Random state UMAP and output for make_cloud

[BernalFA]

Using dedenser as a python package, I noticed the needed for some modifications.

1. The umap implementation on make_cloud does not have a fixed random_state, causing different downsampling outcomes when repeatedly calling the function and Dedenser.downsample() on the same dataset. I added a random seed for umap.
2. To improve ease of use as a python package, I added the output option on make_cloud, which allows returning the point cloud array for direct use in Dedenser.downsample(). The array is still saved to file.

minimal example:

from dedenser import make_cloud, Dedenser

# define random seed
SEED = 21

# create chemical point cloud
point_clouds = make_cloud(
    path=path_to_file, 
    path_out=path_out,
    position=5,
    heady=True,
    random_state=SEED, # new
    output=True # new
)

# run downsampling
dd = Dedenser(
    data=point_clouds,
    target=0.35,
    random_seed=SEED
)
out_cloud = dd.downsample()