fix(guam): date regex was picking wrong values

Solves freelawproject#789

- Validated and improved regexes for date, docket and citation

juriscraper/opinions/united_states/territories/guam.py

@@ -4,11 +4,17 @@
 Author: mmantel
 History:
   2019-12-09: Created by mmantel
+  2024-01-25: updated by grossir
 """
 
 import re
 from datetime import date
+from typing import Optional
 
+from dateutil import parser
+from dateutil.parser import ParserError
+
+from juriscraper.AbstractSite import logger
 from juriscraper.OpinionSiteLinear import OpinionSiteLinear
 
 
@@ -17,50 +23,90 @@ def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         self.court_id = self.__module__
         self.url = "https://guamcourts.org/Supreme-Court-Opinions/Supreme-Court-Opinions.asp"
+        self.status = "Published"
+
         self._year = date.today().year
         self.parameters = {"Year": str(self._year)}
         self.method = "POST"
-        self.status = "Published"
 
         # The year dropdown goes back to 1990, but the Court wasn't
         # created until 1996 and there are no opinions posted for
         # prior years.
         self.back_scrape_iterable = range(1996, self._year)
 
-    def _process_html(self):
-        year = 2019 if self.test_mode_enabled() else self._year
-        middle_of_year = "July 2, %d" % year
+    def _process_html(self) -> None:
+        """Process HTML into case objects
+
+        Some docket numbers are a consolidation of other dockets
+        For example: "CVA12-018 (consolidated with CVA12-030)"
+        Deleting the date and citation from the free text allows us
+        to catch these dockets name
 
-        for table in self.html.xpath(
-            '//a[@id="Opinion"]/following-sibling::table'
-        ):
+        :return: None
+        """
+        row_xpath = '//a[@id="Opinion"]/following-sibling::table'
+        for table in self.html.xpath(row_xpath):
             text = table.xpath(".//td/text()")[0]
-            # Text is typically something like this:
-            # CVA16-016, 2018 Guam 8, July 26, 2018
-            # But formatting varies from year to year, and some elements
-            # are not provided for some cases.
 
-            date_match = re.search(r"[A-Za-z]+\.?\s+[0-9]+,\s+[0-9]+", text)
-            docket_match = re.search(r"[A-Z]+[0-9]+\-+[0-9]+", text)
-            neutral_citation_match = re.search(r"[0-9]+\s+Guam\s+[0-9]+", text)
+            date = self.find_date(text)
+            if not date:
+                logger.error(
+                    "Could not find date for record in guam scraper Year: %s",
+                    self._year,
+                )
+                continue
+
+            # Seen formats: 2021-Guam 3, 2021 Guam 29, 2020 Guam15,
+            # Edge cases which will be left empty: "Guam 7", "014 Guam 31"
+            citation_match = re.search(r"\d{4}[\s-]*Guam[\s-]*\d{1,2}", text)
+            if citation_match:
+                text = text.replace(citation_match.group(0), " ")
+                citation = citation_match.group(0)
+            else:
+                citation = ""
+
+            name = table.xpath(".//a/text()")[0]
+            docket = (
+                text.replace(date, "").replace(" filed ", "").strip(" .,\r\n")
+            )
+            # If the docket is not in the free text, sometimes it is at the end
+            # of the case name. Sometimes, it does not exist
+            if not docket:
+                docket_match = re.search(r"[A-Z]{3}\d{2}-\d{3}", name)
+                if docket_match:
+                    docket = docket_match.group(0)
 
             self.cases.append(
                 {
-                    "date": date_match.group(0)
-                    if date_match
-                    else middle_of_year,
-                    "date_filed_is_approximate": date_match is None,
-                    "docket": docket_match.group(0) if docket_match else "",
-                    "name": table.xpath(".//a/text()")[0],
-                    "citation": neutral_citation_match.group(0)
-                    if neutral_citation_match
-                    else "",
                     "url": table.xpath(".//a/@href")[0],
+                    "name": name,
+                    "docket": docket,
+                    "date": date,
+                    "citation": citation,
                 }
             )
 
-    def _download_backwards(self, year):
+    def find_date(self, text: str) -> Optional[str]:
+        """Find dates on text, and validate that they are indeed dates
+        Sometimes the regex will pick a part of the string that is not a date
+
+        :param text: free text with docket, date and citation info in varying order
+        :return: validated date or None
+        """
+        # Seen formats: "12-28-2023", "October 11, 2023", "Nov. 29, 2023"
+        date_pattern = r"([JFMASONDa-z.]+|\d{1,4})[\s-]+\d{1,2}[,\s-]+\d{2,4}"
+        for date_match in re.finditer(date_pattern, text):
+            try:
+                parser.parse(date_match.group())
+                return date_match.group()
+            except ParserError:
+                pass
+
+    def _download_backwards(self, year: int) -> None:
+        """Sets up the download of past records
+
+        :param year: search filter for the page
+        :return: None
+        """
         self._year = year
         self.parameters = {"Year": str(year)}
-        self.html = self._download()
-        self._process_html()