About pyprojroot2

Want to determine where in the file system your project's base (or root) directory is?

pyprojroot2 provides the solution, and makes it easy to locate data in a project.

This project is a complete rewrite of Daniel Chen's pyprojroot inspired by the issues in the tracker and the TODOs in the code.

tl;dr

Sick of fiddling around with the current directory and searching for the correct path to your data in your git project?

from pyprojroot2 import is_git_root
data_file = is_git_root.find_file("data/your_file.csv")

searches upwards from your current directory for the git root and returns the absolute path name to your data file.

Many other popular project root criteria are available at pyprojroot2.predefined_roots.

Popular Choice: here("data/your_file.csv") is a handy shortcut to locate your data by detecting the root of a collection of python projects.

Install

Try it out:

pip3 install git+https://github.com/achimgaedke/pyprojroot2.git

It works as a drop-in replacement for pyprojroot:

import pyprojroot2 as pyprojroot

The project runs with python-3.7 and upwards and no other dependencies required.

Use here

here is straightforward to use, simply find your files relative to the project's root (or base) directory:

from pyprojroot2 import here

data_path = here("data")

pd.read_csv(data_path / "dataset.csv")

The project is detected with the help of a (pythonic) set of criteria starting from the current path, upwards in the filesystem.

Want the pathname as a string?

data_path = str(here("data"))

If the current directory is somewhere outside the project you want to detect, you can specify the start path for the search:

data_path = here("data", path="/humungous/partiton/some/notebooks/")

If the root could not be determined, the exception FileNotFoundError is raised.

In that case, you can add a .here file into the corresponding directory, giving here() a hint or define your root criteria, as described in the next section.

Want to know what here() chose as root and why?

from pyprojroot2 import get_root_desc, py_here_criteria

get_root_desc(py_here_criteria, here())

Under the hood here() uses the py_here_criteria predefined criterion, bundling a popular selection of project root criteria.

More functions comparable to the R packages rprojroot and here are available from the pyprojroot2 submodules rprojroot and rhere.

Define your root criteria

If you know that your project root will have the (only) data directory in your project, then you can specify:

has_dir("data")

as a root criterion.

You can combine root criteria using the operators & (logical "and") and | (logical "or"). Unsurprisingly, the predefined is_vcs_root criterion is:

is_vcs_root = is_git_root | is_svn_root

from pyprojroot2 import is_git_root, has_dir, is_here

is_my_root = is_git_root | is_here
my_data = is_my_root.find_root_file("data")

This root criterion will first look for a git root in all parent directories, if that isn't found anywhere, then move on to finding a directory with a data directory and finally if data couldn't be found, look out for .here.

If none of those could be located, a FileNotFoundError exception is raised.

If you know that the data directory is in the git root directory, you could write:

is_my_root = (
    is_git_root & has_dir("data") |
    is_here
)

and also keep the .here marker file as a fallback option.

The usual operator precedence applies, & binds stronger than |.

Root finding outside python

The command pyprojroot2 will allow locating the root of a project on the shell. So you can be sure, the result is consistent with the python scripting.

This command prints out the project root directory searched from the current directory using the default criteria.

So, "going back to the root" is simple as this

cd "$(pyprojroot2)"

If no project root is found, it will fail with an error message.

Optionally you can set one of the predefined criteria with -c or specify the start path.