• extraction
  • file

A common approach to this is using a state machine that reads the text until the <START> marker is encountered, then starts a “recording mode”, and extracts the text until the <END> marker is encountered. This process can repeat if multiple sections may appear in the file and have to be extracted.

inRecordingMode = False
for line in file:
    if not inRecordingMode:
        if line.startswith('<START>'):
            inRecordingMode = True
    elif line.startswith('<END>'):
        inRecordingMode = False
    else:
        yield line

For simple cases, this could also be solved with a regular expression.