ben@BenMBP2021 ~ % cat j | jq
{
"jsonrpc": "2.0",
"id": 131,
"result": {
"signatures": [
{
"label": "ufunc()",
"documentation": {
"kind": "plaintext",
"value": "isnan(x, \\/, out=None, *, where=True, casting='same_kind', order='K', dtype=None, subok=True[, signature])\\n\\nTest element-wise for NaN and return result as a boolean array.\\n\\nParameters\\n----------\\nx : array_like\\n\\u00a0\\u00a0\\u00a0\\u00a0Input array.\\nout : ndarray, None, or tuple of ndarray and None, optional\\n\\u00a0\\u00a0\\u00a0\\u00a0A location into which the result is stored. If provided, it must have\\n\\u00a0\\u00a0\\u00a0\\u00a0a shape that the inputs broadcast to. If not provided or None,\\n\\u00a0\\u00a0\\u00a0\\u00a0a freshly-allocated array is returned. A tuple (possible only as a\\n\\u00a0\\u00a0\\u00a0\\u00a0keyword argument) must have length equal to the number of outputs.\\nwhere : array_like, optional\\n\\u00a0\\u00a0\\u00a0\\u00a0This condition is broadcast over the input. At locations where the\\n\\u00a0\\u00a0\\u00a0\\u00a0condition is True, the `out` array will be set to the ufunc result.\\n\\u00a0\\u00a0\\u00a0\\u00a0Elsewhere, the `out` array will retain its original value.\\n\\u00a0\\u00a0\\u00a0\\u00a0Note that if an uninitialized `out` array is created via the default\\n\\u00a0\\u00a0\\u00a0\\u00a0``out=None``, locations within it where the condition is False will\\n\\u00a0\\u00a0\\u00a0\\u00a0remain uninitialized.\\n**kwargs\\n\\u00a0\\u00a0\\u00a0\\u00a0For other keyword-only arguments, see the\\n\\u00a0\\u00a0\\u00a0\\u00a0:ref:`ufunc docs <ufuncs.kwargs>`.\\n\\nReturns\\n-------\\ny : ndarray or bool\\n\\u00a0\\u00a0\\u00a0\\u00a0True where ``x`` is NaN, false otherwise.\\n\\u00a0\\u00a0\\u00a0\\u00a0This is a scalar if `x` is a scalar.\\n\\nSee Also\\n--------\\nisinf, isneginf, isposinf, isfinite, isnat\\n\\nNotes\\n-----\\nNumPy uses the IEEE Standard for Binary Floating-Point for Arithmetic\\n(IEEE 754). This means that Not a Number is not equivalent to infinity.\\n\\nExamples\\n--------\\n>>> import numpy as np\\n>>> np.isnan(np.nan)\\nTrue\\n>>> np.isnan(np.inf)\\nFalse\\n>>> np.isnan([np.log(-1.),1.,np.log(0)])\\narray([ True, False, False])"
},
"parameters": [
{
"label": "args",
"documentation": {
"kind": "plaintext",
"value": ""
}
},
{
"label": "kwargs",
"documentation": {
"kind": "plaintext",
"value": ""
}
}
]
}
],
"activeSignature": 0,
"activeParameter": 0
}
}
Please raise this with the developer of the language server or use YCM's supported built-in Python support via Jedi. Or disable signature help of course.
See ycm-core/YouCompleteMe#4313, which is rooted in an invalid signature message received from the LSP server.
Quoting the YCM developer's analysis: