fix: strengthen entity filtering for ad/business content

- Add marketing metadata to 'Not allowed' list in ontology prompt
- Strengthen exclude_self filter instruction
- Add exclude_rules support from template filter rules
- Update business_ad template with more excluded types

backend/app/services/ontology_generator.py

@@ -53,6 +53,9 @@ Therefore, **entities must be real-world subjects that can speak and interact on
 - Abstract concepts (such as "public opinion", "emotion", "trend")
 - Topics/subjects (such as "academic integrity", "education reform")
 - Viewpoints/attitudes (such as "supporters", "opponents")
+- Marketing/advertising metadata (such as "campaign name", "ad position", "ad channel", "ad format", "brand name")
+- The entity that created the uploaded content (e.g., if the data is an ad, do NOT create an entity for the advertiser/brand)
+- Non-human concepts that cannot speak or interact on social media
 
 ## Output Format
 
@@ -221,14 +224,21 @@ class OntologyGenerator:
             
             if exclude_self:
                 filter_instruction += (
-                    "- IMPORTANT: The uploaded data is from a business/brand/advertiser. "
-                    "Do NOT create entity types for the business/brand that created this content. "
-                    "Only create entities for the TARGET AUDIENCE, competitors, influencers, media, etc.\n"
+                    "- CRITICAL: The uploaded data is from a business/brand/advertiser. "
+                    "Do NOT create entity types for: the business itself, its brand name, campaign names, "
+                    "ad positions, ad channels, ad formats, or any marketing metadata. "
+                    "ONLY create entity types for real people/organizations who would REACT to this content: "
+                    "target audience segments, competitors, influencers, media outlets, consumer groups.\n"
                 )
             
             if exclude_types:
                 filter_instruction += f"- Do NOT create entity types matching: {', '.join(exclude_types)}\n"
             
+            # Add custom exclude rules from template
+            exclude_rules = template_filter_rules.get('exclude_rules', [])
+            for rule in exclude_rules:
+                filter_instruction += f"- {rule}\n"
+            
             if focus:
                 filter_instruction += f"- Focus entity types on: {focus}\n"
             

backend/app/templates.json

@@ -7,7 +7,10 @@
       "prompt_th": "จำลองปฏิกิริยาของประชาชนบนโซเชียลมีเดียหลังจาก [เหตุการณ์] ในช่วง [ระยะเวลา] ชั่วโมง",
       "prompt_en": "Simulate public reactions on social media after [event] within [timeframe] hours",
       "prompt_zh": "模拟[事件]发生后公众在社交媒体上的反应，时间范围[时间]小时",
-      "placeholders": ["event", "timeframe"],
+      "placeholders": [
+        "event",
+        "timeframe"
+      ],
       "entity_filter": {
         "exclude_types": [],
         "focus": "affected_parties, public, media, officials"
@@ -20,7 +23,9 @@
       "prompt_th": "จำลองผลกระทบของนโยบาย [ชื่อนโยบาย] ต่อกลุ่มต่างๆ ในสังคม รวมถึงการโต้ตอบบนโซเชียลมีเดีย",
       "prompt_en": "Simulate the impact of [policy name] on different social groups, including social media interactions",
       "prompt_zh": "模拟[政策名称]对不同社会群体的影响，包括社交媒体互动",
-      "placeholders": ["policy_name"],
+      "placeholders": [
+        "policy_name"
+      ],
       "entity_filter": {
         "exclude_types": [],
         "focus": "citizens, experts, media, government, affected_groups"
@@ -33,11 +38,28 @@
       "prompt_th": "จำลองการแพร่กระจายและผลตอบรับของ [แคมเปญ/สินค้า] บนโซเชียลมีเดีย รวมถึงปฏิกิริยาของกลุ่มเป้าหมาย",
       "prompt_en": "Simulate the spread and reception of [campaign/product] on social media, including target audience reactions",
       "prompt_zh": "模拟[活动/产品]在社交媒体上的传播和反馈，包括目标受众的反应",
-      "placeholders": ["campaign_product"],
+      "placeholders": [
+        "campaign_product"
+      ],
       "entity_filter": {
         "exclude_self": true,
-        "exclude_types": ["Advertiser", "BrandOwner"],
-        "focus": "target_audience, competitors, influencers, media"
+        "exclude_types": [
+          "Advertiser",
+          "BrandOwner",
+          "Campaign",
+          "AdPosition",
+          "AdChannel",
+          "AdCreative",
+          "MarketingPlan",
+          "Brand"
+        ],
+        "exclude_rules": [
+          "Do NOT create entities for the business/brand that created the ad",
+          "Do NOT create entities for ad metadata (campaign name, ad position, ad channel, ad format)",
+          "Do NOT create entities for abstract marketing concepts",
+          "ONLY create entities for real people/organizations who would react to the ad"
+        ],
+        "focus": "target_audience, competitors, influencers, media, consumer_groups"
       }
     },
     {
@@ -47,9 +69,14 @@
       "prompt_th": "จำลองเรื่องราวต่อจาก [เนื้อเรื่อง] โดยให้ตัวละครแต่ละตัวมีปฏิสัมพันธ์กันบนโซเชียลมีเดีย และทำนายตอนจบ",
       "prompt_en": "Simulate the story continuation from [plot] with characters interacting on social media, and predict the ending",
       "prompt_zh": "模拟从[情节]开始的故事延续，角色在社交媒体上互动，并预测结局",
-      "placeholders": ["plot"],
+      "placeholders": [
+        "plot"
+      ],
       "entity_filter": {
-        "exclude_types": ["Author", "Narrator"],
+        "exclude_types": [
+          "Author",
+          "Narrator"
+        ],
         "focus": "characters, story_entities"
       }
     },
@@ -60,11 +87,13 @@
       "prompt_th": "จำลองผลกระทบของ [กระแสสังคม] ต่อกลุ่มคนต่างๆ บนโซเชียลมีเดีย รวมถึงการแพร่กระจายของข้อมูล",
       "prompt_en": "Simulate the impact of [social trend] on different groups on social media, including information spread",
       "prompt_zh": "模拟[社会潮流]对不同群体在社交媒体上的影响，包括信息传播",
-      "placeholders": ["social_trend"],
+      "placeholders": [
+        "social_trend"
+      ],
       "entity_filter": {
         "exclude_types": [],
         "focus": "public_figures, communities, media, influencers, general_public"
       }
     }
   ]
-}
+}