Drop query string from reference PDF URLs

This fixes #24
metanorma · Mar 19, 2022 · 71f5628 · 71f5628
1 parent 5ed8ce0
commit 71f5628
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/exe/bipm-fetch-cgpm b/exe/bipm-fetch-cgpm
@@ -36,7 +36,7 @@ FileUtils.rm_rf "#{BASE_DIR}/cgpm/meetings-en"
 
     binding.pry if date <= Date.parse("0000-01-01") || date >= Date.today
 
-    pdf = meeting.at_css('a.title-third[href*=".pdf"]')&.attr('href')
+    pdf = meeting.at_css('a.title-third[href*=".pdf"]')&.attr('href')&.split('?')&.first
 
     h = {
       "metadata" => {

diff --git a/lib/bipm/data/importer/common.rb b/lib/bipm/data/importer/common.rb
@@ -198,7 +198,7 @@ def parse_resolution res, res_id, date, type = :cgpm, lang = 'en'
           }
 
           if refs.length > 0
-            r["reference"] = res.uri.merge(refs.first.attr('href')).to_s
+            r["reference"] = res.uri.merge(refs.first.attr('href')).to_s.split('?').first
             name, page = refs.first.text.strip.split(/, p(?=[0-9])/)
             r["reference_name"] = name
             if page