Merge pull request #5087 from openstates/usa-more-doc-note-trim

USA: Trim more bill document note property strings
openstates · Nov 16, 2024 · 59f35b6 · 59f35b6
2 parents bb7e0da + eb892d5
commit 59f35b6
Showing 1 changed file with 5 additions and 3 deletions.
diff --git a/scrapers/usa/bills.py b/scrapers/usa/bills.py
@@ -441,7 +441,7 @@ def scrape_amendments(self, bill, xml, session, chamber, bill_id):
                         if not amdt_url.startswith("http"):
                             continue
                         bill.add_document_link(
-                            note=amdt_name,
+                            note=amdt_name[:300],
                             url=amdt_url,
                             media_type="application/pdf",
                         )
@@ -453,7 +453,7 @@ def scrape_amendments(self, bill, xml, session, chamber, bill_id):
     def scrape_cbo(self, bill, xml):
         for row in xml.findall("bill/cboCostEstimates/item"):
             bill.add_document_link(
-                note=f"CBO: {self.get_xpath(row, 'title')}",
+                note=f"CBO: {self.get_xpath(row, 'title')}"[:300],
                 url=self.get_xpath(row, "url"),
                 media_type="text/html",
             )
@@ -468,7 +468,9 @@ def scrape_committee_reports(self, bill, xml):
 
             url = f"https://www.congress.gov/{match.group('session')}/crpt/{match.group('chamber').lower()}rpt{match.group('num')}/CRPT-{match.group('session')}{match.group('chamber').lower()}rpt{match.group('num')}.pdf"
 
-            bill.add_document_link(note=report, url=url, media_type="application/pdf")
+            bill.add_document_link(
+                note=report[:300], url=url, media_type="application/pdf"
+            )
 
     def scrape_cosponsors(self, bill, xml):
         all_sponsors = []