update: handle dodginess

blogs/migrations/0018_blog_dodginess_score.py

@@ -0,0 +1,18 @@
+# Generated by Django 3.1.14 on 2024-06-20 09:05
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('blogs', '0017_post_first_published_at'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='blog',
+            name='dodginess_score',
+            field=models.FloatField(default=0),
+        ),
+    ]

blogs/models.py

@@ -54,6 +54,7 @@ class Blog(models.Model):
     header_directive = models.TextField(blank=True)
     footer_directive = models.TextField(blank=True)
 
+    dodginess_score = models.FloatField(default=0)
     reviewed = models.BooleanField(default=False)
     ignored_date = models.DateTimeField(blank=True, null=True)
     to_review = models.BooleanField(default=False)
@@ -135,13 +136,26 @@ def generate_auth_token(self):
         self.auth_token = ''.join(random.choice(allowed_chars) for _ in range(30))
         self.save()
 
+    def determine_dodginess(self):
+        persistent_store = PersistentStore.load()
+        dodgy_term_count = 0
+        all_content = f"{self.title} {self.content}"
+
+        post = self.posts.first()
+        if post:
+            all_content += f"{post.title} {post.content}"
+
+        for term in persistent_store.highlight_terms:
+            dodgy_term_count += all_content.lower().count(term.lower())
+
+        self.dodginess_score = dodgy_term_count / len(all_content) *10000
+
     def save(self, *args, **kwargs):
-    #     # Create auth_token
-    #     if not self.auth_token:
-    #         allowed_chars = string.ascii_letters.replace('O', '').replace('l', '')
-    #         self.auth_token = ''.join(random.choice(allowed_chars) for _ in range(20))
         if self.user.settings.upgraded:
             self.reviewed = True
+
+        self.determine_dodginess()
+
         super(Blog, self).save(*args, **kwargs)
 
     def __str__(self):
@@ -200,7 +214,6 @@ def update_score(self):
                 else:
                     score = (log_of_upvotes) + ((seconds - 1577811600) / (14 * 86400))
                 self.score = score
-        self.save()
 
     def save(self, *args, **kwargs):
         self.slug = self.slug.lower()
@@ -216,6 +229,12 @@ def save(self, *args, **kwargs):
         if self.publish and self.first_published_at is None:
             self.first_published_at = self.published_date or timezone.now()
 
+        # Update the score for the discover feed
+        self.update_score()
+
+        # Save blog to trigger determine_dodginess
+        self.blog.save()
+
         super(Post, self).save(*args, **kwargs)
 
     def __str__(self):
@@ -231,8 +250,8 @@ def save(self, *args, **kwargs):
         # Save the Upvote instance
         super(Upvote, self).save(*args, **kwargs)
 
-        # Now that the Upvote is saved and counted, update the post score
-        self.post.update_score()
+        # Update the post score on post save
+        self.post.save()
 
     def __str__(self):
         return f"{self.created_date.strftime('%d %b %Y, %X')} - {self.hash_id} - {self.post}"

blogs/views/staff.py

@@ -132,38 +132,19 @@ def get_empty_blogs():
 
 def blogs_to_review():
     # Opted-in for review
-    to_review = Blog.objects.filter(reviewed=False, user__is_active=True, to_review=True)
+    to_review = Blog.objects.filter(reviewed=False, user__is_active=True, to_review=True).order_by('created_date')
 
     if to_review.count() < 1:
-        persistent_store = PersistentStore.load()
-        highlight_terms = persistent_store.highlight_terms
-
-        new_blogs = Blog.objects.filter(
+        to_review = Blog.objects.filter(
             reviewed=False,
             user__is_active=True,
             to_review=False,
+            dodginess_score__gt=5,
             ignored_date__isnull=True
-        ).prefetch_related('posts')
-
-        blog_ids_to_review = []
-
-        for blog in new_blogs:
-            term_count = 0
-            content = f"{blog.title} {blog.content}"
-
-            # post = blog.posts.first()
-            # if post:
-            #     content += f"{post.title} {post.content}"
-            for term in highlight_terms:
-                term_count += content.lower().count(term.lower())
-
-            if term_count >= 2:
-                blog_ids_to_review.append(blog.id)
-
-        to_review = Blog.objects.filter(id__in=blog_ids_to_review)
+        ).prefetch_related('posts').order_by('-dodginess_score')
 
 
-    return to_review.order_by('created_date')
+    return to_review
 
 
 @staff_member_required
@@ -177,7 +158,7 @@ def delete_empty(request):
 
 @staff_member_required
 def review_bulk(request):
-    blogs = blogs_to_review()[:20]
+    blogs = blogs_to_review()[:100]
     still_to_go = blogs.count()
     persistent_store = PersistentStore.load()
 

blogs/views/studio.py

@@ -245,9 +245,6 @@ def post(request, id, uid=None):
                 return post
             else:
                 post.save()
-
-                # Recalculate score to prevent future date spoofing
-                post.update_score()
 
                 if is_new:
                     # Self-upvote

templates/staff/review_bulk.html

@@ -39,6 +39,7 @@ <h3>Still to go: {{ still_to_go }}</h3>
 {% for blog in blogs %}
 <div id="blog-{{ blog.pk }}">
     <h3>{{ blog.title }}</h3>
+    <h4>Dodginess score: {{ blog.dodginess_score }}</h4>
     <a href="{{ blog.useful_domain }}" target="_blank">{{ blog.useful_domain }}</a>
     <br>
     <a href="mailto:{{ blog.user.email }}">{{ blog.user.email }}</a>