Adding pandarallel library (#1005)

* Adding pandarallel library

Pandarallel allows to execute pandas apply method in parallel, which allows to do data preprocessing faster and easier. This is very he;pful in kernel only competiotions.

See:

https://github.com/nalepae/pandarallel

* Adding test to pandarallel

Adding a simple test

Author: manuelsh

Dockerfile

@@ -427,6 +427,7 @@ RUN pip install flashtext && \
     pip install jax==0.2.12 jaxlib==0.1.64 && \
     # ipympl adds interactive widget support for matplotlib
     pip install ipympl==0.7.0 && \
+    pip install pandarallel && \
     /tmp/clean-layer.sh
 
 # Download base easyocr models.

tests/test_pandarralel.py

@@ -0,0 +1,11 @@
+import unittest
+
+import pandas as pd
+from pandarallel import pandarallel
+
+pandarallel.initialize()
+
+class TestPandarallel(unittest.TestCase):    
+    def test_pandarallel(self):
+        data = pd.read_csv("/input/tests/data/train.csv")
+        data['label_converted'] = data['label'].parallel_apply(lambda x: x+1)