[SciPy-User] integrate.quad logical and ?

[josef.pktd at gmail.com]

I was trying to reduce the precision of integrate quad, since I only
need 3 decimals, but it didn't work until I reduce also epsrel and not
just epsabs.
epsabs=1e-4, epsrel=1e-2, limit=150 seems to work well so far.

It looks like they both have to be satisfied for quad to stop. Is this correct?
The docs don't say anything about it, and the recommendation to look
at the detailed explanation doesn't print anything.

>>> scipy.integrate.quad_explain()
>>> #nothing here
>>>

integrate.quadrature doc string explicitly says OR

Josef