feat: implement chatbot engine and prompt builder modules for graduation project recommendation and validation

src/recommendation_engine/chatbot_engine.py

@@ -119,31 +119,32 @@ Do not include any other text or punctuation.
     except Exception as e:
         return "general_chat"
 
-def is_valid_graduation_project_domain(domain: str) -> bool:
+def validate_and_format_domain(domain: str) -> str:
     # 1. Quick local validation for standard domains
     extracted = extract_domain(domain)
     if extracted and extracted.lower() != "others":
-        return True
+        return extracted
 
     # 2. Fall back to LLM validation
     prompt = f"""
 Determine if the following domain/field is a valid academic, engineering, scientific, or technology domain suitable for a university graduation project (e.g., Computer Science, Engineering, Medicine, Business, Agriculture, Biology, etc.).
+Also, correct any typos and format it cleanly (e.g., Title Case).
 
 Domain to evaluate: "{domain}"
 
 Rules:
-- If it is a valid field of study, technology, or academic discipline (e.g., "artificial intelligence", "robotics", "bioinformatics", "educational games") -> Return YES
-- If it is unrelated to academic/technology graduation projects, or contains names of celebrities, sports teams, food, pop culture, or random questions (e.g., "messi", "fc barcelona", "pizza", "what is this") -> Return NO
+- If it is a valid field of study, technology, or academic discipline (e.g., "artificial intelligence", "robotics", "bioinformatics", "educational games"), return ONLY the corrected and formatted domain name (e.g., "Artificial Intelligence").
+- If it is unrelated to academic/technology graduation projects, or contains names of celebrities, sports teams, food, pop culture, or random questions (e.g., "messi", "fc barcelona", "pizza", "what is this"), return exactly "INVALID".
 
-Return ONLY "YES" or "NO".
+Return ONLY the formatted domain name or "INVALID". Do not include any other text.
 """
     try:
-        res = generate_text(prompt, task="intent").strip().upper()
-        if not res:
-            return True
-        return "YES" in res
+        res = generate_text(prompt, task="intent").strip()
+        if not res or res.upper() == "INVALID":
+            return ""
+        return res.strip('"').strip("'")
     except Exception:
-        return True
+        return ""
 
 def is_weak_project_title(title: str) -> bool:
 
@@ -532,10 +533,16 @@ def chatbot(user_id: str, user_input: str):
         "ideas"
     ]
 
-    if any(x in text for x in explicit_idea_requests) or text in ["change domain", "new domain", "reset domain", "reset", "restart"]:
-        memory = get_user_memory(user_id)
-        current_state = memory.get("state") or default_state()
-        
+    memory = get_user_memory(user_id)
+    current_state = memory.get("state") or default_state()
+    
+    is_explicit_idea = False
+    if text in explicit_idea_requests or text in ["change domain", "new domain", "reset domain", "reset", "restart"]:
+        is_explicit_idea = True
+    elif any(x in text for x in explicit_idea_requests) and len(text.split()) <= 3 and not current_state.get("project_chat_mode"):
+        is_explicit_idea = True
+
+    if is_explicit_idea:
         if not current_state.get("waiting_for_niche_domain"):
             state = default_state()
             state["waiting_for_domain"] = True
@@ -715,6 +722,11 @@ def chatbot(user_id: str, user_input: str):
         state["waiting_for_project_idea_confirm"] = False
 
         if is_yes:
+            if state.get("from_full_project_generation"):
+                state["from_full_project_generation"] = False
+                save_user_memory(user_id, {"history": history, "state": state})
+                return "Do you want to add this to Submission?"
+
             title = state.get("project_title", "")
             abstract = state.get("abstract", "")
 
@@ -734,11 +746,9 @@ def chatbot(user_id: str, user_input: str):
                     f"📌 Title: {title}\n"
                     f"{domain_line}\n\n"
                     f"What would you like to do next?\n"
-                    f"1️⃣ Generate features\n"
-                    f"2️⃣ Talk with chatbot about the idea\n\n"
+                    f"1️⃣ Generate features\n\n"
                     f"👉 You can also say:\n"
-                    f"- generate features\n"
-                    f"- discuss project"
+                    f"- generate features"
                 )
             else:
                 reason = validation.get("reason", "")
@@ -814,11 +824,9 @@ def chatbot(user_id: str, user_input: str):
                 f"{project_title}\n\n"
                 "Choose what you want next:\n"
                 "1️⃣ Generate features\n"
-                "2️⃣ Talk with chatbot about the idea\n"
-                "3️⃣ Generate full project\n\n"
+                "2️⃣ Generate full project\n\n"
                 "👉 You can also say:\n"
                 "- generate features\n"
-                "- discuss project\n"
                 "- generate full project"
             )
 
@@ -947,6 +955,7 @@ def chatbot(user_id: str, user_input: str):
                 state["category"] = state.get("domain")
 
             state["waiting_for_project_idea_confirm"] = True
+            state["from_full_project_generation"] = True
 
             response = f"""
 📦 Full Project Generated
@@ -1098,7 +1107,6 @@ Reply **yes** to validate & categorize it, or **no** to explore other ideas.
         ]
 
         chat_words = [
-            "2",
             "chat",
             "talk",
             "discussion",
@@ -1109,7 +1117,7 @@ Reply **yes** to validate & categorize it, or **no** to explore other ideas.
         ]
 
         full_project_words = [
-    "3",
+    "2",
     "full project",
     "generate full project",
     "complete project",
@@ -1235,6 +1243,7 @@ Reply **yes** to validate & categorize it, or **no** to explore other ideas.
                 state["category"] = state.get("domain")
 
             state["waiting_for_project_idea_confirm"] = True
+            state["from_full_project_generation"] = True
 
             response = f"""
 📦 Full Project Generated
@@ -1546,12 +1555,10 @@ Reply **yes** to validate & categorize it, or **no** to explore other ideas.
 Choose what you want:
 
 1️⃣ Generate Features
-2️⃣ Discuss Project
-3️⃣ Generate Full Project
+2️⃣ Generate Full Project
 
 👉 You can also type:
 - features
-- discuss
 - full project
 """,
 
@@ -1575,11 +1582,9 @@ Choose what you want:
                 f"📌 Project selected:\n"
                 f"{idea}\n\n"
                 "Choose what you want next:\n"
-                "1️⃣ Generate features\n"
-                "2️⃣ Talk with chatbot about the idea\n\n"
+                "1️⃣ Generate features\n\n"
                 "👉 You can also say:\n"
-                "- generate features\n"
-                "- discuss project"
+                "- generate features"
             )
 
     
@@ -1631,14 +1636,15 @@ Choose what you want:
                 return finalize_response(user_input, response, history, state, user_id)
             else:
                 domain = user_input.strip()
-                if not is_valid_graduation_project_domain(domain):
+                validated_domain = validate_and_format_domain(domain)
+                if not validated_domain:
                     return (
                         "⚠️ That does not seem to be a valid technology or academic domain.\n\n"
                         "Please enter a valid domain (e.g. AI, Healthcare, Fintech, or 'Others')."
                     )
                 state["waiting_for_niche_domain"] = False
                 state["suggested_domains"] = []
-                analysis["domain"] = domain
+                analysis["domain"] = validated_domain
                 analysis["intent"] = "idea"
                 if "project_title" in analysis:
                     del analysis["project_title"]
@@ -1680,10 +1686,18 @@ Choose what you want:
                 if "project_title" in analysis:
                     del analysis["project_title"]
             else:
-                return (
-                    "⚠️ Please enter a valid domain "
-                    "(AI, healthcare, fintech...)"
-                )
+                validated_domain = validate_and_format_domain(user_input.strip())
+                if validated_domain:
+                    state["waiting_for_domain"] = False
+                    analysis["domain"] = validated_domain
+                    analysis["intent"] = "idea"
+                    if "project_title" in analysis:
+                        del analysis["project_title"]
+                else:
+                    return (
+                        "⚠️ Please enter a valid domain "
+                        "(AI, healthcare, fintech...)"
+                    )
 
     intent = analysis.get("intent", "chat")
 
@@ -1691,6 +1705,12 @@ Choose what you want:
 
     user_project = analysis.get("project_title")
 
+    if state.get("project_chat_mode"):
+        user_project = None
+        exact_cmds = {"idea", "ideas", "menu", "exit", "quit", "generate features", "features", "generate full project"}
+        if user_input.strip().lower() not in exact_cmds:
+            intent = "chat"
+
     # Override intent using the LLM-derived semantic context when no project is active yet
     if not state.get("project_title") and not is_in_domain_or_selection_state:
         if context == "idea_request":
@@ -1894,11 +1914,9 @@ Choose what you want:
     f"{user_project}\n\n"
     "Choose what you want next:\n"
     "1️⃣ Generate features\n"
-    "2️⃣ Talk with chatbot about the idea\n"
-    "3️⃣ Generate full project\n\n"
+    "2️⃣ Generate full project\n\n"
     "👉 You can also say:\n"
     "- generate features\n"
-    "- discuss project\n"
     "- generate full project"
 )
     
@@ -2106,6 +2124,7 @@ Choose what you want:
             state["category"] = state.get("domain")
 
         state["waiting_for_project_idea_confirm"] = True
+        state["from_full_project_generation"] = True
 
         response = f"""
 📦 Project Specification Ready!
@@ -2206,7 +2225,7 @@ User:
     
     
     
-    if not is_project_related(text):
+    if not is_project_related(text) and not state.get("project_title"):
 
         return (
             "I'm designed specifically for graduation projects 🎓\n\n"

src/recommendation_engine/prompt_builder.py

@@ -243,6 +243,7 @@ RULES:
 - Be realistic
 - Be technically accurate
 - Keep all answers related to the current project
+- STRICTLY REFUSE to discuss any illegal, inappropriate, explicit (e.g., sexual content), or unethical topics. If the user mentions such topics, reply immediately with a professional warning that you can only discuss academic graduation projects.
 """.strip()
 
 def build_full_project_prompt(context):