getting error from "aorta_data_analysis.ipynb"

[btlantis]

Dear jingcheng-du,
I am trying to run "aorta_data_analysis.ipynb" on google colab. However, I am getting error when I run cell [18]. Here is the error I got as following:

TypeError Traceback (most recent call last)
in <cell line: 24>()
29 else:
30 idx, sim = get_similar_vectors(test_emebd[k][np.newaxis, ...], ref_cell_embeddings)
---> 31 pred = mode(y_train[idx], axis=0)
32 neighbors_list_gpt_v1.append(y_train[idx])
33 gt_list.append(gt)

1 frames
/usr/local/lib/python3.10/dist-packages/scipy/stats/_stats_py.py in mode(a, axis, nan_policy, keepdims)
509 "array was deprecated in SciPy 1.9.0 and removed in SciPy "
510 "1.11.0. Please consider np.unique.")
--> 511 raise TypeError(message)
512
513 if a.size == 0:

TypeError: Argument a is not recognized as numeric. Support for input that cannot be coerced to a numeric array was deprecated in SciPy 1.9.0 and removed in SciPy 1.11.0. Please consider np.unique.

PS: I only modify "N_TRUNC_GENE = 1000" into "N_TRUNC_GENE = 20" (try to reducing embedding process??)
Could you please help me with that?
Thank you so much!

Best Regards,
Ben

[ShengZhaoDE]

change
pred = mode(y_train[idx], axis=0)
to
vals,counts = np.unique(y_train[idx], return_counts=True) pred = vals[np.argmax(counts)]

works for me.

[SmileSpring1823]

Hey there,

I made the change. But I get "0.0" for the accuracy score.

Seems like the predictions just return the first letter:

print("gt")
print(gt_list[0:23])

print("pred")
print(pred_list[0:23])

gt
['Tcell', 'SMC1', 'NK', 'SMC1', 'MonoMaphDC', 'NK', 'Tcell', 'Tcell', 'Tcell', 'MonoMaphDC', 'Tcell', 'MonoMaphDC', 'MonoMaphDC', 'SMC2', 'Tcell', 'SMC1', 'Tcell', 'Tcell', 'MonoMaphDC', 'SMC1', 'SMC1', 'NK', 'Tcell']
pred
['T', 'S', 'T', 'S', 'M', 'M', 'T', 'T', 'T', 'M', 'T', 'M', 'M', 'S', 'T', 'S', 'T', 'T', 'M', 'S', 'S', 'T', 'T']

Any suggestions?

EDIT: You just need to remove the [0][0] at the append for pred

[HelloWorldLTY]

Hi, it seems that it is a problem related to the neighbor searching. I suggest installing this package directly from github rather than pypl:

https://github.com/nmslib/hnswlib