Calculate has_* fields before saving Messages

Before saving a Message object, call update_calculated_fields()
to set the has_attachment/has_image/has_link fields.

Note that the pre_save hook we added here does not get called
if you call bulk_create, hence the explicit call to
update_calculated_fields() in do_send_messages().

(imported from commit 1d60ae5908ef186aa5ff1e39277dbb2b765e60d4)

zerver/lib/actions.py

@@ -376,6 +376,7 @@ def do_send_messages(messages):
         message['active_recipients'] = filter(lambda user_profile: user_profile.is_active,
                                               message['recipients'])
         message['message'].maybe_render_content(None)
+        message['message'].update_calculated_fields()
 
     # Save the message receipts in the database
     user_message_flags = defaultdict(dict)

zerver/models.py

@@ -4,6 +4,7 @@ from django.db import models
 from django.conf import settings
 from django.contrib.auth.models import AbstractBaseUser, UserManager, \
     PermissionsMixin
+from django.dispatch import receiver
 from zerver.lib.cache import cache_with_key, flush_user_profile, flush_realm, \
     user_profile_by_id_cache_key, user_profile_by_email_cache_key, \
     generic_bulk_cached_fetch, cache_set, flush_stream, \
@@ -15,13 +16,14 @@ from zerver.lib.avatar import gravatar_hash, get_avatar_url
 from django.utils import timezone
 from django.contrib.sessions.models import Session
 from zerver.lib.timestamp import datetime_to_timestamp
-from django.db.models.signals import post_save, post_delete
+from django.db.models.signals import pre_save, post_save, post_delete
 from guardian.shortcuts import get_users_with_perms
 import zlib
 
 from bitfield import BitField
 from collections import defaultdict
 import pylibmc
+import re
 import ujson
 import logging
 
@@ -935,6 +937,31 @@ class Message(models.Model):
                                    'website', 'ios', 'android')) or \
                                    ('desktop app' in sending_client)
 
+    @staticmethod
+    def content_has_attachment(content):
+        return '/user_uploads/' in content
+
+    @staticmethod
+    def content_has_image(content):
+        return bool(re.search('/user_uploads/\S+\.(bmp|gif|jpg|jpeg|png|webp)', content, re.IGNORECASE))
+
+    @staticmethod
+    def content_has_link(content):
+        return 'http://' in content or 'https://' in content or '/user_uploads' in content
+
+    def update_calculated_fields(self):
+        # TODO: rendered_content could also be considered a calculated field
+        content = self.content
+        self.has_attachment = bool(Message.content_has_attachment(content))
+        self.has_image = bool(Message.content_has_image(content))
+        self.has_link = bool(Message.content_has_link(content))
+
+@receiver(pre_save, sender=Message)
+def pre_save_message(sender, **kwargs):
+    if kwargs['update_fields'] is None or "content" in kwargs['update_fields']:
+        message = kwargs['instance']
+        message.update_calculated_fields()
+
 class UserMessage(models.Model):
     user_profile = models.ForeignKey(UserProfile)
     message = models.ForeignKey(Message)

zerver/test_messages.py

@@ -4,6 +4,7 @@ from django.db.models import Q
 from sqlalchemy.sql import (
     and_, select, column, compiler
 )
+from django.test import TestCase
 from zerver.lib import bugdown
 from zerver.decorator import JsonableError
 from zerver.lib.test_runner import slow
@@ -1277,6 +1278,25 @@ class StarTests(AuthedTestCase):
         self.assertEqual(sent_message.message.content, content)
         self.assertFalse(sent_message.flags.starred)
 
+class AttachmentTest(TestCase):
+    def test_basics(self):
+        self.assertFalse(Message.content_has_attachment('whatever'))
+        self.assertFalse(Message.content_has_attachment('yo http://foo.com'))
+        self.assertTrue(Message.content_has_attachment('yo\n https://staging.zulip.com/user_uploads/'))
+        self.assertTrue(Message.content_has_attachment('yo\n /user_uploads/1/wEAnI-PEmVmCjo15xxNaQbnj/photo-10.jpg foo'))
+
+        self.assertFalse(Message.content_has_image('whatever'))
+        self.assertFalse(Message.content_has_image('yo http://foo.com'))
+        self.assertFalse(Message.content_has_image('yo\n /user_uploads/1/wEAnI-PEmVmCjo15xxNaQbnj/photo-10.pdf foo'))
+        for ext in [".bmp", ".gif", ".jpg", "jpeg", ".png", ".webp", ".JPG"]:
+            content = 'yo\n /user_uploads/1/wEAnI-PEmVmCjo15xxNaQbnj/photo-10.%s foo' % (ext,)
+            self.assertTrue(Message.content_has_image(content))
+
+        self.assertFalse(Message.content_has_link('whatever'))
+        self.assertTrue(Message.content_has_link('yo\n http://foo.com'))
+        self.assertTrue(Message.content_has_link('yo\n https://example.com?spam=1&eggs=2'))
+        self.assertTrue(Message.content_has_link('yo /user_uploads/1/wEAnI-PEmVmCjo15xxNaQbnj/photo-10.pdf foo'))
+
 class CheckMessageTest(AuthedTestCase):
     def test_basic_check_message_call(self):
         sender = get_user_profile_by_email('othello@zulip.com')