Address several numpy/pandas deprecation warnings; clean up test output (#1930)

kandersolar · web-flow · commit 8057393e6eaa · 2023-12-19T09:46:29.000-05:00
* install pytest-remotedata from conda-forge

The version in default channel is old.  Version in conda-forge is up to date and contains a fix for this deprecation warning:

DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.

* nix `unit='T'` in pd.to_timedelta

Use `unit='minutes'` instead of `unit='T'` ('T' was deprecated in pandas 2.1).  Fixes:

FutureWarning: Unit 'T' is deprecated and will be removed in a future version.

* Fix StringIO error regex for python 3.12

* use iloc with pandas Series

Fixes:

FutureWarning: Series.__setitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To set a value by position, use `ser.iloc[pos] = value`

* use pandas bfill()/ffill() instead of fillna(method='...')

Fixes:

FutureWarning: DataFrame.fillna with 'method' is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.

* avoid incompatible dtypes in solarposition.ephemeris

Fixes:

FutureWarning: Setting an item of incompatible dtype is deprecated and will raise in a future error of pandas. Value '[69.43454873]' has dtype incompatible with int64, please explicitly cast to a compatible dtype first.

* convert np array of length 1 to scalar

Fixes:

DeprecationWarning: Conversion of an array with ndim &gt; 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)

* use numpy.errstate in bifacial.utils tests

Fixes:

RuntimeWarning: invalid value encountered in multiply
RuntimeWarning: invalid value encountered in scalar multiply

* Fix another pandas iloc warning

Fixes:

FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
diff --git a/ci/requirements-py3.10.yml b/ci/requirements-py3.10.yml
@@ -17,7 +17,7 @@ dependencies:
     - requests-mock
     - pytest-timeout
     - pytest-rerunfailures
-    - pytest-remotedata
+    - conda-forge::pytest-remotedata  # version in default channel is old
     - python=3.10
     - pytz
     - requests
diff --git a/ci/requirements-py3.11.yml b/ci/requirements-py3.11.yml
@@ -17,7 +17,7 @@ dependencies:
     - requests-mock
     - pytest-timeout
     - pytest-rerunfailures
-    - pytest-remotedata
+    - conda-forge::pytest-remotedata  # version in default channel is old
     - python=3.11
     - pytz
     - requests
diff --git a/ci/requirements-py3.12.yml b/ci/requirements-py3.12.yml
@@ -17,7 +17,7 @@ dependencies:
     - requests-mock
     - pytest-timeout
     - pytest-rerunfailures
-    - pytest-remotedata
+    - conda-forge::pytest-remotedata  # version in default channel is old
     - python=3.12
     - pytz
     - requests
diff --git a/ci/requirements-py3.7.yml b/ci/requirements-py3.7.yml
@@ -17,7 +17,7 @@ dependencies:
     - requests-mock
     - pytest-timeout
     - pytest-rerunfailures
-    - pytest-remotedata
+    - conda-forge::pytest-remotedata  # version in default channel is old
     - python=3.7
     - pytz
     - requests
diff --git a/ci/requirements-py3.8.yml b/ci/requirements-py3.8.yml
@@ -17,7 +17,7 @@ dependencies:
     - requests-mock
     - pytest-timeout
     - pytest-rerunfailures
-    - pytest-remotedata
+    - conda-forge::pytest-remotedata  # version in default channel is old
     - python=3.8
     - pytz
     - requests
diff --git a/ci/requirements-py3.9.yml b/ci/requirements-py3.9.yml
@@ -17,7 +17,7 @@ dependencies:
     - requests-mock
     - pytest-timeout
     - pytest-rerunfailures
-    - pytest-remotedata
+    - conda-forge::pytest-remotedata  # version in default channel is old
     - python=3.9
     - pytz
     - requests
diff --git a/pvlib/iotools/bsrn.py b/pvlib/iotools/bsrn.py
@@ -321,7 +321,7 @@ def parse_bsrn(fbuf, logical_records=('0100',)):
         LR_0100.columns = BSRN_LR0100_COLUMNS
         # Set datetime index
         LR_0100.index = (start_date+pd.to_timedelta(LR_0100['day']-1, unit='d')
-                         + pd.to_timedelta(LR_0100['minute'], unit='T'))
+                         + pd.to_timedelta(LR_0100['minute'], unit='minutes'))
         # Drop empty, minute, and day columns
         LR_0100 = LR_0100.drop(columns=['empty', 'day', 'minute'])
         dfs.append(LR_0100)
@@ -335,7 +335,7 @@ def parse_bsrn(fbuf, logical_records=('0100',)):
                               colspecs=BSRN_LR0300_COL_SPECS,
                               names=BSRN_LR0300_COLUMNS)
         LR_0300.index = (start_date+pd.to_timedelta(LR_0300['day']-1, unit='d')
-                         + pd.to_timedelta(LR_0300['minute'], unit='T'))
+                         + pd.to_timedelta(LR_0300['minute'], unit='minutes'))
         LR_0300 = LR_0300.drop(columns=['day', 'minute']).astype(float)
         dfs.append(LR_0300)
 
@@ -352,7 +352,7 @@ def parse_bsrn(fbuf, logical_records=('0100',)):
         LR_0500 = LR_0500.reindex(sorted(LR_0500.columns), axis='columns')
         LR_0500.columns = BSRN_LR0500_COLUMNS
         LR_0500.index = (start_date+pd.to_timedelta(LR_0500['day']-1, unit='d')
-                         + pd.to_timedelta(LR_0500['minute'], unit='T'))
+                         + pd.to_timedelta(LR_0500['minute'], unit='minutes'))
         LR_0500 = LR_0500.drop(columns=['empty', 'day', 'minute'])
         dfs.append(LR_0500)
 
diff --git a/pvlib/ivtools/sdm.py b/pvlib/ivtools/sdm.py
@@ -680,7 +680,7 @@ def _fit_desoto_sandia_diode(ee, voc, vth, tc, specs, const):
     y = voc - specs['beta_voc'] * (tc - const['T0'])
     new_x = sm.add_constant(x)
     res = sm.RLM(y, new_x).fit()
-    return res.params[1]
+    return np.array(res.params)[1]
 
 
 def _initial_iv_params(ivcurves, ee, voc, isc, rsh, nnsvth):
diff --git a/pvlib/scaling.py b/pvlib/scaling.py
@@ -134,6 +134,7 @@ def fn(x):
         return np.abs((x ** 2 - x) / 2 - n_pairs)
 
     n_dist = np.round(scipy.optimize.fmin(fn, np.sqrt(n_pairs), disp=False))
+    n_dist = n_dist.item()
     # Compute VR
     A = cloud_speed / 2  # Resultant fit for A from [2]
     vr = np.zeros(tmscales.shape)
@@ -276,7 +277,7 @@ def _compute_wavelet(clearsky_index, dt=None):
         # Produces slightly different end effects than the MATLAB version
         df = cs_long.rolling(window=intvlen, center=True, min_periods=1).mean()
         # Fill nan's in both directions
-        df = df.fillna(method='bfill').fillna(method='ffill')
+        df = df.bfill().ffill()
         # Pop values back out of the dataframe and store
         csi_mean[i, :] = df.values.flatten()
         # Shift to account for different indexing in MATLAB moving average
diff --git a/pvlib/snow.py b/pvlib/snow.py
@@ -53,9 +53,9 @@ def fully_covered_nrel(snowfall, threshold_snowfall=1.):
     freq = pd.infer_freq(snowfall.index)
     if freq is not None:
         timedelta = pd.tseries.frequencies.to_offset(freq) / pd.Timedelta('1h')
-        hourly_snow_rate.iloc[0] = snowfall[0] / timedelta
+        hourly_snow_rate.iloc[0] = snowfall.iloc[0] / timedelta
     else:  # can't infer frequency from index
-        hourly_snow_rate[0] = 0  # replaces NaN
+        hourly_snow_rate.iloc[0] = 0  # replaces NaN
     return hourly_snow_rate > threshold_snowfall
 
 
diff --git a/pvlib/soiling.py b/pvlib/soiling.py
@@ -91,7 +91,7 @@ def hsu(rainfall, cleaning_threshold, surface_tilt, pm2_5, pm10,
     mass_no_cleaning = pd.Series(index=rainfall.index, data=tms_cumsum)
     # specify dtype so pandas doesn't assume object
     mass_removed = pd.Series(index=rainfall.index, dtype='float64')
-    mass_removed[0] = 0.
+    mass_removed.iloc[0] = 0.
     mass_removed[cleaning_times] = mass_no_cleaning[cleaning_times]
     accum_mass = mass_no_cleaning - mass_removed.ffill()
 
diff --git a/pvlib/solarposition.py b/pvlib/solarposition.py
@@ -832,7 +832,7 @@ def ephemeris(time, latitude, longitude, pressure=101325, temperature=12):
     # Calculate refraction correction
     Elevation = SunEl
     TanEl = pd.Series(np.tan(np.radians(Elevation)), index=time_utc)
-    Refract = pd.Series(0, index=time_utc)
+    Refract = pd.Series(0., index=time_utc)
 
     Refract[(Elevation > 5) & (Elevation <= 85)] = (
         58.1/TanEl - 0.07/(TanEl**3) + 8.6e-05/(TanEl**5))
diff --git a/pvlib/tests/bifacial/test_utils.py b/pvlib/tests/bifacial/test_utils.py
@@ -120,13 +120,15 @@ def test_vf_row_sky_2d(test_system_fixed_tilt):
 def test_vf_row_sky_2d_integ(test_system_fixed_tilt):
     ts, _, _ = test_system_fixed_tilt
     # with float input, check end position
-    vf = utils.vf_row_sky_2d_integ(ts['surface_tilt'], ts['gcr'], 1., 1.)
+    with np.errstate(invalid='ignore'):
+        vf = utils.vf_row_sky_2d_integ(ts['surface_tilt'], ts['gcr'], 1., 1.)
     expected = utils.vf_row_sky_2d(ts['surface_tilt'], ts['gcr'], 1.)
     assert np.isclose(vf, expected)
     # with array input
     fx0 = np.array([0., 0.5])
     fx1 = np.array([0., 0.8])
-    vf = utils.vf_row_sky_2d_integ(ts['surface_tilt'], ts['gcr'], fx0, fx1)
+    with np.errstate(invalid='ignore'):
+        vf = utils.vf_row_sky_2d_integ(ts['surface_tilt'], ts['gcr'], fx0, fx1)
     phi = masking_angle(ts['surface_tilt'], ts['gcr'], fx0[0])
     y0 = 0.5 * (1 + cosd(ts['surface_tilt'] + phi))
     x = np.arange(fx0[1], fx1[1], 1e-4)
@@ -161,13 +163,17 @@ def test_vf_row_ground_2d(test_system_fixed_tilt):
 def test_vf_ground_2d_integ(test_system_fixed_tilt):
     ts, _, _ = test_system_fixed_tilt
     # with float input, check end position
-    vf = utils.vf_row_ground_2d_integ(ts['surface_tilt'], ts['gcr'], 0., 0.)
+    with np.errstate(invalid='ignore'):
+        vf = utils.vf_row_ground_2d_integ(ts['surface_tilt'], ts['gcr'],
+                                          0., 0.)
     expected = utils.vf_row_ground_2d(ts['surface_tilt'], ts['gcr'], 0.)
     assert np.isclose(vf, expected)
     # with array input
     fx0 = np.array([0., 0.5])
     fx1 = np.array([0., 0.8])
-    vf = utils.vf_row_ground_2d_integ(ts['surface_tilt'], ts['gcr'], fx0, fx1)
+    with np.errstate(invalid='ignore'):
+        vf = utils.vf_row_ground_2d_integ(ts['surface_tilt'], ts['gcr'],
+                                          fx0, fx1)
     phi = ground_angle(ts['surface_tilt'], ts['gcr'], fx0[0])
     y0 = 0.5 * (1 - cosd(phi - ts['surface_tilt']))
     x = np.arange(fx0[1], fx1[1], 1e-4)
diff --git a/pvlib/tests/iotools/test_pvgis.py b/pvlib/tests/iotools/test_pvgis.py
@@ -201,8 +201,9 @@ def test_read_pvgis_hourly_bad_extension():
     # Test if ValueError is raised if an unkonwn pvgis_format is specified
     with pytest.raises(ValueError, match="pvgis format 'txt' was unknown"):
         read_pvgis_hourly(testfile_pv_json, pvgis_format='txt')
-    # Test if TypeError is raised if input is a buffer and pvgis_format=None
-    with pytest.raises(TypeError, match="expected str, bytes or os.PathLike"):
+    # Test if TypeError is raised if input is a buffer and pvgis_format=None.
+    # The error text changed in python 3.12. This regex matches both versions:
+    with pytest.raises(TypeError, match="str.*os.PathLike"):
         read_pvgis_hourly(io.StringIO())
 
 
