I think you can just use an Integer instead of PhotoScan.HighFaceCount to specify a target count.
from the docs:
buildModel(surface=Arbitrary, interpolation=EnabledInterpolation,
face_count=MediumFaceCount[, source ][, classes])
Generate model for the chunk frame.
Parameters
• surface (PhotoScan.SurfaceType) – Type of object to be reconstructed.
• interpolation (PhotoScan.Interpolation) – Interpolation mode.
• face_count (PhotoScan.FaceCount or int) – Target face count.
• source (PhotoScan.PointsSource) – Selects between dense point cloud and sparse
point cloud. If not specified, uses dense cloud if available.
• classes (list of int) – List of dense point classes to be used for surface extraction.
Returns Success of operation.
Return type boolean