Code fix for WV Event Time Wrong

scrapers/wv/events.py

@@ -10,6 +10,11 @@
 class WVEventScraper(Scraper, LXMLMixin):
     verify = False
     _tz = pytz.timezone("US/Eastern")
+    # Matches a time of day such as "1:00 PM", "9 AM" or "10:30a.m."
+    _time_re = re.compile(
+        r"\d{1,2}(:\d{2})?\s*[ap]\.?m\.?",
+        re.IGNORECASE,
+    )
 
     def scrape(self):
         com_urls = [
@@ -48,8 +53,9 @@ def scrape_committee_page(self, url):
         page = self.lxmlize(url)
         page.make_links_absolute(url)
 
-        # if the page starts w/ a meeting
-        if page.xpath('//div[@id="wrapleftcol"]/h1'):
+        # if the page starts w/ a meeting. The h1 is always the committee title;
+        # a meeting's date/time lives in the h2, so use that to detect a meeting.
+        if page.xpath('//div[@id="wrapleftcol"]/h2'):
             yield from self.scrape_meeting_page(url)
 
         for row in page.xpath('//td/a[contains(@href, "agendas.cfm")]'):
@@ -59,7 +65,7 @@ def scrape_committee_page(self, url):
             # meeting pages show events going back years
             # so just grab this cal year and later
             when = row.xpath("text()")[0].strip()
-            when = when.split("-")[0]
+            when = self.strip_date_range(when)
             when = self.clean_date(when)
 
             # manual fix for un-yeared leap year meeting
@@ -79,12 +85,14 @@ def scrape_meeting_page(self, url):
         if page.xpath('//div[text()="Error"]'):
             return
 
-        if not page.xpath('//div[@id="wrapleftcol"]/h3'):
+        if not page.xpath('//div[@id="wrapleftcol"]/h2'):
             return
 
-        com = page.xpath('//div[@id="wrapleftcol"]/h3[1]/text()')[0].strip()
-        com = re.sub(r"[\s\-]+Agenda", "", com)
-        when = page.xpath('//div[@id="wrapleftcol"]/h1[1]/text()')[0].strip()
+        # The committee name is in the h1 (e.g. "Senate Finance Committee - Agenda")
+        # and the meeting date/time is in the h2 (e.g. "March 12, 2026, 3:00 PM").
+        com = page.xpath('//div[@id="wrapleftcol"]/h1[1]/text()')[0].strip()
+        com = re.sub(r"\s*-\s*Agenda\s*$", "", com).strip()
+        when = page.xpath('//div[@id="wrapleftcol"]/h2[1]/text()')[0].strip()
 
         if when == "test, test" or when == ",":
             # Ignore test page
@@ -106,7 +114,7 @@ def scrape_meeting_page(self, url):
         )
         when = re.sub(r",?\s+After Floor", "", when, flags=re.IGNORECASE)
 
-        when = when.split("-")[0]
+        when = self.strip_date_range(when)
         when = self.clean_date(when)
         when = dateutil.parser.parse(when)
         when = self._tz.localize(when)
@@ -171,11 +179,41 @@ def scrape_meeting_page(self, url):
 
         yield event
 
+    def strip_date_range(self, when):
+        """
+        Strip a trailing date range and keep only the start date.
+
+        Agenda dates are occasionally expressed as a range using a hyphen
+        surrounded by whitespace (e.g. "March 3, 2023 - March 4, 2023"), in
+        which case we only want the first date.
+
+        We must NOT split on a bare hyphen because numeric dates use hyphens
+        as separators (e.g. "1-14-14", "2-11-20"). Splitting those on "-"
+        would leave just the month component (e.g. "1"), which dateutil then
+        parses using today's month/year as defaults, producing wildly wrong
+        dates. Only treat a hyphen surrounded by whitespace as a range
+        delimiter.
+        """
+        return re.split(r"\s+-\s+", when)[0].strip()
+
     def clean_date(self, when):
-        # Remove all text after the third comma to make sure no extra text
-        # is included in the date. Required to correctly parse text like this:
-        # "Friday, March 3, 2023, Following wrap up of morning agenda"
-        when = ",".join(when.split(",")[:3])
+        """
+        Cleans the meeting date string by removing trailing non-date text while
+        preserving any valid meeting time.
+
+        The date occupies the first three comma-separated segments (weekday,
+        month/day, year). Some agenda pages include additional text such as
+        "Following wrap up of morning agenda", while others include the meeting
+        time (e.g. "1:00 PM") after the date. Preserve any segment containing a
+        valid time and discard other trailing text so the datetime is parsed
+        correctly.
+        """
+        segments = when.split(",")
+        kept = segments[:3]
+        for segment in segments[3:]:
+            if self._time_re.search(segment):
+                kept.append(segment)
+        when = ",".join(kept)
 
         removals = [
             r"(\d+|Thirty) (min\.|mins\.|minutes) After (.*)",