I figured out my problem. In the "re.sub" command, I was putting the "re.I" flag, which tells commands to ignore case, into the field that was expecting the count. Basically, I was calling the command without ignoring case, thus never triggering any matches.
Looking at the Stanford NLP parser again. Having a hard time understanding its output. Wondering if I need to add more clarifying text during my preprocessing stages. For example, when I change "8-9:30am" to "8 9:30am", should I throw in a a "to" or a "through" or "until" in there?
No comments:
Post a Comment