fix: prefer PDF in (+refactor) materials_document view (#8136)

* fix: prefer PDF in materials_document(); refactor

* style: Black

* refactor: split urls entry into simpler pair

It seems Django cannot reverse a URL pattern that uses "|" to
combine options when one inclues a named variable and the other
does not.

* test: test extension choice

* fix: fix test failures

* refactor: get rid of io.open()

* refactor: reunite url patterns

Adds option for a trailing "/" when using an extension,
which was not previously supported but should be somewhere
between harmless and a feature.

ietf/meeting/tests_views.py

@@ -125,8 +125,12 @@ class BaseMeetingTestCase(TestCase):
         settings.MEETINGHOST_LOGO_PATH = self.saved_meetinghost_logo_path
         super().tearDown()
 
-    def write_materials_file(self, meeting, doc, content, charset="utf-8"):
+    def write_materials_file(self, meeting, doc, content, charset="utf-8", with_ext=None):
-        path = os.path.join(self.materials_dir, "%s/%s/%s" % (meeting.number, doc.type_id, doc.uploaded_filename))
+        if with_ext is None:
+            filename = doc.uploaded_filename
+        else:
+            filename = Path(doc.uploaded_filename).with_suffix(with_ext)
+        path = os.path.join(self.materials_dir, "%s/%s/%s" % (meeting.number, doc.type_id, filename))
 
         dirname = os.path.dirname(path)
         if not os.path.exists(dirname):
@@ -753,7 +757,56 @@ class MeetingTests(BaseMeetingTestCase):
         )
         self.assertEqual(len(q(f'a[href^="{edit_url}#session"]')), 1, f'Link to session_details page for {acro}')
 
+    def test_materials_document_extension_choice(self):
+        def _url(**kwargs):
+            return urlreverse("ietf.meeting.views.materials_document", kwargs=kwargs)
 
+        presentation = SessionPresentationFactory(
+            document__rev="00",
+            document__name="slides-whatever",
+            document__uploaded_filename="slides-whatever-00.txt",
+            document__type_id="slides",
+            document__states=(("reuse_policy", "single"),)
+        )
+        session = presentation.session
+        meeting = session.meeting
+        # This is not a realistic set of files to exist, but is useful for testing. Normally,
+        # we'd have _either_ txt, pdf, or pptx + pdf.
+        self.write_materials_file(meeting, presentation.document, "Hi I'm a txt", with_ext=".txt")
+        self.write_materials_file(meeting, presentation.document, "Hi I'm a pptx", with_ext=".pptx")
+
+        # with no rev, prefers the uploaded_filename
+        r = self.client.get(_url(document="slides-whatever", num=meeting.number))  # no rev
+        self.assertEqual(r.status_code, 200)
+        self.assertEqual(r.content.decode(), "Hi I'm a txt")
+        
+        # with a rev, prefers pptx because it comes first alphabetically
+        r = self.client.get(_url(document="slides-whatever-00", num=meeting.number))
+        self.assertEqual(r.status_code, 200)
+        self.assertEqual(r.content.decode(), "Hi I'm a pptx")
+
+        # now create a pdf
+        self.write_materials_file(meeting, presentation.document, "Hi I'm a pdf", with_ext=".pdf")
+
+        # with no rev, still prefers uploaded_filename
+        r = self.client.get(_url(document="slides-whatever", num=meeting.number))  # no rev
+        self.assertEqual(r.status_code, 200)
+        self.assertEqual(r.content.decode(), "Hi I'm a txt")
+
+        # pdf should be preferred with a rev
+        r = self.client.get(_url(document="slides-whatever-00", num=meeting.number))
+        self.assertEqual(r.status_code, 200)
+        self.assertEqual(r.content.decode(), "Hi I'm a pdf")
+        
+        # and explicit extensions should, of course, be respected
+        for ext in ["pdf", "pptx", "txt"]:
+            r = self.client.get(_url(document="slides-whatever-00", num=meeting.number, ext=f".{ext}"))
+            self.assertEqual(r.status_code, 200)
+            self.assertEqual(r.content.decode(), f"Hi I'm a {ext}")
+        
+        # and 404 should come up if the ext is not found
+        r = self.client.get(_url(document="slides-whatever-00", num=meeting.number, ext=".docx"))
+        self.assertEqual(r.status_code, 404)
 
     def test_materials_editable_groups(self):
         meeting = make_meeting_test_data()

ietf/meeting/urls.py

@@ -83,7 +83,7 @@ type_ietf_only_patterns_id_optional = [
     url(r'^week-view(?:.html)?/?$', AgendaRedirectView.as_view(pattern_name='agenda', permanent=True)),
     url(r'^materials(?:.html)?/?$', views.materials),
     url(r'^request_minutes/?$', views.request_minutes),
-    url(r'^materials/%(document)s((?P<ext>\.[a-z0-9]+)|/)?$' % settings.URL_REGEXPS, views.materials_document),
+    url(r'^materials/%(document)s(?P<ext>\.[a-z0-9]+)?/?$' % settings.URL_REGEXPS, views.materials_document),
     url(r'^session/?$', views.materials_editable_groups),
     url(r'^proceedings(?:.html)?/?$', views.proceedings),
     url(r'^proceedings(?:.html)?/finalize/?$', views.finalize_proceedings),

ietf/meeting/views.py

@@ -4,7 +4,6 @@
 
 import csv
 import datetime
-import glob
 import io
 import itertools
 import json
@@ -20,6 +19,7 @@ from calendar import timegm
 from collections import OrderedDict, Counter, deque, defaultdict, namedtuple
 from functools import partialmethod
 import jsonschema
+from pathlib import Path
 from urllib.parse import parse_qs, unquote, urlencode, urlsplit, urlunsplit
 from tempfile import mkstemp
 from wsgiref.handlers import format_date_time
@@ -250,7 +250,14 @@ def _get_materials_doc(meeting, name):
 
 @cache_page(1 * 60)
 def materials_document(request, document, num=None, ext=None):
-    meeting=get_meeting(num,type_in=['ietf','interim'])
+    """Materials document view
+
+    :param request: Django request
+    :param document: Name of document without an extension
+    :param num: meeting number
+    :param ext: extension including preceding '.'
+    """
+    meeting = get_meeting(num, type_in=["ietf", "interim"])
     num = meeting.number
     try:
         doc, rev = _get_materials_doc(meeting=meeting, name=document)
@@ -258,32 +265,34 @@ def materials_document(request, document, num=None, ext=None):
         raise Http404("No such document for meeting %s" % num)
 
     if not rev:
-        filename = doc.get_file_name()
+        filename = Path(doc.get_file_name())
     else:
-        filename = os.path.join(doc.get_file_path(), document)
+        filename = Path(doc.get_file_path()) / document
     if ext:
-        if not filename.endswith(ext):
+        filename = filename.with_suffix(ext)
-            name, _ = os.path.splitext(filename)
+    elif filename.suffix == "":
-            filename = name + ext
+        # If we don't already have an extension, try to add one
+        ext_choices = {
+            # Construct a map from suffix to full filename
+            fn.suffix: fn
+            for fn in sorted(filename.parent.glob(filename.stem + ".*"))
+        }
+        if len(ext_choices) > 0:
+            if ".pdf" in ext_choices:
+                filename = ext_choices[".pdf"]
             else:
-        filenames = glob.glob(filename+'.*')
+                filename = list(ext_choices.values())[0]
-        if filenames:
+    if not filename.exists():
-            filename = filenames[0]
+        raise Http404(f"File not found: {filename}")
-    _, basename = os.path.split(filename)
-    if not os.path.exists(filename):
-        raise Http404("File not found: %s" % filename)
 
     old_proceedings_format = meeting.number.isdigit() and int(meeting.number) <= 96
     if settings.MEETING_MATERIALS_SERVE_LOCALLY or old_proceedings_format:
-        with io.open(filename, 'rb') as file:
+        bytes = filename.read_bytes()
-            bytes = file.read()
-        
         mtype, chset = get_mime_type(bytes)
         content_type = "%s; charset=%s" % (mtype, chset)
 
-        file_ext = os.path.splitext(filename)
+        if filename.suffix == ".md" and mtype == "text/plain":
-        if len(file_ext) == 2 and file_ext[1] == '.md' and mtype == 'text/plain':
+            sorted_accept = sort_accept_tuple(request.META.get("HTTP_ACCEPT"))
-            sorted_accept = sort_accept_tuple(request.META.get('HTTP_ACCEPT'))
             for atype in sorted_accept:
                 if atype[0] == "text/markdown":
                     content_type = content_type.replace("plain", "markdown", 1)
@@ -293,7 +302,7 @@ def materials_document(request, document, num=None, ext=None):
                         "minimal.html",
                         {
                             "content": markdown.markdown(bytes.decode(encoding=chset)),
-                            "title": basename,
+                            "title": filename.name,
                         },
                     )
                     content_type = content_type.replace("plain", "html", 1)
@@ -302,11 +311,12 @@ def materials_document(request, document, num=None, ext=None):
                     break
 
         response = HttpResponse(bytes, content_type=content_type)
-        response['Content-Disposition'] = 'inline; filename="%s"' % basename
+        response["Content-Disposition"] = f'inline; filename="{filename.name}"'
         return response
     else:
         return HttpResponseRedirect(redirect_to=doc.get_href(meeting=meeting))
 
+
 @login_required
 def materials_editable_groups(request, num=None):
     meeting = get_meeting(num)