Parametric file paths to access and build structured folder trees.
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.github
hansel
.editorconfig
.envrc
.gitignore
.scrutinizer.yml
.travis.yml
CHANGES.rst
LICENCE.txt
MANIFEST.in
Makefile
Pipfile
README.rst
conftest.py
setup.cfg
setup.py

README.rst

hansel

Parametric file paths to access and build structured folder trees.

PyPI Build Status Coverage Status Code Health

It almost doesn't have Dependencies, check how to Install it.

Github repository: https://github.com/alexsavio/hansel

Usage

Quick Intro

Imagine this folder tree:

data
└── raw
    ├── 0040000
    │   └── session_1
    │       ├── anat_1
    │       └── rest_1
    ├── 0040001
    │   └── session_1
    │       ├── anat_1
    │       └── rest_1
    ├── 0040002
    │   └── session_1
    │       ├── anat_1
    │       └── rest_1
    ├── 0040003
    │   └── session_1
    │       ├── anat_1
    │       └── rest_1
    ├── 0040004
    │   └── session_1
    │       ├── anat_1
    │       └── rest_1
>>> from pprint import pprint
>>> from hansel import Crumb

# create the crumb
>>> crumb = Crumb("{base_dir}/raw/{subject_id}/{session_id}/{image_type}/{image}")

# set the base_dir path
>>> crumb = crumb.replace(base_dir='/tmp/hansel/data')
>>> print(str(crumb))
/tmp/hansel/data/raw/{subject_id}/{session_id}/{image_type}/{image}

# get the ids of the subjects
>>> subj_ids = crumb['subject_id']
>>> print(subj_ids)
['0040000', '0040001', '0040002', '0040003', '0040004', '0040005', ...

# get the paths to the subject folders, the output can be strings or crumbs,
# you choose with the ``make_crumbs`` boolean argument. Default: True.
>>> subj_paths = crumb.ls('subject_id', make_crumbs=True)
>>> pprint(subj_paths)
[Crumb("/tmp/hansel/data/raw/0040000/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040001/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040002/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040003/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040004/{session_id}/{image_type}/{image}"),
 ...

# set the image_type
>>> anat_crumb = crumb.replace(image_type='anat_1')
>>> print(anat_crumb)
/tmp/hansel/data/raw/{subject_id}/{session_id}/anat_1/{image}

# get the paths to the images inside the anat_1 folders
>>> anat_paths = anat_crumb.ls('image')
>>> pprint(anat_paths)
[Crumb("/tmp/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040003/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040004/session_1/anat_1/mprage.nii.gz"),
 ...

# get the ``session_id`` of each of these ``anat_paths``
>>> sessions = [cr['session_id'][0] for cr in anat_paths]
>>> print(sessions)
['session_1', 'session_1', 'session_1', 'session_1', 'session_1', ...

# if you don't want the the output to be ``Crumbs`` but string paths:
>>> anat_paths = anat_crumb.ls('image', make_crumbs=False)
>>> pprint(anat_paths)
['/tmp/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz',
 '/tmp/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz',
 '/tmp/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz',
 '/tmp/hansel/data/raw/0040003/session_1/anat_1/mprage.nii.gz',
 '/tmp/hansel/data/raw/0040004/session_1/anat_1/mprage.nii.gz',
 ...

# you can also use a list of ``fnmatch`` expressions to ignore certain files patterns
# using the ``ignore_list`` argument in the constructor.
# For example, the files that start with '.'.
>>> crumb = Crumb("{base_dir}/data/raw/{subject_id}/{session_id}/{image_type}/{image}", ignore_list=['.*'])

Once you have a fully defined Crumb, you can use its path for operations with the corresponding file. For that you have to convert it to string by using str(crumb) or crumb.path.

See more quick examples after the Long Intro check More features and tricks.


Long Intro

I often find myself in a work related with structured folder paths, such as the one shown above.

I have tried many ways of solving these situations: loops, dictionaries, configuration files, etc. I always end up doing a different thing for the same problem over and over again.

This week I grew tired of it and decided to make a representation of a structured folder tree in a string and access it the most easy way.

If you look at the folder structure above I have:

  • the root directory from where it is hanging: ...data/raw,
  • many identifiers (in this case a subject identification), e.g., 0040000,
  • session identification, session_1 and
  • a data type (in this case an image type), anat_1 and rest_1.

With hansel I can represent this folder structure like this:

>>> from hansel import Crumb
>>> crumb = Crumb("{base_dir}/data/raw/{subject_id}/{session_id}/{image_type}/{image}")

Let's say we have the structure above hanging from a base directory like /home/hansel/.

I can use the replace function to make set the base_dir parameter:

>>> crumb = crumb.replace(base_dir='/home/hansel')
>>> print(str(crumb))
/home/hansel/data/raw/{subject_id}/{session_id}/{image_type}/{image}

if I don't need a copy of crumb, I can use the [] operator:

>>> crumb['base_dir'] = '/tmp/hansel'
>>> print(str(crumb))
/tmp/hansel/data/raw/{subject_id}/{session_id}/{image_type}/{image}

Now that the root path of my dataset is set, I can start querying my crumb path.

If I want to know the path to the existing subject_id folders:

We can use the ls function. Its output can be str or Crumb. I can choose this using the make_crumbs argument (default: True):

>>> subj_crumbs = crumb.ls('subject_id')
>>> pprint(subj_crumbs)
[Crumb("/tmp/hansel/data/raw/0040000/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040001/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040002/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040003/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/0040004/{session_id}/{image_type}/{image}"),
 ...

>>> subj_paths = crumb.ls('subject_id', make_crumbs=False)
>>> pprint(subj_paths)
['/tmp/hansel/data/raw/0040000/{session_id}/{image_type}/{image}',
 '/tmp/hansel/data/raw/0040001/{session_id}/{image_type}/{image}',
 '/tmp/hansel/data/raw/0040002/{session_id}/{image_type}/{image}',
 '/tmp/hansel/data/raw/0040003/{session_id}/{image_type}/{image}',
 '/tmp/hansel/data/raw/0040004/{session_id}/{image_type}/{image}',
 ...

If I want to know what are the existing subject_id:

>>> subj_ids = crumb.ls('subject_id', fullpath=False)
>>> print(subj_ids)
['0040000', '0040001', '0040002', '0040003', '0040004', '0040005', ...

or

>>> subj_ids = crumb['subject_id']
>>> print(subj_ids)
['0040000', '0040001', '0040002', '0040003', '0040004', '0040005', ...

Now, if I wanted to get the path to all the images inside the anat_1 folders, I could do this:

>>> anat_crumb = crumb.replace(image_type='anat_1')
>>> print(anat_crumb)
/tmp/hansel/data/raw/{subject_id}/{session_id}/anat_1/{image}

or if I don't need to keep a copy of crumb:

>>> crumb['image_type'] = 'anat_1'

# get the paths to the images inside the anat_1 folders
>>> anat_paths = crumb.ls('image')
>>> pprint(anat_paths)
[Crumb("/tmp/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040003/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040004/session_1/anat_1/mprage.nii.gz"),
 ...

Remember that I can still access the replaced crumb arguments in each of the previous crumbs in anat_paths.

>>> subj_ids = [cr['subject_id'][0] for cr in anat_paths]
>>> print(subj_ids)
['0040000', '0040001', '0040002', '0040003', '0040004', ...

>>> files = [cr['image'][0] for cr in anat_paths]
>>> print(files)
['mprage.nii.gz', 'mprage.nii.gz', 'mprage.nii.gz', 'mprage.nii.gz', ...

More features and tricks

There are more possibilities such as:

Creating folder trees

Use mktree and ParameterGrid to create a tree of folders.

>>> from hansel import mktree
>>> from hansel.utils import ParameterGrid

>>> crumb = Crumb("/tmp/hansel/data/raw/{subject_id}/{session_id}/{image_type}/{image}")

>>> session_ids = ["session_{}".format(i) for i in range(2)]
>>> subject_ids = ["subj_{}".format(i) for i in range(3)]

>>> values_map = dict(session_id=session_ids, subject_id=subject_ids)

>>> crumbs = mktree(crumb, list(ParameterGrid(values_map)))
>>> pprint(crumbs)
[Crumb("/tmp/hansel/data/raw/subj_0/session_0/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/subj_1/session_0/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/subj_2/session_0/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/subj_0/session_1/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/subj_1/session_1/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/raw/subj_2/session_1/{image_type}/{image}")]

Check the feasibility of a crumb path

>>> crumb = Crumb("/tmp/hansel/raw/{subject_id}/{session_id}/{image_type}/{image}")

# ask if there is any subject with the image 'lollipop.png'.
>>> crumb['image'] = 'lollipop.png'
>>> assert not crumb.exists()

Check which subjects have 'jujube.png' and 'toffee.png' files

>>> crumb = Crumb("/tmp/hansel/raw/{subject_id}/{session_id}/{image_type}/{image}")

>>> toffee_crumb = crumb.replace(image='toffee.png')
>>> jujube_crumb = crumb.replace(image='jujube.png')

# using sets functionality
>>> gluttons = set(toffee_crumb['subject_id']).intersection(set(jujube_crumb['subject_id']) # doctest: +SKIP
>>> print(gluttons)  # doctest: +SKIP
['gretel', 'hansel']

Use the intersection function

Use it for comparisons on more than one crumb argument. This can be used to compare datasets with the same structure in different folders.

One argument

Imagine that we have two working folders of subjects for two different projects: proj1 and proj2. If I want to check what subjects are common to both projects:

>>> from hansel import intersection

# using one argument
>>> cr_proj1 = Crumb("/tmp/hansel/data/proj1/{subject_id}/{session_id}/{image_type}/{image}")
>>> cr_proj2 = Crumb("/tmp/hansel/data/proj2/{subject_id}/{session_id}/{image_type}/{image}")

# set the `on` argument in `intersection` to specify which crumb arguments to merge.
>>> merged = intersection(cr_proj1, cr_proj2, on=['subject_id'])
>>> pprint(merged)
[(('subject_id', '0040006'),),
 (('subject_id', '0040007'),),
 (('subject_id', '0040008'),),
 (('subject_id', '0040009'),)]

# I can pick these subject crumbs from this result using the `build_paths` function.
>>> proj1_merged_paths = cr_proj1.build_paths(merged, make_crumbs=True)
>>> type(proj1_merged_paths)
<class 'generator'>

>>> pprint(list(proj1_merged_paths))
[Crumb("/tmp/hansel/data/proj1/0040006/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/proj1/0040007/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/proj1/0040008/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/proj1/0040009/{session_id}/{image_type}/{image}")]

>>> pprint(list(cr_proj2.build_paths(merged, make_crumbs=True)))
[Crumb("/tmp/hansel/data/proj2/0040006/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/proj2/0040007/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/proj2/0040008/{session_id}/{image_type}/{image}"),
 Crumb("/tmp/hansel/data/proj2/0040009/{session_id}/{image_type}/{image}")]

Two arguments

Now, imagine that I have different sets of {image} for these subjects. I want to check which of those subjects have exactly the same images. Let's say that the subject 0040000 has a anatomical.nii.gz instead of mprage.nii.gz.

>>> from hansel import intersection

# using one argument
>>> cr_proj3 = Crumb("/tmp/hansel/data/proj3/{subject_id}/{session_id}/anat_1/{image}")
>>> cr_proj4 = Crumb("/tmp/hansel/data/proj4/{subject_id}/{session_id}/anat_1/{image}")

# set the `on` argument in `intersection` to specify which crumb arguments to merge.
>>> merged = intersection(cr_proj3, cr_proj4, on=['subject_id', 'image'])
>>> pprint(merged)
[(('subject_id', '0040001'), ('image', 'mprage.nii.gz')),
 (('subject_id', '0040002'), ('image', 'mprage.nii.gz')),
 (('subject_id', '0040003'), ('image', 'mprage.nii.gz')),
 (('subject_id', '0040004'), ('image', 'mprage.nii.gz')),
...

# I can pick these image crumbs from this result using the `build_paths` function.
>>> pprint(list(cr_proj3.build_paths(merged, make_crumbs=True)))
[Crumb("/tmp/hansel/data/proj3/0040001/{session_id}/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/proj3/0040002/{session_id}/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/proj3/0040003/{session_id}/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/proj3/0040004/{session_id}/anat_1/mprage.nii.gz"),
 ...

>>> pprint(list(cr_proj4.build_paths(merged, make_crumbs=True)))
[Crumb("/tmp/hansel/data/proj4/0040001/{session_id}/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/proj4/0040002/{session_id}/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/proj4/0040003/{session_id}/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/proj4/0040004/{session_id}/anat_1/mprage.nii.gz"),
...

# adding 'mod' to the intersection would be:
>>> common_values = intersection(cr_proj3, cr_proj4, on=['subject_id', 'session_id', 'image'])
>>> pprint(common_values, width=120)
[(('subject_id', '0040001'), ('session_id', 'session_1'), ('image', 'mprage.nii.gz')),
 (('subject_id', '0040002'), ('session_id', 'session_1'), ('image', 'mprage.nii.gz')),
 (('subject_id', '0040003'), ('session_id', 'session_1'), ('image', 'mprage.nii.gz')),
 (('subject_id', '0040004'), ('session_id', 'session_1'), ('image', 'mprage.nii.gz')),
 ...

The unfold and ls functions

Unfold the whole crumb path to get the whole file tree in a list of paths:

>>> all_images = Crumb("/tmp/hansel/data/raw/{subject_id}/{session_id}/{image_type}/{image}")
>>> all_images = all_images.unfold()
>>> pprint(all_images)
[Crumb("/tmp/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040000/session_1/rest_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/rest_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040002/session_1/rest_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040003/session_1/anat_1/mprage.nii.gz"),
...

# and you can ask for the value of the crumb argument in each element
>>> print(all_images[0]['subject_id'])
['0040000']

Note that unfold is the same as calling ls function without arguments.

Use regular expressions

Use re.match or fnmatch expressions to filter the paths:

The syntax for crumb arguments with a regular expression is: "{<arg_name>:<arg_regex>}"

# only the session_1 folders
>>> session1_cr = Crumb("/tmp/hansel/data/raw/{subject_id}/{session_id:*_1}/{image_type}/{image}")
>>> session1_imgs = session1_cr.ls()
>>> pprint(session1_imgs)
[Crumb("/tmp/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040000/session_1/rest_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/rest_1/mprage.nii.gz"),
...

The default is for fnmatch expressions. If you prefer using re.match for filtering, set the regex argument to 're' or 're.ignorecase' in the constructor.

# only the rest images from the subject ``040000``
>>> s0_rest_cr = Crumb("/tmp/hansel/data/raw/{subject_id:.*00$}/{session_id}/{image_type:rest.*}/{image}", regex='re')
>>> s0_rest_imgs = s0_rest_cr.ls()
>>> print(s0_rest_imgs)
[Crumb("/tmp/hansel/data/raw/0040000/session_1/rest_1/mprage.nii.gz")]

The regular expressions can be checked with the patterns property.

>>> pprint(s0_rest_cr.patterns)
{'image_type': 'rest.*', 'subject_id': '.*00$'}

And can be also modified with the set_pattern function.

>>> s0_rest_cr.set_pattern('session_id', '.*_1$')
>>> pprint(s0_rest_cr.patterns)
{'image_type': 'rest.*', 'session_id': '.*_1$', 'subject_id': '.*00$'}
>>> s0_rest_cr.path
'/tmp/hansel/data/raw/{subject_id:.*00$}/{session_id:.*_1$}/{image_type:rest.*}/{image}'

A regular expression can be temporarily set with the ls function and the [] operator.

>>> crumb = Crumb("/tmp/hansel/data/raw/{subject_id}/{session_id}/{image_type}/{image}")
>>> mprage_crumb = crumb.ls('image:mprage.*')
>>> pprint(mprage_crumb)
[Crumb("/tmp/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040000/session_1/rest_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040001/session_1/rest_1/mprage.nii.gz"),
 Crumb("/tmp/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz"),
...

>>> pprint(crumb['image:mprage.*'])
['mprage.nii.gz',
 'mprage.nii.gz',
 'mprage.nii.gz',
 'mprage.nii.gz',
 'mprage.nii.gz',
 'mprage.nii.gz',
...

Copy and modify folder structure with crumb_copy

Copy a folder structure from one crumb to the other. The source crumb must be fully specified, i.e., all crumb arguments must get an existing value. In addition the destination crumb can only have a subset of the crumb arguments of the source crumb.

>>> from hansel import Crumb, crumb_copy
>>> src_cr = Crumb("/tmp/hansel/data/raw/{subj_id}/{sess}/{type}/{img}")
>>> dst_cr = Crumb("/tmp/hansel/data/copy/{subj_id}/{sess}/{type}")
>>> crumb_copy(src_cr, dst_cr)

More functionalities, ideas and comments are welcome.

Command Line

hansel will install a command called crumb. This CLI has been made with Click, so try crumb -h to see more details.

You can use Crumb.ls:

crumb ls "/data/hansel/cobre/{sid:4*100}/{session}/{img}"

Copy one file tree to another file tree with crumb copy:

crumb copy "/data/hansel/cobre/{sid}/{session}/{img}" "/data/hansel/cobre2/{sid}/{img}"

Link one file tree to another file tree with link:

crumb link "/data/hansel/cobre/{sid}/{session}/{img}" "/data/hansel/cobre2/{sid}/{img}"

Return the intersection between crumb1 and crumb2 on a given argument with the intersect function:

crumb intersect --on "sid" "/data/hansel/cobre/{sid}/{session}/{img}" "/data/hansel/cobre2/{sid}/{img}"

Return the difference crumb1 - crumb2 on a given argument with the diff function:

crumb diff --on "sid" "/data/hansel/cobre/{sid}/{session}/{img}" "/data/hansel/cobre2/{sid}/{img}"

Dependencies

Please see the requirements.txt file. Before installing this package, install its dependencies with:

pip install -r requirements.txt

Install

It works on Python 3.4, 3.5 and 2.7. For Python 2.7 install pathlib2 as well.

This package uses setuptools. You can install it running:

python setup.py install

If you already have the dependencies listed in requirements.txt installed, to install in your home directory, use:

python setup.py install --user

To install for all users on Unix/Linux:

python setup.py build
sudo python setup.py install

You can also install it in development mode with:

python setup.py develop

Development

Code

Github

You can check the latest sources with the command:

git clone https://www.github.com/alexsavio/hansel.git

or if you have write privileges:

git clone git@github.com:alexsavio/hansel.git

If you are going to create patches for this project, create a branch for it from the master branch.

We tag stable releases in the repository with the version number.

Testing

We are using py.test to help us with the testing.

Otherwise you can run the tests executing:

python setup.py test

or

py.test

or

make test