Merge pull request #123 from seermedical/optional-dependencies

Move matplotlib and scipy into optional dependencies

README.md

@@ -7,7 +7,7 @@ Note that by default there is no public access to any data on the Seer Cloud. Yo
 
 ## Install
 
-To install, simply clone or download this repository, then type `pip install .` which will install all the dependencies.
+To install, simply clone or download this repository, then type `pip install .` which will install all the dependencies. To enable plotting signal data (required by the example notebook), use `pip install .[viz]`
 
 ### Epilepsy Ecosystem Data
 

requirements.txt

@@ -1,9 +1,7 @@
 gql
 requests
-matplotlib
 numpy
 pandas
-scipy
 # TODO: graphql-core 3.0.0 has a problem as at 02/12/2019
 # revisit in the future, we may be able to remove the restriction
 graphql-core<3.0.0

seerpy/utils.py

@@ -9,13 +9,9 @@
 from multiprocessing import Pool
 import os
 
-from matplotlib.collections import LineCollection
-from matplotlib import gridspec
-from matplotlib import pyplot as plt
 import numpy as np
 import pandas as pd
 import requests
-from scipy.signal import butter, sosfilt
 
 
 # pylint:disable=too-many-locals,too-many-statements
@@ -417,6 +413,14 @@ def plot_eeg(x, y=None, pred=None, squeeze=5.0, scaling_factor=None):
     >>> data_series = data_df.iloc[:, 0]
     >>> plot_eeg(x=data_series)
     """
+    try:
+        from matplotlib.collections import LineCollection
+        from matplotlib import gridspec
+        from matplotlib import pyplot as plt
+    except ModuleNotFoundError:
+        raise ModuleNotFoundError(
+            'Must have `matplotlib` installed. Try `pip install -U seerpy[viz]`')
+
     if not isinstance(x, np.ndarray):
         x = np.asarray(x)
 
@@ -469,118 +473,6 @@ def plot_eeg(x, y=None, pred=None, squeeze=5.0, scaling_factor=None):
     return plt
 
 
-def butter_bandstop(lowcut, highcut, fs, order=5):
-    """
-    Get second-order-sections representation of an IIR Butterworth digital
-    bandstop filter.
-
-    Parameters
-    ----------
-    lowcut : float
-        The lowcut critical frequency
-    highcut : float
-        The highcut critical frequency
-    fs : float
-        The sampling frequency of the digital system
-    order : int
-        The order of the filter
-
-    Returns
-    -------
-    filter_params : np.ndarray
-        Second-order-section filter parameters
-    """
-    nyq = 0.5 * fs
-    low = lowcut / nyq
-    high = highcut / nyq
-    sos = butter(order, [low, high], analog=False, btype='bandstop', output='sos')
-    return sos
-
-
-def butter_bandstop_filter(data, lowcut, highcut, fs, order=5):
-    """
-    Apply a bandstop filter to data along one dimension using cascaded
-    second-order sections.
-
-    Parameters
-    ----------
-    data : np.ndarray
-        Array of data to apply filter to
-    lowcut : float
-        The lowcut critical frequency
-    highcut : float
-        The highcut critical frequency
-    fs : float
-        The sampling frequency of the digital system
-    order : int
-        The order of the filter
-
-    Returns
-    -------
-    filtered_data : np.ndarray
-        The original data after applying the filter
-    """
-    sos = butter_bandstop(lowcut, highcut, fs, order=order)
-    y = sosfilt(sos, data)
-    return y
-
-
-def butter_bandpass(lowcut, highcut, fs, order=5):
-    """
-    Get second-order-sections representation of an IIR Butterworth digital
-    bandpass filter.
-
-    Parameters
-    ----------
-    lowcut : float
-        The lowcut critical frequency
-    highcut : float
-        The highcut critical frequency
-    fs : float
-        The sampling frequency of the digital system
-    order : int
-        The order of the filter
-
-    Returns
-    -------
-    filter_params : np.ndarray
-        Second-order-section filter parameters
-    """
-    nyq = 0.5 * fs
-    low = lowcut / nyq
-    high = highcut / nyq
-    sos = butter(order, [low, high], analog=False, btype='bandpass', output='sos')
-    return sos
-
-
-def butter_bandpass_filter(data, lowcut, highcut, fs, order=5):
-    """
-    Apply a bandpass filter to data along one dimension using cascaded
-    second-order sections.
-
-    Parameters
-    ----------
-    data : np.ndarray
-        Array of data to apply filter to
-    lowcut : float
-        The lowcut critical frequency
-    highcut : float
-        The highcut critical frequency
-    fs : float
-        The sampling frequency of the digital system
-    order : int
-        The order of the filter
-
-    Returns
-    -------
-    filtered_data : np.ndarray
-        The original data after applying the filter
-    """
-    sos = butter_bandpass(lowcut, highcut, fs, order=order)
-    y = sosfilt(sos, data)
-    return y
-
-
 def get_diary_fitbit_data(data_url):
     """
     Download Fitbit data from a given URL and return as a DataFrame.

setup.py

@@ -2,7 +2,7 @@
 
 setup(
     name='seerpy',
-    version='0.3.0',
+    version='0.4.0',
     description='Seer Platform SDK for Python',
     long_description=open('README.md').read(),
     url='https://github.com/seermedical/seer-py',
@@ -18,16 +18,7 @@
     ],
     keywords='api seer eeg ecg client',
     packages=find_packages(include=["seerpy*"]),
-    install_requires=[
-        'gql',
-        'requests',
-        'matplotlib',
-        'numpy',
-        'pandas',
-        'scipy',
-        'pyjwt[crypto]'
-    ],
-    tests_require=[
-        'pytest'
-    ],
+    install_requires=['gql', 'requests', 'numpy', 'pandas', 'pyjwt[crypto]'],
+    extras_require={'viz': ['matplotlib']},
+    tests_require=['pytest'],
 )