Similarities

Open up compare. Suffice it to say that file’s name doesn’t end in .py, even though the file contains a program written in Python. But that’s okay! Notice the "shebang" atop the file:

That line tells a computer to interpret (i.e., run) the program using python3 (aka python on CS50 IDE), an interpreter that understands Python 3.

Notice how the file defines a function called main and calls that function toward the bottom of the file. Defining main isn’t strictly necessary in Python, but it is necessary to define functions before you call them. Accordingly, because main calls a function called positive, and because we wanted to keep the "main" part of this program atop the file, it made sense to implement main as a function as well. That way, main doesn’t get called until the bottom of the file (after positive has been implemented), even though main is implemented atop the file.

No need to understand each of the lines in compare, but notice, per its comments, what it does overall: it parses its command-line arguments, reads two files into variables as strings, and compares those strings, and then prints a list of similarities. The strings themselves are compared in one of three ways, as specified by a command-line argument: line by line, sentence by sentence, or substring by substring.

Open up requirements.txt (without changing it, though you can later if you’d like). This file specifies the libraries, one per line, on which all of this functionality depends.

Implement lines in such a way that, given two strings, a and b, it returns a list of the lines that are, identically, in both a and b. The list should not contain any duplicates. Assume that lines in a and b will be be separated by \n, but the strings in the returned list should not end in \n. If both a and b contain one or more blank lines (i.e., a \n immediately preceded by no other characters), the returned list should include an empty string (i.e., "").

Implement sentences in such a way that, given two strings, a and b, it returns a list of the unique English sentences that are, identically, present in both a and b. The list should not contain any duplicates. Use sent_tokenize from the Natural Language Toolkit to "tokenize" (i.e., separate) each string into a list of sentences. It can be imported with:

Per its documentation, sent_tokenize, given a str as input, returns a list of English sentences therein. It assumes that its input is indeed English text (and not, e.g., code, which might coincidentally have periods too).

Implement substrings in such a way that, given two strings, a and b, and an integer, n, it returns a list of all substrings of length n that are, identically, present in both a and b. The list should not contain any duplicates.

Recall that a substring of length n of some string is just a sequence of n characters from that string. For instance, if n is 2 and the string is Yale, there are three possible substrings of length 2: Ya, al, and le. Meanwhile, if n is 1 and the string is Harvard, there are seven possible substrings of length 1: H, a, r, v, a, r, and d. But once we eliminate duplicates, there are only five unique substrings: H, a, r, v, and d.