scipy / stats : quantiles using sample weights from survey data

[Christopher Barrington-Leigh]

There is a function scipy.stats.mstats.mquantiles  that returns
quantiles for a vector of data.
But my data should not be uniformly weighted in an estimate of the
distribution, since they are from a survey and come with estimated
sampling weights based on the stratification used in sampling.

Is there a routine to calculate these quantiles taking into account
the survey weights?  I can find nothing, so maybe you have had the
same problem and written something.

Thanks!