Merge pull request #156 from pycroscopy/main

Getting the recent changes

.github/workflows/actions.yml

@@ -1,7 +1,7 @@
 name: build
 
 env:
-  PYTHON_MAIN_VERSION: 3.8
+  PYTHON_MAIN_VERSION: 3.9
 
 on:
   pull_request:
@@ -19,7 +19,7 @@ jobs:
     strategy:
       max-parallel: 5
       matrix:
-        python-version: [3.8, 3.7, 3.6, 3.9]
+        python-version: [3.8, 3.9, "3.10", 3.7, 3.6]
 
     steps:
     - uses: actions/checkout@v2
@@ -69,15 +69,15 @@ jobs:
         verbose: true
 
     - name: Documentation build
-      if: ${{ matrix.python-version == env.PYTHON_MAIN_VERSION && github.ref == 'refs/heads/master'}}
+      if: ${{ matrix.python-version == env.PYTHON_MAIN_VERSION && github.ref == 'refs/heads/main'}}
       run: |
         pip install sphinx>=3.1.1 sphinx-gallery sphinx-rtd-theme>=0.5.0 sphinx-autodoc-typehints numpydoc wget pysptools cvxopt scipy nbsphinx
         sudo apt-get install pandoc
         sphinx-build -b html -aET docs/source docs/_build/html
         touch docs/_build/html/.nojekyll
         
     - name: Deploy to GitHub Pages
-      if: ${{ matrix.python-version == env.PYTHON_MAIN_VERSION && github.ref == 'refs/heads/master'}}
+      if: ${{ matrix.python-version == env.PYTHON_MAIN_VERSION && github.ref == 'refs/heads/main'}}
       uses: crazy-max/ghaction-github-pages@v2
       with:
         target_branch: gh-pages

README.rst

@@ -8,6 +8,10 @@ sidpy
 .. image:: https://img.shields.io/pypi/v/sidpy.svg
     :target: https://pypi.org/project/sidpy/
     :alt: PyPI
+
+.. image:: https://img.shields.io/conda/vn/conda-forge/sidpy.svg
+    :target: https://github.com/conda-forge/sidpy-feedstock
+    :alt: conda-forge
 
 .. image:: https://codecov.io/gh/pycroscopy/sidpy/branch/master/graph/badge.svg?token=BCFR4FR6AL
     :target: https://codecov.io/gh/pycroscopy/sidpy

sidpy/__version__.py

@@ -1,2 +1,2 @@
-version = '0.0.8'
-time = '2022-01-07 17:08:00'
+version = '0.10'
+time = '2022-04-29 17:08:00'

sidpy/proc/fitter.py

@@ -6,12 +6,15 @@
 @author: Rama Vasudevan, Mani Valleti
 """
 
+from xml.dom import NotFoundErr
 from dask.distributed import Client
 import numpy as np
 import dask
 import inspect
 from ..sid import Dimension, Dataset
 from ..sid.dimension import DimensionType
+from ..viz.dataset_viz import SpectralImageFitVisualizer
+from ..sid.dataset import DataType
 
 try:
     from scipy.optimize import curve_fit
@@ -47,7 +50,7 @@ def __init__(self, sidpy_dataset, fit_fn, xvec=None, ind_dims=None, guess_fn=Non
             If NOT provided, it is assumed that all the non-spectral dimensions are independent dimensions.
 
         guess_fn: (function) (optional) This optional function should be utilized to generate priors for the full fit
-        It takes the same arguments as the fitting function and should return the same type of results array.
+        It takes (xvec,yvec) as inputs and should return the fit parameters.
         If the guess_fn is NOT provided, then the user MUST input the num_fit_parms.
 
         num_fit_parms: (int) Number of fitting parameters. This is needed IF the guess function is not provided to set
@@ -170,6 +173,7 @@ def __init__(self, sidpy_dataset, fit_fn, xvec=None, ind_dims=None, guess_fn=Non
         self.return_std = return_std
         self.return_cov = return_cov
         self.return_fit = return_fit
+        self.fitted_dset = None
 
         self.mean_fit_results = []
         if self.return_cov:
@@ -251,7 +255,7 @@ def do_fit(self, **kwargs):
                     p0 = self.prior[ind, :]
 
                 lazy_result = dask.delayed(SidFitter.default_curve_fit)(self.fit_fn, self.dep_vec,
-                                                                        self.folded_dataset[ind, :],
+                                                                        self.folded_dataset[ind, :],self.num_fit_parms,
                                                                         return_cov=(self.return_cov or self.return_std),
                                                                         p0=p0, **kwargs)
                 fit_results.append(lazy_result)
@@ -263,7 +267,7 @@ def do_fit(self, **kwargs):
             self.get_km_priors()
             for ind in range(self.num_computations):
                 lazy_result = dask.delayed(SidFitter.default_curve_fit)(self.fit_fn, self.dep_vec,
-                                                                        self.folded_dataset[ind, :],
+                                                                        self.folded_dataset[ind, :], self.num_fit_parms,
                                                                         return_cov=(self.return_cov or self.return_std),
                                                                         p0=self.km_priors[self.km_labels[ind]],
                                                                         **kwargs)
@@ -394,7 +398,7 @@ def get_fitted_dataset(self):
         fitted_sid_dset_folded = fitted_dset_fold.like_data(np_folded_arr, title=fitted_dset_fold.title)
         fitted_sid_dset = fitted_sid_dset_folded.unfold()
         fitted_sid_dset.original_metadata = self.dataset.original_metadata.copy()
-
+        self.fitted_dset = fitted_sid_dset
         return fitted_sid_dset
 
     def get_km_priors(self):
@@ -422,23 +426,48 @@ def get_km_priors(self):
             else:
                 p0 = np.random.normal(loc=0.5, scale=0.1, size=self.num_fit_parms)
 
-            km_priors.append(SidFitter.default_curve_fit(self.fit_fn, self.dep_vec, cen,
+            km_priors.append(SidFitter.default_curve_fit(self.fit_fn, self.dep_vec, cen, self.num_fit_parms,
                                                          return_cov=False,
                                                          p0=p0, maxfev=10000))
         self.km_priors = np.array(km_priors)
         self.num_fit_parms = self.km_priors.shape[-1]
 
+    def visualize_fit_results(self, figure = None, horizontal = True):
+        '''
+        Calls the interactive visualizer for comparing raw and fit datasets.
+
+        Inputs:
+            - figure: (Optional, default None) - handle to existing figure
+            - horiziontal: (Optional, default True) - whether spectrum should be plotted horizontally
+
+        '''
+        dset_type = self.dataset.data_type 
+        supported_types = ['SPECTRAL_IMAGE']
+        if self.fitted_dset == None:
+            raise NotFoundErr("No fitted dataset found. Re-run with return_fit=True to use this feature")
+        if dset_type == DataType.SPECTRAL_IMAGE:
+            visualizer = SpectralImageFitVisualizer(self.dataset, self.fitted_dset,
+                    figure=figure, horizontal=horizontal)
+        else:
+            raise NotImplementedError("Data type is {} but currently we only support types {}".format(dset_type, supported_types))
+
+        return visualizer
+
+
     @staticmethod
-    def default_curve_fit(fit_fn, xvec, yvec, return_cov=True, **kwargs):
+    def default_curve_fit(fit_fn, xvec, yvec, num_fit_parms, return_cov=True, **kwargs):
         xvec = np.array(xvec)
         yvec = np.array(yvec)
         yvec = yvec.ravel()
         xvec = xvec.ravel()
         if curve_fit is None:
             raise ModuleNotFoundError("scipy is not installed")
         else:
-            popt, pcov = curve_fit(fit_fn, xvec, yvec, **kwargs)
-
+            try:
+                popt, pcov = curve_fit(fit_fn, xvec, yvec, **kwargs)
+            except:
+                popt = np.zeros(num_fit_parms)
+                pcov = np.zeros((num_fit_parms, num_fit_parms))
         if return_cov:
             return popt, pcov
         else:

sidpy/sid/dimension.py

@@ -31,6 +31,7 @@ class DimensionType(Enum):
     RECIPROCAL = 2
     SPECTRAL = 3
     TEMPORAL = 4
+    CHANNEL = 5
 
 
 class Dimension(np.ndarray):
@@ -53,7 +54,7 @@ def __new__(cls, values, name='none', quantity='generic', units='generic',
             set of values will be generated if an integer is provided instead
             of an array.
         dimension_type : str or sidpy.sid.dimension.DimensionType
-            For example: 'spectral', 'spatial', 'reciprocal', or 'UNKNOWN'
+            For example: 'spectral', 'spatial', 'reciprocal', 'channel', or 'UNKNOWN',
             'time', 'frame', 'reciprocal'
             This will determine how the data are visualized. 'spatial' are
             image dimensions. 'spectral' indicate spectroscopy data dimensions.