A semi-classical inverse problem I: Taylor expansions.