Stemming - Ingestion and Search handling

[Aksha1812]

Ingestion

happy and happiness stem to -> happi
personal and person stem to -> person 

→ payload: {doc:1 happy, 
            doc:2 happiness, 
            doc:3 happi,
            doc:4 personal,
            doc:5 person
            }
            
Prefix Trie path:                  ()
                                /       \
                           happ      person   
                          /     \          \
                      iness      y          al  
                  
Posting objects to be stored :

 happy : Posting {doc:1},
 happi : Posting {doc:3} ,
 happiness : Posting {doc:2}, 
 personal: Posting {doc:4}, 
 person: Posting {doc:5}, 

Stem Trie path:                    ()
                                /       \
                           happi      person   

Stem Targets to be stored :

happi -> {happy, happiness} 
person -> {personal}

(even 'happi' stems to 'happi' but mapping a word to itslef doesnt make sense)

Search

A non exact term search should return both original word and all stem parents of the stemmed root of original word

For instance for query FT.SEARCH testindex '@desc:happi'

Results should include the original word + All stem targets of word happi + Exact Stem root if exists

Other things added with this PR

• Addressed comments onGetLanguageString() function

[KarthikSubbarao]

For stemming, it would be good to consider a separate word iterator. This code should be re-used through the GetStemmedWordIterator and not embedded in each search operation

Also, as discussed, we only will need to use this StemmedWordIterator during the initialization of the search operations once the open PR is merged in