I'm +0 for changing "sum" into "integrate".

I'm +1 on showing the formula, however, it should use the nomenclature of the return (which currently is n, not h) and also should not use the assigment =, so probably np.sum(n * np.diff(bins)) == 1

I'm -1 on deleting the sentence on how the density is computed; although it could of course be reformulated to be less ambiguous.

I don't see that the "how" its computed really makes things clearer. For reference, here is numpy's doc:

If False, the result will contain the number of samples in each bin. If True, the result is the value of the probability density function at the bin, normalized such that the integral over the range is 1. Note that the sum of the histogram values will not be equal to 1 unless bins of unity width are chosen; it is not a probability mass function.

https://docs.scipy.org/doc/numpy/reference/generated/numpy.histogram.html

OK< changed to

              If ``True``, the first element of the return tuple will
           be the raw counts per bin normalized to form a probability density
             so that the area under the histogram will integrate to 1, i.e.
             ``np.sum(n * np.diff(bins)) == 1``.  (Computed by dividing the
             raw counts ``n0`` by the total number of data points and
             multiplying by the bin width: ``n = n0 / sum(n0) * np.diff(bins)``)

Thanks @jklymak !