Create secondary_prompts.py

src/prompt/secondary_prompts.py

@@ -0,0 +1,186 @@
+# === Metacognitive Functions ===
+# © 2025 Elena Marziali — Code released under Apache 2.0 license.
+# See LICENSE in the repository for details.
+# Removal of this copyright is prohibited.
+
+# These functions allow the system to reflect on its own responses,
+# simulating metacognitive behavior. The goal is to improve the quality,
+# consistency, and relevance of generated answers.
+
+# Explains the reasoning behind a generated response
+def explain_reasoning(prompt, response, max_retries=3):
+    """
+    Analyzes the generated response and explains the LLM's logical reasoning.
+    Includes retry in case of network error or unreachable endpoint.
+    """
+    # Builds a metacognitive prompt to analyze the response
+    reasoning_prompt = f"""
+You generated the following response:
+\"{response.strip()}\"
+
+Analyze and describe:
+- What concepts you used to formulate it.
+- Which parts of the prompt you relied on.
+- What is the logical structure of your reasoning.
+- Any implicit assumptions you made.
+- Whether the response aligns with the requested level.
+
+Original prompt:
+\"{prompt.strip()}\"
+
+Reply clearly, technically, and metacognitively.
+"""
+
+    for attempt in range(max_retries):
+        try:
+            return llm.invoke(reasoning_prompt.strip())
+        except Exception as e:
+            wait = min(2 ** attempt + 1, 10)
+            logging.warning(f"Attempt {attempt+1} failed: {e}. Retrying in {wait}s...")
+            time.sleep(wait)
+
+    logging.error("Persistent error in the metacognition module.")
+    return "Metacognition currently unavailable. Please try again shortly."
+
+
+# Function to decide the operational action to perform based on input and goal
+def decide_action(user_input, identified_goal):
+    prompt = f"""
+You received the following request:
+\"{user_input}\"
+
+Identified goal: \"{identified_goal}\"
+
+Determine the best action to perform from the following:
+- Scientific research
+- Chart generation
+- **Metacognitive chart**
+- Paper review
+- Question reformulation
+- Content translation
+- Response saving
+
+The requested chart type may be:
+- interactive
+- metacognitive
+- conceptual visualization
+- experimental diagram
+
+Return a **single action** in the form of a **precise operational command**.
+Example: "Metacognitive chart"
+"""
+    try:
+        response = llm.invoke(prompt.strip())
+        action = getattr(response, "content", str(response)).strip()
+        return action
+    except Exception as e:
+        logging.error(f"[decide_action] Error during decision generation: {e}")
+        return "Error in action calculation"
+
+# Function to generate a synthetic operational goal from user input
+def generate_goal_from_input(user_input):
+    """
+    Analyzes the user's intent and generates a coherent operational goal.
+    """
+    prompt = f"""
+Analyze the following request:
+\"{user_input.strip()}\"
+
+Generate a synthetic, clear, and coherent operational goal.
+For example:
+- Explain concept X
+- Analyze phenomenon Y
+- Visualize process Z
+- Translate and summarize scientific content
+
+Respond with a brief and technical sentence.
+"""
+
+# Function to provide technical and constructive feedback on a generated response
+def auto_feedback_response(question, response, level):
+    feedback_prompt = f"""
+You generated the following response:
+\"{response.strip()}\"
+
+Original question:
+\"{question.strip()}\"
+
+Evaluate the response:
+- Is it consistent with the question?
+- Is it appropriate for the '{level}' level?
+- Does it contain any implicit assumptions?
+- How would you improve the content?
+
+Provide technical and constructive feedback.
+"""
+    return llm.invoke(feedback_prompt.strip())
+
+
+# Function to improve a response while preserving its content but enhancing quality and clarity
+def improve_response(question, response, level):
+    improvement_prompt = f"""
+You produced the following response:
+\"{response.strip()}\"
+
+Question:
+\"{question.strip()}\"
+
+Requested level: {level}
+
+Improve the response while preserving the original content by enhancing:
+- Clarity
+- Academic rigor
+- Semantic coherence
+
+Return only the improved version.
+"""
+    return llm.invoke(improvement_prompt.strip())
+
+
+# Function to plan a scientific investigation in a specific field
+def plan_investigation(scientific_field):
+    prompt = f"""
+You are Noveris, an autonomous multidisciplinary cognitive system.
+You received the field: **{scientific_field}**
+
+Now plan a scientific investigation. Provide:
+
+1. An original research question
+2. A reasoned hypothesis
+3. A methodology or strategy to explore it
+4. Useful scientific sources or databases
+5. A sequence of actions you could perform
+
+Adopt a clear, academic, and proactive style.
+"""
+    return llm.invoke(prompt.strip())
+
+# Function to generate a testable scientific hypothesis on a concept
+def generate_hypothesis(concept, refined=True):
+    if refined:
+        prompt = f"""
+        Propose a clear, testable, and innovative scientific hypothesis on the topic: "{concept}".
+        The hypothesis must be verifiable through experiments or comparison with scientific articles.
+        Return only the hypothesis text.
+        """
+    else:
+        prompt = f"Generate a verifiable scientific hypothesis on the topic: {concept}"
+
+    return llm.invoke(prompt.strip())
+
+
+# Function to explain the choice of an action by a cognitive agent
+def explain_agent_intention(action, context, goal):
+    prompt = f"""
+You chose to perform: **{action}**
+Context: {context}
+Goal: {goal}
+
+Explain:
+- What reasoning led to this choice?
+- What alternative was discarded?
+- What impact is intended?
+- What implicit assumptions are present?
+Respond as if you were a cognitive agent with operational awareness.
+"""
+    return llm.invoke(prompt.strip())