Merge pull request #15 from t4guw/data-formatting

Complete data formatting

src/.gitignore

@@ -0,0 +1 @@
+__pycache__

src/data/statements.txt

@@ -1,4 +1,3 @@
-~~~STATEMENT_START~~~
 1
 Given an array of integers, return indices of the two numbers such that they add up to a specific target.
 You may assume that each input would have exactly one solution, and you may not use the same element twice.

src/parser.py

@@ -0,0 +1,73 @@
+#!/usr/bin/env python
+
+'''
+Parses solutions.txt and statements.txt.
+Creates problem objects for each LeetCode problem that are serialized as JSON.
+Produces JSON file that is intended to be the input of the labeling functions
+5/27/2020
+'''
+
+import json
+from problem import Problem
+from itertools import islice
+
+OUTPUT_FILE="problems.json"
+PROBLEMS = {}
+
+'''
+Example solution:
+    ~~~START_SOLUTION~~~
+    <problem_number>
+    public int{} ... 
+    ....
+    }
+    ~~~END_SOLUTION~~~
+'''
+def process_solutions(filename):
+    with open(filename) as f:
+        while line := f.readline():
+            if "~~~START_SOLUTION~~~" in line:
+                continue
+            else:
+                number = int(line)
+                solution = read_section(f, "~~~END_SOLUTION~~~")
+                PROBLEMS[number].solutions.append(solution)
+
+
+'''
+Example statement:
+    ~~~STATEMENT_START~~~
+    1
+    Given an array of integers....
+'''
+def process_statements(filename):
+    with open(filename) as f:
+        while line := f.readline():
+            number = int(line)
+            statement = read_section(f, "~~~STATEMENT_START~~~")
+            PROBLEMS[number] = Problem(number, statement)
+
+
+def read_section(file, marker=''):
+    content = str()
+    while line := file.readline():
+        if marker in line:
+            return content
+        else:
+            content += line
+
+def create_output_file():
+    with open(OUTPUT_FILE, 'w') as outfile:
+        for k, v in PROBLEMS.items():
+            json.dump(v.serialize(), outfile)
+
+def main():
+    process_statements("data/statements.txt")
+    process_solutions("data/solutions.txt")
+    for k,v in PROBLEMS.items():
+        print(k, v.serialize())
+        break
+    create_output_file()
+
+if __name__ == "__main__":
+    main()

src/problem.py

@@ -0,0 +1,19 @@
+import json
+
+# Class for serializing custom objects into JSON. 
+# See https://docs.python.org/3/library/json.html#json.JSONEncoder.default
+class ComplexEncoder(json.JSONEncoder):
+    def default(self, obj):
+        if isinstance(obj, Problem):
+            return obj.__dict__
+        return json.JSONEncoder(self, obj)
+
+# Class that defines the properties of each Problem object
+class Problem(object):
+    def __init__(self, number, statement):
+        self.number = number
+        self.statement = statement
+        self.solutions = []
+
+    def serialize(self):
+        return json.dumps(self, cls=ComplexEncoder)