AttributeError: module 'cellrank' has no attribute 'kernels'

[jwalewski]

When I attempt to run cellrank in a conda enviornment that also contains velocyto (so that I can generate my count matrices and then proceed) I am running into an issue where cellrank claims it has no kernel module, despite me copying and pasting the vignette into my shell script.

# ------- 1. installing packages common to all vignettes -------
print("Step 2: cellrank preprocessing has begun")
import numpy as np
import sys
import pandas as pd
import cellrank as cr
import scvelo as scv
import scanpy as sc
import warnings
warnings.simplefilter("ignore", category=UserWarning)

## ------- 2. Configuring basic packages -------
scv.settings.verbosity = 3
sc.settings.set_figure_params("scvelo") #these may change as we wish
cr.settings.verbosity = 2
warnings.simplefilter("ignore", category=UserWarning)

# ------- 3. read input data from loom file generated from velocyto -------
adata = scv.read("<my_data>.loom", cache=True) 
# Loading in test data & confirming
scv.pl.proportions(adata)
adata

# ------- 4. Running scVelo -------
# Running scvelo
scv.tl.recover_dynamics(adata, n_jobs=8)
scv.tl.velocity(adata, mode="dynamical")

# ------ 5. main RNA Velocity Kernel --------
vk = cr.kernels.VelocityKernel(adata)
...

AttributeError: module 'cellrank' has no attribute 'kernels'
...

Any ideas about what could be going on here?