Remember if a page has previously been matched

This is a small step towards preventing false positives: remember if a
page has previously been matched to a ward.

This means that if we match "broadheath" and then later try to find
"heath" in the same PDF, we'll never try to match it against the page
containing "broadheath".

This has the other advantage of being able to list unlatched pages, that
might come in handy for debugging later.

ynr/apps/sopn_parsing/helpers/pdf_helpers.py

@@ -53,21 +53,26 @@ def parse_pages(self):
     def get_pages_by_ward_name(self, ward):
         ward = clean_text(ward)
         matched_pages = []
-        for page in self.pages:
+        for page in self.unmatched_pages():
             if page.is_top_page:
                 if matched_pages:
                     return matched_pages
                 search_text = clean_text(page.get_page_heading())
                 wards = ward.split("/")
                 for ward in wards:
                     if ward in search_text:
+                        page.matched = ward
                         matched_pages.append(page)
             else:
                 if matched_pages:
+                    page.matched = ward
                     matched_pages.append(page)
         if matched_pages:
             return matched_pages
 
+    def unmatched_pages(self):
+        return [p for p in self.pages if not p.matched]
+
 
 class SOPNPageText:
     """
@@ -79,6 +84,7 @@ def __init__(self, page_number, text):
         self.raw_text = text
         self.text = clean_text(text)
         self.is_top_page = True
+        self.matched = None
 
     def get_page_heading_set(self):
         """