data_reader.py

import sys
import re
USAGE = """
    python data_reader.py <input_file>
"""

"""
File format:
<pattern>
<freq>3</freq>
<score> 36.242 </score><what>sleek</what>
<where>245 380 576</where>
</pattern>
<pattern>
<freq>3</freq>
<score> 36.242 </score><what>sleek</what>
<where>245 380 576</where>
</pattern>
<pattern>
<freq>3</freq>
<score> 36.242 </score><what>sleek</what>
<where>245 380 576</where>
</pattern>
<pattern>
<freq>3</freq>
<score> 36.242 </score><what>sleek</what>
<where>245 380 576</where>
</pattern>
...
"""

if len(sys.argv) < 2:
    print(USAGE)
    exit(0)


def main(input_file):
    records  = []
    with open(input_file, 'r') as f:
        while True:
            line = f.readline()
            # Got EOF
            if not line:
                break
            line = line.strip()
            if line.startswith("<pattern>"):
                freq = f.readline().strip()
                assert freq.startswith("<freq>")
                freq = int(re.sub(r'<[^>]+>', '', freq))
                score_and_what = f.readline().strip()
                assert score_and_what.startswith("<score>")
                score, what = score_and_what.split("</score>")
                what = re.sub(r'<[^>]+>', '', what).strip()
                score = re.sub(r'<[^>]+>', '', score)
                score = float(score)
                where = f.readline().strip()
                assert where.startswith("<where>")
                where = [int(x) for x in re.sub(r'<[^>]+>', '', where).split()]
                # skip </pattern>
                f.readline()
                records.append((freq, score, what, where))
    return records

if __name__ == '__main__':
    records = main(sys.argv[1])
    print(records)