NaN checks: use math.isnan() to verify/test for NaN values

nlp_profiler/grammar_quality_check.py

@@ -1,6 +1,8 @@
 import language_tool_python
+
 language_tool = language_tool_python.LanguageTool('en-GB')
 import pandas as pd
+import math
 
 from nlp_profiler.constants import NOT_APPLICABLE, NaN, DEFAULT_PARALLEL_METHOD, \
     GRAMMAR_CHECK_SCORE_COL, GRAMMAR_CHECK_COL
@@ -31,13 +33,13 @@ def grammar_check_score(text: str) -> int:
     return len(matches)
 
 
-def grammar_quality(score: float) -> str:
-    if score is NaN:
+def grammar_quality(score: int) -> str:
+    if math.isnan(score):
         return NOT_APPLICABLE
 
     if score == 1:
         return "1 issue"
     elif score > 1:
-        return f"{score} issues"
+        return f"{int(score)} issues"
 
     return "No issues"

nlp_profiler/sentiment_subjectivity.py

@@ -1,6 +1,7 @@
 from textblob import TextBlob
 
 from nlp_profiler.constants import NOT_APPLICABLE, NaN
+import math
 
 
 ### Sentiment Subjectivity
@@ -31,7 +32,7 @@ def sentiment_subjectivity_summarised(subjectivity: str) -> str:
 
 
 def sentiment_subjectivity(score: float) -> str:
-    if score is NaN:
+    if math.isnan(score):
         return NOT_APPLICABLE
 
     score = float(score) * 100

nlp_profiler/spelling_quality_check.py

@@ -4,6 +4,7 @@
 from joblib import Memory
 from nltk.tokenize import word_tokenize
 from textblob import Word
+import math
 
 from nlp_profiler.constants import \
     DEFAULT_PARALLEL_METHOD
@@ -91,7 +92,7 @@ def actual_spell_check(each_word: str) -> str:  # pragma: no cover
 
 
 def spelling_quality(score: float) -> str:
-    if score is NaN:
+    if math.isnan(score):
         return NOT_APPLICABLE
 
     score = float(score) * 100

tests/high_level/test_sentiment_polarity.py

@@ -40,7 +40,7 @@ def test_given_a_text_when_sentiment_analysis_is_applied_then_sentiment_analysis
     # given, when
     actual_score = sentiment_polarity_score(text)
     # then
-    if expected_polarity_score is NaN:
+    if math.isnan(expected_polarity_score):
         assert actual_score is expected_polarity_score
     else:
         assert math.isclose(expected_polarity_score, actual_score,