Python: How to ARMA(p,q)-GARCH(r,s) fitting using ARCH Lib's mean model

Question

I'm trying, similar to R's ugarch

# standard GARCH model with optional ARMA part
        spec <- ugarchspec(variance.model = list(model = "sGARCH",    garchOrder = c(r,s)),
                           mean.model     = list(armaOrder = c(p,q)), distribution.model = dist[1])

ugarchfit(spec, data = x[,i], solver = "hybrid", fit.control = list(scale = 1),
                             numderiv.control = list(hess.eps = 1e-2))

fitting joint ARIMA(p,0,q)-GARCH(r,s) to several time series using ARCH library. Based on several test methods I would like to find out best fit parameters for p,q,r,s
Based on ARCH Documentation mean Models one can either choose No Mean,Constant Mean, Autoregressions and Heterogeneous Autoregressions.

arch.arch_model(y, x=None, mean='Constant', lags=0, vol='Garch', p=1, o=0, q=1, power=2.0, dist='Normal', hold_back=None)[source]

How can I specify optional MA components additional to AR in Python similar to statsmodels.tsa.arima_model.ARMA?

Thank you very much in advance.

Xerxes