improvements from review

pvlib/ivtools/utility.py

@@ -6,18 +6,13 @@
 
 import numpy as np
 import pandas as pd
-from collections import OrderedDict
 
 
 # A small number used to decide when a slope is equivalent to zero
 EPS = np.finfo('float').eps**(1/3)
 
 
-constants = OrderedDict()
-constants['E0'] = 1000.0
-constants['T0'] = 25.0
-constants['k'] = 1.38066e-23
-constants['q'] = 1.60218e-19
+constants = {'E0': 1000.0, 'T0': 25.0, 'k': 1.38066e-23, 'q': 1.60218e-19}
 
 
 def numdiff(x, f):
@@ -166,9 +161,6 @@ def rectify_iv_curve(voltage, current, decimals=None):
           equal to the average of current at duplicated voltages.
     """
 
-    if len(voltage) != len(current):
-        raise ValueError('voltage and current must have the same length')
-
     df = pd.DataFrame(data=np.vstack((voltage, current)).T, columns=['v', 'i'])
     # restrict to first quadrant
     df.dropna(inplace=True)
@@ -240,7 +232,6 @@ def schumaker_qspline(x, y):
     y = y.flatten()
 
     n = x.size
-    assert n == y.size
 
     # compute various values used by the algorithm: differences, length of line
     # segments between data points, and ratios of differences.
@@ -313,12 +304,12 @@ def schumaker_qspline(x, y):
 
     # MATLAB differs from NumPy, boolean indices must be same size as
     # array
-    xk[:(n-1)][u] = tmpx[u]
-    yk[:(n-1)][u] = tmpy[u]
+    xk[:(n - 1)][u] = tmpx[u]
+    yk[:(n - 1)][u] = tmpy[u]
     # constant term for each polynomial for intervals without knots
-    a[:(n-1), 2][u] = tmpy[u]
-    a[:(n-1), 1][u] = s[:-1][u]
-    a[:(n-1), 0][u] = 0.5 * diffs[u] / delx[u]  # leading coefficients
+    a[:(n - 1), 2][u] = tmpy[u]
+    a[:(n - 1), 1][u] = s[:-1][u]
+    a[:(n - 1), 0][u] = 0.5 * diffs[u] / delx[u]  # leading coefficients
 
     # [2], Algorithm 4.1 subpart 2 of Step 5
     # original x values that are left points of intervals with internal knots
@@ -355,7 +346,7 @@ def schumaker_qspline(x, y):
         uu[(n + q - 1):(n + q + r - 1), :] = np.array([tmpx[w], tmpx2[w],
                                                        tmpy[w], tmps[w],
                                                        tmps2[w], delta[w]]).T
-        xi[:(n-1)][w] = xk[(n + q - 1):(n + q + r - 1)]
+        xi[:(n - 1)][w] = xk[(n + q - 1):(n + q + r - 1)]
 
     z = np.logical_and(~u, ~v)  # last 'else' in Algorithm 4.1 Step 5
     z = np.logical_and(z, ~w)
@@ -372,13 +363,13 @@ def schumaker_qspline(x, y):
     # define polynomial coefficients for intervals with added knots
     ff = ~u
     sbar[:(n-1)][ff] = (
-        (2 * uu[:(n-1), 5][ff] - uu[:(n-1), 4][ff])
-        + (uu[:(n-1), 4][ff] - uu[:(n-1), 3][ff])
-        * (xi[:(n-1)][ff] - uu[:(n-1), 0][ff])
-        / (uu[:(n-1), 1][ff] - uu[:(n-1), 0][ff]))
+        (2 * uu[:(n - 1), 5][ff] - uu[:(n-1), 4][ff])
+        + (uu[:(n - 1), 4][ff] - uu[:(n-1), 3][ff])
+        * (xi[:(n - 1)][ff] - uu[:(n-1), 0][ff])
+        / (uu[:(n - 1), 1][ff] - uu[:(n-1), 0][ff]))
     eta[:(n-1)][ff] = (
-        (sbar[:(n-1)][ff] - uu[:(n-1), 3][ff])
-        / (xi[:(n-1)][ff] - uu[:(n-1), 0][ff]))
+        (sbar[:(n - 1)][ff] - uu[:(n-1), 3][ff])
+        / (xi[:(n - 1)][ff] - uu[:(n-1), 0][ff]))
 
     sbar[(n - 1):(n + q + r + ss - 1)] = \
         (2 * uu[(n - 1):(n + q + r + ss - 1), 5] -
@@ -396,9 +387,9 @@ def schumaker_qspline(x, y):
          uu[(n - 1):(n + q + r + ss - 1), 0])
 
     # constant term for polynomial for intervals with internal knots
-    a[:(n-1), 2][~u] = uu[:(n-1), 2][~u]
-    a[:(n-1), 1][~u] = uu[:(n-1), 3][~u]
-    a[:(n-1), 0][~u] = 0.5 * eta[:(n-1)][~u]  # leading coefficient
+    a[:(n - 1), 2][~u] = uu[:(n - 1), 2][~u]
+    a[:(n - 1), 1][~u] = uu[:(n - 1), 3][~u]
+    a[:(n - 1), 0][~u] = 0.5 * eta[:(n - 1)][~u]  # leading coefficient
 
     a[(n - 1):(n + q + r + ss - 1), 2] = \
         uu[(n - 1):(n + q + r + ss - 1), 2] + \

pvlib/tests/ivtools/test_sdm.py

@@ -217,14 +217,14 @@ def test_fit_pvsyst_sandia(npts=3000):
 
     pvsyst.update(zip(varlist, [Iph, Io, Rs, Rsh, u]))
 
-    expected = sdm.fit_pvsyst_sandia(ivcurves, iv_specs)
+    modeled = sdm.fit_pvsyst_sandia(ivcurves, iv_specs)
     param_res = pvsystem.calcparams_pvsyst(
         effective_irradiance=ivcurves['ee'], temp_cell=ivcurves['tc'],
-        alpha_sc=iv_specs['alpha_sc'], gamma_ref=expected['gamma_ref'],
-        mu_gamma=expected['mu_gamma'], I_L_ref=expected['I_L_ref'],
-        I_o_ref=expected['I_o_ref'], R_sh_ref=expected['R_sh_ref'],
-        R_sh_0=expected['R_sh_0'], R_s=expected['R_s'],
-        cells_in_series=iv_specs['cells_in_series'], EgRef=expected['EgRef'])
+        alpha_sc=iv_specs['alpha_sc'], gamma_ref=modeled['gamma_ref'],
+        mu_gamma=modeled['mu_gamma'], I_L_ref=modeled['I_L_ref'],
+        I_o_ref=modeled['I_o_ref'], R_sh_ref=modeled['R_sh_ref'],
+        R_sh_0=modeled['R_sh_0'], R_s=modeled['R_s'],
+        cells_in_series=iv_specs['cells_in_series'], EgRef=modeled['EgRef'])
     iv_res = pvsystem.singlediode(*param_res)
 
     ivcurves['p_mp'] = ivcurves['v_mp'] * ivcurves['i_mp']  # power
@@ -244,35 +244,35 @@ def test_fit_pvsyst_sandia(npts=3000):
     assert all((iv_specs[spec] == pvsyst_specs[spec]) for spec in spec_list)
     # I_L_ref, I_o_ref, EgRef, R_sh_ref, R_sh_0, R_sh_exp, R_s, gamma_ref,
     # mu_gamma
-    assert np.isclose(expected['I_L_ref'], pvsyst['I_L_ref'], rtol=6.5e-5)
-    assert np.isclose(expected['I_o_ref'], pvsyst['I_o_ref'], rtol=0.15)
-    assert np.isclose(expected['R_s'], pvsyst['R_s'], rtol=0.0035)
-    assert np.isclose(expected['R_sh_ref'], pvsyst['R_sh_ref'], rtol=0.091)
-    assert np.isclose(expected['R_sh_0'], pvsyst['R_sh_0'], rtol=0.013)
-    assert np.isclose(expected['EgRef'], pvsyst['EgRef'], rtol=0.037)
-    assert np.isclose(expected['gamma_ref'], pvsyst['gamma_ref'], rtol=0.0045)
-    assert np.isclose(expected['mu_gamma'], pvsyst['mu_gamma'], rtol=0.064)
+    assert np.isclose(modeled['I_L_ref'], pvsyst['I_L_ref'], rtol=6.5e-5)
+    assert np.isclose(modeled['I_o_ref'], pvsyst['I_o_ref'], rtol=0.15)
+    assert np.isclose(modeled['R_s'], pvsyst['R_s'], rtol=0.0035)
+    assert np.isclose(modeled['R_sh_ref'], pvsyst['R_sh_ref'], rtol=0.091)
+    assert np.isclose(modeled['R_sh_0'], pvsyst['R_sh_0'], rtol=0.013)
+    assert np.isclose(modeled['EgRef'], pvsyst['EgRef'], rtol=0.037)
+    assert np.isclose(modeled['gamma_ref'], pvsyst['gamma_ref'], rtol=0.0045)
+    assert np.isclose(modeled['mu_gamma'], pvsyst['mu_gamma'], rtol=0.064)
 
     # Iph, Io, Rsh, Rs, u
-    mask = np.ones(expected['u'].shape, dtype=bool)
+    mask = np.ones(modeled['u'].shape, dtype=bool)
     # exclude one curve with different convergence
     umask = mask.copy()
     umask[2540] = False
-    assert all(expected['u'][umask] == pvsyst['u'][:npts][umask])
+    assert all(modeled['u'][umask] == pvsyst['u'][:npts][umask])
     assert np.allclose(
-        expected['iph'][expected['u']], pvsyst['iph'][:npts][expected['u']],
+        modeled['iph'][modeled['u']], pvsyst['iph'][:npts][modeled['u']],
         equal_nan=True, rtol=0.0009)
     assert np.allclose(
-        expected['io'][expected['u']], pvsyst['io'][:npts][expected['u']],
+        modeled['io'][modeled['u']], pvsyst['io'][:npts][modeled['u']],
         equal_nan=True, rtol=0.096)
     assert np.allclose(
-        expected['rs'][expected['u']], pvsyst['rs'][:npts][expected['u']],
+        modeled['rs'][modeled['u']], pvsyst['rs'][:npts][modeled['u']],
         equal_nan=True, rtol=0.035)
     # exclude one curve with Rsh outside 63% tolerance
-    rshmask = expected['u'].copy()
+    rshmask = modeled['u'].copy()
     rshmask[2545] = False
     assert np.allclose(
-        expected['rsh'][rshmask], pvsyst['rsh'][:npts][rshmask],
+        modeled['rsh'][rshmask], pvsyst['rsh'][:npts][rshmask],
         equal_nan=True, rtol=0.63)
 
 

setup.py

@@ -44,8 +44,9 @@
 TESTS_REQUIRE = ['nose', 'pytest', 'pytest-cov', 'pytest-mock',
                  'pytest-timeout', 'pytest-rerunfailures', 'pytest-remotedata']
 EXTRAS_REQUIRE = {
-    'optional': ['ephem', 'cython', 'netcdf4', 'nrel-pysam', 'numba',
-                 'pvfactors', 'scipy', 'siphon', 'tables', 'cftime >= 1.1.1'],
+    'optional': ['cython', 'ephem', 'netcdf4', 'nrel-pysam', 'numba',
+                 'pvfactors', 'scipy', 'siphon', 'statsmodels', 'tables',
+                 'cftime >= 1.1.1'],
     'doc': ['ipython', 'matplotlib', 'sphinx == 1.8.5', 'sphinx_rtd_theme',
             'sphinx-gallery', 'docutils == 0.15.2', 'pillow', 'scipy',
             'netcdf4', 'siphon', 'tables'],