I always get error messages with the function "sample_entropy" that the types at _numba_sampen are not the correct ones:
"No matching definition for argument type(s) readonly array(float64, 1d, C), int64, float64".
I was able to fix the error by hard specifying the data types:
Old:
if metric == "chebyshev" and x.size < 5000: return _numba_sampen(x, order=order, r=0.2 * x.std(ddof=0))
New:
if metric == "chebyshev" and x.size < 5000: return _numba_sampen(x.astype(np.float64), order=np.int32(order), r=np.float64(0.2 * x.std(ddof=0)))
Can you share an example code that would produce the error?
That's surprising because we already enforce that x is a float64. But better safe than sorry, please feel free to submit the above modifications as a PR!
I always get error messages with the function "sample_entropy" that the types at _numba_sampen are not the correct ones: "No matching definition for argument type(s) readonly array(float64, 1d, C), int64, float64".
I was able to fix the error by hard specifying the data types: Old:
if metric == "chebyshev" and x.size < 5000: return _numba_sampen(x, order=order, r=0.2 * x.std(ddof=0))
New:
if metric == "chebyshev" and x.size < 5000: return _numba_sampen(x.astype(np.float64), order=np.int32(order), r=np.float64(0.2 * x.std(ddof=0)))