grabrastervalues.py

import rasterio, fiona
from rasterio import features

uidfield = #whatever field contains unique values for extraction
infile = #input shapefile
inraster = #input raster

with rasterio.open(inraster, 'r') as src:
    rasArr = src.read_band(band_no)
    oshape = src.shape
    otrans = src.transform


with fiona.open(infile, 'r') as shp:
    fields = list(set(feat['properties'][uidfield] for feat in shp))
    sampleRaster = features.rasterize(
                    ((feat['geometry'], feat['properties'][uidfield]) for feat in shp),
                    out_shape=oshape,
                    transform=otrans)

#list of unique values
values = list(np.mean(rasArr[np.where(sampleRaster == i)]) for i in fields)

print values