[WIP] Preliminary version of script to extract dates from emails

1 files changed, 181 insertions, 0 deletions

diff --git a/event.py b/event.py
new file mode 100755
index 0000000..978d0bd
--- /dev/null
+++ b/event.py
@@ -0,0 +1,181 @@
+from calendar import day_name, month_name
+import re
+from datetime import datetime, timedelta, time
+import sys
+import email
+from email.utils import parsedate
+from itertools import takewhile, dropwhile
+
+days = [day.lower() for day in day_name]
+days_abbr = [day.lower()[:3] for day in day_name]
+days_abbr2 = [day.lower()[:4] for day in day_name]
+months = [month.lower() for month in month_name][1:]
+months_abbr = [month.lower()[:3] for month in month_name][1:]
+
+
+def clean(text):
+    lines = [line.strip() for line in text.splitlines()]
+    lines = list(takewhile(lambda x: not (x.startswith("---") or x == "--"),
+                           lines))
+    lines.reverse()
+    lines = list(dropwhile(lambda x: not x, lines))
+    if lines[0].startswith(">"):
+        lines = list(dropwhile(lambda x: x.startswith(">"), lines))
+        lines = list(dropwhile(lambda x: not x, lines))
+        lines = list(dropwhile(lambda x: x, lines))
+    lines.reverse()
+    lines = [line.strip() for line in lines if line]
+    text = " ".join(line for line in lines if not line.startswith(">"))
+    text = text.translate(None, ",.!?*\"()").lower()
+    text = " ".join(text.split())
+    return text
+
+
+def parse_dates(text, ref=None):
+    if not ref:
+        ref = datetime.now().date()
+
+    l = []
+
+    def sub(match):
+        groups = match.groupdict()
+        date = ref.replace(day=int(groups["day"]), month=int(groups["month"]))
+        if groups["year"]:
+            date = date.replace(year=int(groups["year"]))
+        l.append((date, match.span()))
+        return "#" * (match.end() - match.start())
+
+    text = re.sub(r"(?P<month>\d{1,2})/(?P<day>\d{2})"
+                  "(/(?P<year>\d{2,4}))?", sub, text)
+    text = re.sub(r"(?P<year>\d{4})-(?P<month>\d{2})"
+                  "-(?P<day>\d{2})", sub, text)
+
+    def sub(match):
+        groups = match.groupdict()
+        try:
+            idx = months.index(groups["month"])
+        except ValueError:
+            idx = months_abbr.index(groups["month"])
+        date = ref.replace(day=int(groups["day_num"]),
+                           month=idx + 1)
+        l.append((date, match.span()))
+        return "#" * (match.end() - match.start())
+
+    text = re.sub(r"((?P<day>" + "|".join(days) + "|".join(days_abbr)
+                  + "|".join(days_abbr2) + ") )?"
+                  + r"(?P<month>" + "|".join(months)
+                  + "|".join(months_abbr) + ")"
+                  + r" (?P<day_num>\d{1,2})(st|rd|th|nd)?( \d{4})?",
+                  sub, text)
+
+    def sub(match):
+        delta = timedelta((days.index(match.group("day")) - ref.weekday()) % 7)
+        l.append((ref + delta, match.span()))
+        return "#" * (match.end() - match.start())
+
+    text = re.sub(r"(?P<day>" + "|".join(days) + ")", sub, text)
+
+    def sub(match):
+        delta = timedelta(1)
+        l.append((ref + delta, match.span()))
+        return "#" * (match.end() - match.start())
+
+    text = re.sub(r"tomorrow", sub, text)
+
+    l = [d for d in l if d[0] >= ref]
+
+    if not l:
+        l.append((ref, (0, 0)))
+
+    return text, l
+
+
+def create_time(hours, mins, hd):
+    if not hd:
+        if 8 <= hours <= 11:
+            hd = "am"
+        elif 1 <= hours < 8:
+            hd = "pm"
+        elif hours == 12:
+            hd = "pm"
+        elif hours > 12:
+            hours -= 12
+            hd = "pm"
+    hours = hours + (hd == "pm" and hours < 12) * 12
+    try:
+        return time(hours, mins)
+    except ValueError:
+        return None
+
+
+def parse_times(text):
+    l = []
+
+    def sub(match):
+        groups = match.groupdict()
+        groups["min"] = int(groups["min"]) if groups["min"] else 0
+        t1 = create_time(int(groups["hour"]), groups["min"], groups["hd"])
+        groups["min2"] = int(groups["min2"]) if groups["min2"] else 0
+        t2 = create_time(int(groups["hour2"]), groups["min2"], groups["hd2"])
+        if t1:
+            l.append((t1, match.span()))
+        if t2:
+            l.append((t2, match.span()))
+        return "#" * (match.end() - match.start())
+
+    text = re.sub(r"\b(?P<hour>\d{1,2})(:(?P<min>\d{2}))? ?(?P<hd>pm|am)? ?"
+                  "-{1,2} ?(?P<hour2>\d{1,2})(:(?P<min2>\d{2}))? ?"
+                  "(?P<hd2>pm|am)?\b", sub, text)
+
+    def sub(match):
+        if not match.end() - match.start() >= 3:
+            return text[match.start():match.end()]
+        groups = match.groupdict()
+        groups["min"] = int(groups["min"]) if groups["min"] else 0
+        t = create_time(int(groups["hour"]), groups["min"], groups["hd"])
+        if t:
+            l.append((t, match.span()))
+        return "#" * (match.end() - match.start() - 2)
+
+    text = re.sub(r"\b(?P<hour>\d{1,2})(:(?P<min>\d{2}))?( ?(?P<hd>pm|am))?\b",
+                  sub, text)
+    return text, l
+
+
+def align(dates, times):
+    r = []
+    matched = []
+
+    def k(d):
+        if t[1][1] < d[1][0]:
+            return d[1][0] - t[1][1]
+        else:
+            return t[1][0] - d[1][1]
+
+    for t in times:
+        d = min(dates, key=k)
+        matched.append(d[0])
+        r.append(datetime.combine(d[0], t[0]))
+
+    unmatched = [d[0] for d in dates if d[0] not in matched]
+    return r, unmatched
+
+
+def parse_text(text, ref=None):
+    text, dates = parse_dates(text, ref)
+    text, times = parse_times(text)
+    m, u = align(dates, times)
+    print "\n".join(d.isoformat() for d in m)
+
+
+def parse_email(fp):
+    mail = email.message_from_file(fp)
+    values = parsedate(mail["date"])
+    date = datetime(*values[:6]).date()
+    text = clean(mail.get_payload())
+    parse_text(text, ref=date)
+
+
+if __name__ == "__main__":
+    parse_email(sys.stdin)
+    a = raw_input()