Update main.py

main.py

@@ -22,25 +22,27 @@ async def analyze_text(payload: TextPayload):
     doc = nlp(payload.text)
     entity_freq = Counter([ent.text.lower() for ent in doc.ents])
     
-    brand_safety = ['johnson & johnson', 'at&t', 'h&m']
+    # Brands with '&' or 'and' that should NOT be split
+    brand_safety = [
+        'johnson & johnson', 'ben & jerry', 'tiffany & co', 'h&m', 'at&t', 
+        'dolce & gabbana', 'marks & spencer', 'bed bath & beyond'
+    ]
     
+    # UI Noise and Number Words to ignore
+    trash_words = ['zero', 'one', 'two', 'three', 'privacy & cookie', 'accept decline']
+
     processed_entities = []
     seen = set()
 
     for ent in doc.ents:
         raw_text = ent.text.strip()
         e_lower = raw_text.lower()
-        words = e_lower.split()
-        
-        # --- NEW BUG FILTERING LOGIC ---
-        # 1. Remove "Repetition Bugs" (like Social Media Social)
-        if len(words) > 1 and words[-1] == words[0]:
-            continue
         
-        # 2. Skip generic "SEO Fluff" phrases that aren't real entities
-        if "social media social" in e_lower:
+        # 1. Filter out pure numbers, "zero", and UI trash
+        if e_lower in trash_words or raw_text.isdigit():
             continue
 
+        # 2. Smart Splitting Logic (Bing and Yandex -> Bing, Yandex)
         sub_items = [raw_text]
         if e_lower not in brand_safety:
             if " and " in e_lower or " & " in raw_text:
@@ -53,18 +55,25 @@ async def analyze_text(payload: TextPayload):
             if f_lower in seen or len(final_text) < 2:
                 continue
 
+            # 3. Bug Filter: Avoid repetition bugs like "Social Media Social"
+            words = f_lower.split()
+            if len(words) > 1 and words[-1] == words[0]:
+                continue
+
             label = ent.label_
             
-            # Category forcing
+            # 4. Category Force Mapping
             concepts = ['privacy', 'cookies', 'cookie', 'booster', 'get', 'noticed', 'free', 'social media']
-            tech_items = ['exif', 'mediaif', 'cookiesif', 'json', 'api', 'schema', 'vps', 'server']
+            tech_items = ['exif', 'mediaif', 'cookiesif', 'json', 'api', 'schema', 'vps', 'server', 'gravatar']
 
             if any(c in f_lower for c in concepts):
                 label = "CONCEPT"
             elif any(t in f_lower for t in tech_items):
                 label = "PRODUCT"
+            elif any(brand in f_lower for brand in brand_safety):
+                label = "ORG"
 
-            # Block long UI fragments
+            # 5. Length Filter (Max 3 words for entities)
             if len(final_text.split()) >= 4:
                 continue
 
@@ -76,6 +85,7 @@ async def analyze_text(payload: TextPayload):
             })
             seen.add(f_lower)
 
+    # Return up to 50 results
     sorted_entities = sorted(processed_entities, key=lambda x: x['frequency'], reverse=True)
 
     return {