BUG: core.missing._akima_interpolate will raise AttributeError

[jbrockmendel]

    from scipy import interpolate

    P = interpolate.Akima1DInterpolator(xi, yi, axis=axis)

    if der == 0:
        return P(x)
    elif interpolate._isscalar(der):
        return P(x, der=der)
    else:
        return [P(x, nu) for nu in der]

There is no interpolate._isscalar so this would raise if it were ever reached. AFAICT we have only one test that gets to this function, must always have der == 0

[jbrockmendel]

I'm guessing this is the function that we're trying to get at:

def _isscalar(x):
    """Check whether x is if a scalar type, or 0-dim"""
    return np.isscalar(x) or hasattr(x, 'shape') and x.shape == ()

[fifthist]

We could

1. Import _isscalar directly inside the definition of _akima_interpolate.

from scipy.interpolate.polyint import _isscalar

2. Use is_scalar, which is already imported in the module. Looking at it's definition, it seems to check both cases that _isscalar checks. However, this considers datetime objects and None to be scalars, so it may be too nonrestrictive.

[jbrockmendel]

we'll want to use is_scalar, will probably also need a lib.item_from_zerodim.

[hkennyv]

take

[hkennyv]

Hey all, I'm trying to add to the test_interpolate_akima test so that it covers cases where der is a list but I'm having trouble debugging this error:

Exception has occurred: TypeError
NumPy boolean array indexing assignment requires a 0 or 1-dimensional input, input has 2 dimensions
  File "/Users/khuynh/github/pandas/pandas/core/missing.py", line 303, in interpolate_1d
    result[invalid] = _interpolate_scipy_wrapper(
  File "/Users/khuynh/github/pandas/pandas/core/internals/blocks.py", line 1203, in func
    return missing.interpolate_1d(
  File "<__array_function__ internals>", line 5, in apply_along_axis
  File "/Users/khuynh/github/pandas/pandas/core/internals/blocks.py", line 1216, in _interpolate
    interp_values = np.apply_along_axis(func, axis, data)
  File "/Users/khuynh/github/pandas/pandas/core/internals/blocks.py", line 1113, in interpolate
    return self._interpolate(
  File "/Users/khuynh/github/pandas/pandas/core/internals/managers.py", line 397, in apply
    applied = getattr(b, f)(**kwargs)
  File "/Users/khuynh/github/pandas/pandas/core/internals/managers.py", line 558, in interpolate
    return self.apply("interpolate", **kwargs)
  File "/Users/khuynh/github/pandas/pandas/core/generic.py", line 6897, in interpolate
    new_data = data.interpolate(
  File "/Users/khuynh/github/pandas/test.py", line 33, in <module>
    interp_s = ser.reindex(new_index).interpolate(method="akima", der=[1, 2])

I'm getting this error when running this test file I made so I could step through with my debugger (using the editable install btw, you'll run into other errors if you run it with the current master. see #33426 6c42496 and #33426 a147bb0):

import numpy as np
import pytest

import pandas.util._test_decorators as td

import pandas as pd
from pandas import Index, MultiIndex, Series, date_range, isna
import pandas._testing as tm

print("hello world!")

ser = Series([10, 11, 12, 13])

expected = Series(
    [11.00, 11.25, 11.50, 11.75, 12.00, 12.25, 12.50, 12.75, 13.00],
    index=Index([1.0, 1.25, 1.5, 1.75, 2.0, 2.25, 2.5, 2.75, 3.0]),
)
# interpolate at new_index
new_index = ser.index.union(Index([1.25, 1.5, 1.75, 2.25, 2.5, 2.75])).astype(
    float
)
interp_s = ser.reindex(new_index).interpolate(method="akima")
# tm.assert_series_equal(interp_s[1:3], expected)
print(interp_s)

# test interpolate when a non-zero int
interp_s = ser.reindex(new_index).interpolate(method="akima", der=1)
# tm.assert_series_equal(interp_s[1:3], expected)
print(interp_s)
print(interp_s[1:3])

# test interpolate when der is a list
interp_s = ser.reindex(new_index).interpolate(method="akima", der=[1, 2])
# tm.assert_series_equal(interp_s[1:3], expected)

print(interp_s)
print('success!')

Any pointers?

[jbrockmendel]

test so that it covers cases where der is a list

why?

[hkennyv]

because the docstring for core.missing._akima_interpolate specififes der to be "int or list". But I see upon checking scipy.interpolate.PPoly.__call__ that the docstring is incorrect.

I can modify the docstring for _akima_interpolate and remove that else branch and submit a PR if everything else checks out?

[jbrockmendel]

I can modify the docstring for _akima_interpolate and remove that else branch and submit a PR if everything else checks out?

sounds good