Untree inverts the action of tree by converting tree diagrams of directory structures back into directory structures. Given a directory structure, tree produces a tree diagram, and given a tree diagram, untree produces a directory structure.
Let's say you have the following directory structure, created by running tree
in the root of this project:
. ├── Cargo.lock ├── Cargo.toml ├── inputs │ └── test1.tree ├── lib │ ├── either.rs │ ├── errors.rs │ ├── functions.rs │ ├── mod.rs │ ├── more_context.rs │ ├── path_action.rs │ └── types.rs ├── LICENSE.txt ├── media │ ├── image1.png │ └── image2.png ├── README.md └── src └── main.rs
untree can create a mirror that directory structure, just based on that input:
tree | untree --dir path/to/output/dir
Here, test
is the destination directory where untree
is supposed to create
files. Now, if we tree
the newly created directory, we can see that it has the
same structure as the repository:
path/to/output/dir ├── Cargo.lock ├── Cargo.toml ├── inputs │ └── test1.tree ├── lib │ ├── either.rs │ ├── errors.rs │ ├── functions.rs │ ├── mod.rs │ ├── more_context.rs │ ├── path_action.rs │ └── types.rs ├── LICENSE.txt ├── media │ ├── image1.png │ └── image2.png ├── README.md └── src └── main.rs 4 directories, 15 files
untree
can also read in the tree from an input file, or you can paste it in
directly since it accepts input from standard input:
I've noticed that in the past I've had to recreate directory structures in order to answer questions or run tests on the directory. For example, this question asks about ignoring certain kinds of files, and it provides a directory structure as reference.
The files themselves aren't provided, nor do they need to be, but the directory structure itself is relevant to the question.
untree
allows you to replicate the structure of a directory printed with tree,
making it easy to answer questions about programs that traverse the directory
tree. This means that untree is also good for quickly creating directory
structures for the purpose of mocking input to other programs.
You can use untree as a library if you need that functionality included in your
program. In order to create a tree, invoke [create_tree
] with the given
directory, Lines
buffer, and options.
These options are very simple - there's [UntreeOptions::verbose
], which will
tell [create_tree
] and [create_path
] to print out any directories or files
that were created when set, and [UntreeOptions::dry_run
], which will print out
any directories or files without actually creating them (dry_run
implies
verbose
).
Below is an example usage:
use untree::*;
use std::io::{BufRead, BufReader, stdin, Lines};
let options = UntreeOptions::new()
.dry_run(true) // Set dry_run to true
.verbose(true); // Set verbose to true
let lines = BufReader::new(stdin()).lines();
create_tree("path/to/directory", lines, options)?;
# Ok::<(), Error>(())
Additional functions include
- [
create_path
], used to create a file or path with the given options, - [
get_entry
], used to parse a line in a tree file, - [
touch_directory
], used to create a directory, - [
touch_file
], used to touch a file (does the same thing as unix touch)
The primary error type used by untree is [Error
], which holds information
about a path and the action being done on it, in addition to the normal error
information provided by io::Error
.
When asked about untree, my friend said:
I retroactively want that for my time trying to get Conan to work. It woulda made certain things just a little less painful.
— some guy (He asked to be referred to as "some guy")
I'm in the progress of learning rust, so any feedback you have is greatly
appreciated! Also, if untree
is useful to you, please let me know!