hansel 0.7.3

hansel
======

Parametric file paths to access and build structured folder trees.

|PyPI| |Build Status| |Coverage Status| |PyPI Downloads| |Code Health| |Scrutinizer|

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

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


.. code:: python

>>> from hansel import Crumb

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

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

# 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)
>>> print(subj_paths)
[Crumb("/home/hansel/data/raw/0040000/{session_id}/{image_type}/{image}"),
Crumb("/home/hansel/data/raw/0040001/{session_id}/{image_type}/{image}"),
Crumb("/home/hansel/data/raw/0040002/{session_id}/{image_type}/{image}"),
Crumb("/home/hansel/data/raw/0040003/{session_id}/{image_type}/{image}"),
Crumb("/home/hansel/data/raw/0040004/{session_id}/{image_type}/{image}"),
...

# set the image_type
>>> anat_crumb = crumb.replace(image_type='anat_1')
>>> print(anat_crumb)
/home/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')
>>> print(anat_paths)
[Crumb("/home/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040003/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/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)
>>> print(anat_paths)
["/home/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz",
"/home/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz",
"/home/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz",
"/home/hansel/data/raw/0040003/session_1/anat_1/mprage.nii.gz",
"/home/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=['.*'])

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:

.. code:: python

>>> 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:

.. code:: python

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

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

.. code:: python

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

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):

.. code:: python

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

>>> subj_paths = anat_crumb.ls('subject_id', make_crumbs=False)
>>> print(subj_paths)
["/home/hansel/data/raw/0040000/{session_id}/{image_type}/{image}",
"/home/hansel/data/raw/0040001/{session_id}/{image_type}/{image}",
"/home/hansel/data/raw/0040002/{session_id}/{image_type}/{image}",
"/home/hansel/data/raw/0040003/{session_id}/{image_type}/{image}",
"/home/hansel/data/raw/0040004/{session_id}/{image_type}/{image}",
...


If I want to know what are the existing ``subject_id``:

.. code:: python

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

or

.. code:: python

>>> 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:

.. code:: python

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

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

.. code:: python

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

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

.. code:: python

>>> subj_ids = [cr['subject_id'][0] for cr in anat_paths]
>>> print(subj_ids)
['0040000', '0040001', '0040002', '0040003', '0040004', '0040005', ...

>>> 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.

.. code:: python

>>> from hansel import mktree, ParameterGrid

>>> crumb = Crumb("/home/hansel/raw/{subject_id}/{session_id}/{modality}/{image}")

>>> values_map = {'session_id': ['session_' + str(i) for i in range(2)],
>>> 'subject_id': ['subj_' + str(i) for i in range(3)]}

>>> mktree(crumb, list(ParameterGrid(values_map)))


Check the feasibility of a crumb path
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. code:: python

>>> crumb = Crumb("/home/hansel/raw/{subject_id}/{session_id}/{modality}/{image}")

# ask if there is any subject with the image 'lollipop.png'.
>>> crumb['image'] = 'lollipop.png'
>>> assert crumb.exists()


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

.. code:: python

>>> crumb = Crumb("/home/hansel/raw/{subject_id}/{session_id}/{modality}/{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'])
>>> print(gluttons)
['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:

.. code:: python

>>> from hansel import intersection

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

# set the `on` argument in `intersection` to specify which crumb arguments to merge.
>>> merged = intersection(cr_proj1, cr_proj2, on=['subject_id'])
>>> print(merged)
[(('subject_id', '0040000'),), (('subject_id', '0040001'),), (('subject_id', '0040001'),)]

# I can pick these subject crumbs from this result using the `build_paths` function.
>>> cr1.build_paths(merged, make_crumbs=True)
[Crumb("/home/hansel/proj1/0040010/{session}/{mod}/{image}"),
Crumb("/home/hansel/proj1/0040110/{session}/{mod}/{image}")]

>>> cr2.build_paths(merged, make_crumbs=True)
[Crumb("/home/hansel/proj2/0040010/{session}/{mod}/{image}"),
Crumb("/home/hansel/proj2/0040110/{session}/{mod}/{image}")]


Two arguments
`````````````

Now, imagine that I have different sets of `{image}` for these subjects.
I want to check what of those subjects have exactly the same images.
Let's say that the subject `0040001` has a `anatomical.nii.gz` instead of `mprage.nii.gz`.

.. code:: python

>>> from hansel import intersection

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

# set the `on` argument in `intersection` to specify which crumb arguments to merge.
>>> merged = intersection(cr_proj1, cr_proj2, on=['subject_id', 'image'])
>>> print(merged)
[(('subject_id', '0040000'), ('image', 'mprage.nii.gz')),
(('subject_id', '0040000'), ('image', 'rest.nii.gz')),
(('subject_id', '0040001'), ('image', 'rest.nii.gz')),
(('subject_id', '0040002'), ('image', 'mprage.nii.gz')),
(('subject_id', '0040002'), ('image', 'rest.nii.gz'))]


# I can pick these image crumbs from this result using the `build_paths` function.
>>> cr1.build_paths(merged, make_crumbs=True)
[Crumb("/home/hansel/proj1/0040000/{session}/{mod}/mprage.nii.gz"),
Crumb("/home/hansel/proj1/0040000/{session}/{mod}/rest.nii.gz"),
Crumb("/home/hansel/proj1/0040001/{session}/{mod}/rest.nii.gz"),
Crumb("/home/hansel/proj1/0040002/{session}/{mod}/mprage.nii.gz"),
Crumb("/home/hansel/proj1/0040002/{session}/{mod}/rest.nii.gz")]

>>> cr2.build_paths(merged, make_crumbs=True)
[Crumb("/home/alexandre/data/cobre/proj2/0040000/{session}/{mod}/mprage.nii.gz"),
Crumb("/home/alexandre/data/cobre/proj2/0040000/{session}/{mod}/rest.nii.gz"),
Crumb("/home/alexandre/data/cobre/proj2/0040001/{session}/{mod}/rest.nii.gz"),
Crumb("/home/alexandre/data/cobre/proj2/0040002/{session}/{mod}/mprage.nii.gz"),
Crumb("/home/alexandre/data/cobre/proj2/0040002/{session}/{mod}/rest.nii.gz")]

# adding 'mod' to the intersection would be:
>>> intersection(cr1, cr2, on=['subject_id', 'mod', 'image'])
[(('subject_id', '0040000'), ('mod', 'anat_1'), ('image', 'mprage.nii.gz')),
(('subject_id', '0040000'), ('mod', 'rest_1'), ('image', 'rest.nii.gz')),
(('subject_id', '0040001'), ('mod', 'rest_1'), ('image', 'rest.nii.gz')),
(('subject_id', '0040002'), ('mod', 'anat_1'), ('image', 'mprage.nii.gz')),
(('subject_id', '0040002'), ('mod', 'rest_1'), ('image', 'rest.nii.gz'))]


The `unfold` function
~~~~~~~~~~~~~~~~~~~~~

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

.. code:: python

>>> all_images = Crumb("/home/hansel/raw/{subject_id}/{session_id}/{modality}/{image}")
>>> all_images = crumb.unfold()
>>> print(all_images)
[Crumb("/home/hansel/data/raw/0040000/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040000/session_1/rest_1/rest.nii.gz"),
Crumb("/home/hansel/data/raw/0040001/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040001/session_1/rest_1/rest.nii.gz"),
Crumb("/home/hansel/data/raw/0040002/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040002/session_1/rest_1/rest.nii.gz"),
Crumb("/home/hansel/data/raw/0040003/session_1/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040003/session_1/rest_1/rest.nii.gz"),
...

# and you can ask for the value of the crumb argument in each element
>>> print(crumbs[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>}"``

.. code:: python

# only the session_0 folders
>>> s0_imgs = Crumb("/home/hansel/raw/{subject_id}/{session_id:*_0}/{modality}/{image}")
>>> s0_imgs = crumb.unfold()
>>> print(s0_imgs)
[Crumb("/home/hansel/data/raw/0040000/session_0/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040000/session_0/rest_1/rest.nii.gz"),
Crumb("/home/hansel/data/raw/0040001/session_0/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040001/session_0/rest_1/rest.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.

.. code:: python

# only the ``session_0`` folders
>>> s0_imgs = Crumb("/home/hansel/raw/{subject_id}/{session_id:^.*_0$}/{modality}/{image}",
>>> regex='re')
>>> s0_imgs = crumb.unfold()
>>> print(s0_imgs)
[Crumb("/home/hansel/data/raw/0040000/session_0/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040000/session_0/rest_1/rest.nii.gz"),
Crumb("/home/hansel/data/raw/0040001/session_0/anat_1/mprage.nii.gz"),
Crumb("/home/hansel/data/raw/0040001/session_0/rest_1/rest.nii.gz"),
...

The regular expressions can be checked with the `patterns` property.

.. code:: python

>>> print(s0_imgs.patterns)
{'session_id': '^.*_0$', 'modality': '', 'image': '', 'subject_id': ''}

And can be also modified with the `set_pattern` function.

.. code:: python

>>> s0_imgs.set_pattern('modality', 'a.*')
>>> print(s0_imgs.patterns)
{'session_id': '^.*_0$', 'modality': 'a.*', 'image': '', 'subject_id': ''}
>>> print(s0_imgs.path)
/home/hansel/raw/{subject_id}/{session_id:^.*_0$}/{modality:a.*}/{image}



More functionalities, ideas and comments are welcome.


Dependencies
============

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

.. code:: bash

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:

.. code:: bash

python setup.py install


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

.. code:: bash

python setup.py install --user

To install for all users on Unix/Linux:

.. code:: bash

python setup.py build
sudo python setup.py install


You can also install it in development mode with:

.. code:: bash

python setup.py develop


Development
===========

Code
----

Github
~~~~~~

You can check the latest sources with the command:

.. code:: bash

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


or if you have write privileges:

.. code:: bash

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 <http://pytest.org/>`__ to help us with the testing.

Otherwise you can run the tests executing:

.. code:: bash

python setup.py test

or

.. code:: bash

py.test

or

.. code:: bash

make test


.. |PyPI| image:: https://img.shields.io/pypi/v/hansel.svg
:target: https://pypi-hypernode.com/pypi/hansel

.. |Build Status| image:: https://travis-ci.org/alexsavio/hansel.svg?branch=master
:target: https://travis-ci.org/alexsavio/hansel

.. |Coverage Status| image:: https://coveralls.io/repos/alexsavio/hansel/badge.svg?branch=master&service=github
:target: https://coveralls.io/github/alexsavio/hansel?branch=master

.. |PyPI Downloads| image:: https://img.shields.io/pypi/dm/hansel.svg
:target: https://pypi-hypernode.com/pypi/hansel

.. |Code Health| image:: https://landscape.io/github/alexsavio/hansel/master/landscape.svg?style=flat
:target: https://landscape.io/github/alexsavio/hansel/master
:alt: Code Health

.. |Scrutinizer| image:: https://img.shields.io/scrutinizer/g/alexsavio/hansel.svg
:target: https://scrutinizer-ci.com/g/alexsavio/hansel/?branch=master
:alt: Scrutinizer Code Quality



Changelog
=========

Version 0.7.3
-------------
- Fix bug for Python 2.7


Version 0.7.2
-------------
- Fix README.rst because of bad syntax for PyPI.


Version 0.7.1
-------------
- Update README.rst


Version 0.7.0
-------------
- Refactoring of how Crumb works, now using string.Formatter. This will help with new features due to simpler logic.
Now it is not possible to change the syntax of the Crumbs, although I guess nobody is interested in that.
- Fixed a few bugs from previous versions.
- Now `copy` function is not a classmethod anymore, so you can do `crumb.copy()` as well as `Crumb.copy(crumb)`.
- `patterns` is not a dictionary anymore, the regexes are embedded in the `_path` string. The property `patterns`
returns the dictionary as before. The function `set_pattern` must be used instead to set a different pattern
to a given argument.


Version 0.6.2
-------------
- Add `pandas` helper functions.
- Add `utils` to convert from values_maps to dicts.
- Improve docstrings.


Version 0.6.1
-------------
- Change the behaviour or `intersection` with `len(arg_names) == 1` for compatibility with `crumb.build_path` function.
- Improve README, update with new examples using `intersection`.


Version 0.6.0
-------------
- Added `intersection` function in `utils.py`.
- Change of behaviour in `__getitem__`, now it returns a list of values even if is only the one replace string from `_argval`.
- General renaming of the private functions inside Crumbs, more in accordance to the `open_args`/`all_args` idea.
- Fixed a few bugs and now the generated crumbs from `unfold` and `ls` will have the same parameters as the original Crumb.


Version 0.5.5
-------------
- Added CHANGES.rst to MANIFEST.in


Version 0.5.4
-------------
- Deprecating `Crumbs.keys()` function.
- Renamed `Crumbs.keys()` to `Crumbs.open_args()` and added `Crumbs.all_args()`.
- Substitute the internal logic of Crumbs to work with `Crumbs.open_args()`, made it a bit faster.


Version 0.5.3
-------------
- Add `Crumbs.keys()` function.
- Rename `utils.remove_duplicates()` to `utils.rm_dups()`.


Version 0.5.2
-------------
- Add `utils.check_path` function.
- Fix `Crumb.split` function to return the not defined part of the crumb.


Version 0.5.1
-------------
- Add 're.ignorecase' option for the `regex` argument in the constructor.


Version 0.5.0
-------------
- Add Python 2.7 compatibility. Friends don't let friends use Python 2.7!


Version 0.4.2
-------------
- Improve documentation in README.
- Rename member `_argreg` to `patterns`, so the user can use it to manage the argument patterns.


Version 0.4.1
-------------

- Fix CHANGES.rst to correct restview in PyPI.
- Thanks to restview: https://pypi-hypernode.com/pypi/restview.


Version 0.4.0
-------------

- Fill CHANGES.rst.
- All outputs from `Crumb.ls` function will be sorted.
- Add regular expressions or `fnmatch` option for crumb arguments.
- Change `exists` behaviour. Now the empty crumb arguments will return False when `exist()`.
- Code clean up.
- Fix bugs.


Version 0.3.1
-------------

- Fix README.
- Code clean up.


Version 0.3.0
-------------

- Add `_argval` member, a dict which stores crumb arguments replacements.
- Add tests.
- Remove `rm_dups` option in `Crumb.ls` function.
- Remove conversion to `Paths` when `Crumb` has no crumb arguments in `Crumb.ls`.


Version 0.2.0
-------------

- Add `ignore_list` parameter in `Crumb` constructor.


Version 0.1.1
-------------

- Add `Crumb.unfold` function.
- Move `mktree` out of `Crumb` class.


Version 0.1.0
-------------

- Simplify code.
- Increase test coverage.
- Add `exist_check` to `Crumb.ls` function.
- Fix bugs.