) works, and is quite useful, too - you can use it to find all files which are either tagged "foo" or "bar", but not both.

Here’s a simple overview of the language in Pseudo-BNFL:

EXAMPLE QUERIES

foo and bar

foo not bar

foo xor bar not baz spam

foo xor bar not baz not spam

If you have wondered what ilp stands for, its an acronym for Index Librorum Prohibitorum - the infamous (and since 1966 defunct) list of books prohibited by the Vatican. It’s really just the first thing that came to mind in relation to indexing files, and "ilp" sounded catchy and is short to type. Besides, omnia dicta fortiora si dicta Latina ("everything sounds more impressive when said in Latin").