Chi_square_check for discrete distribution fix (apache#13543)

* check for bucket instead of index

* enumerate instead of range(len())

* count instead of sum to solve attribute error

* revert to sum

* seperate discrete and continuous

* Trigger CI

python/mxnet/test_utils.py

@@ -1911,12 +1911,15 @@ def chi_square_check(generator, buckets, probs, nsamples=1000000):
     if continuous_dist:
         sample_bucket_ids = np.searchsorted(buckets_npy, samples, side='right')
     else:
-        sample_bucket_ids = samples
+        sample_bucket_ids = np.array(samples)
     if continuous_dist:
         sample_bucket_ids = sample_bucket_ids // 2
     obs_freq = np.zeros(shape=len(buckets), dtype=np.int)
-    for i in range(len(buckets)):
-        obs_freq[i] = (sample_bucket_ids == i).sum()
+    for i, _ in enumerate(buckets):
+        if continuous_dist:
+            obs_freq[i] = (sample_bucket_ids == i).sum()
+        else:
+            obs_freq[i] = (sample_bucket_ids == buckets[i]).sum()
     _, p = ss.chisquare(f_obs=obs_freq, f_exp=expected_freq)
     return p, obs_freq, expected_freq