How do I create a pyspark program that extracts all the N-grams from a csv file?

So I use this

DF = spark.read.format('csv').option('header', "true").load("file.csv")

I then make the DF an RDD by doing

RDD_DF = DF.rdd.map(lambda x:x[0])

I only want the first column. Now how do I create an N-gram filter?

So after running the code above I get something like this

['This was a good time. Apples.',

'dogs. mason prop',

There are many cats']

I wanna get a code that selects all n-grams so all the unigrams, -bigrams, trigrams, quadgrams etc. so if the code collected all unigrams it would return

Apples

dogs

Bigrams would return

mason prop

etc.