This script is useful for reconstruction of a document from individual sentences when those have been modified (cleaned). It relies on levenshtein distance of substrings.
It takes two files as an input: original document with arbitrary doc_IDs and file with individual modified sentences. Script returns sentences with original doc_ids, sentence number in document, and sentence.
Assumptions: Sentences are in the same order as in original documents. But they can be modified (capitalization, punctuation, removal of words, etc).