This is a python package for converting xy coordinates (in lng,lat order) in numpy array between WGS-84, GCJ-02 and BD-09 system.
The conversion in numpy is much more efficient (50-70 times speedup) compared with previous Python implementations such as eviltransform and coordTransform_py that convert only a single point at each call.
pip install xyconvert
from xyconvert import gcj2wgs
import numpy as np
gcj_xy = np.asarray([[104.07008157, 30.73199687],
[104.07008159, 30.73177709],
[104.06999188, 30.73147758]])
wgs_xy = gcj2wgs(gcj_xy)
print wgs_xy
'''
[[104.06756309 30.73437796]
[104.06756312 30.73415829]
[104.06747357 30.73385904]]
'''
The code is partly copied/modified from https://github.com/Argons/nextlocation and https://github.com/wandergis/coordTransform_py.