Update synthetic_data.py

synthetic_data.py

@@ -1,11 +1,32 @@
-import pandas as pd
-import random
-
-def generate_synthetic_dataset(num_agents=5, num_samples=20):
-    data = []
-    for agent in range(num_agents):
-        for i in range(num_samples):
-            instr = f"Task {i}: Summarize topic {random.randint(1,10)}"
-            resp = f"This is a response to {instr}. Agent {agent} provides details."
-            data.append({"agent": f"Agent_{agent}", "instruction": instr, "response": resp})
-    return pd.DataFrame(data)
+# synthetic_data.py
+import pandas as pd
+import random
+
+def generate_synthetic_dataset(num_agents=4, num_samples=20):
+    data = []
+    tasks = ["qa", "summarization", "reasoning"]
+    for a in range(num_agents):
+        agent_name = f"Agent_{a}"
+        for i in range(num_samples):
+            task = random.choice(tasks)
+            if task == "qa":
+                prompt = f"What is the capital of country {i}?"
+                reference = "Paris is the capital of France." if i % 2 == 0 else "Rome is the capital of Italy."
+                response = reference if random.random() > 0.2 else "I think it's Berlin."
+            elif task == "summarization":
+                prompt = "Summarize: 'The sun is a star that provides light and heat to the Earth.'"
+                reference = "The sun provides light and heat essential for life on Earth."
+                response = "The sun provides light and heat for Earth."
+            else:
+                prompt = "Explain why the sky is blue."
+                reference = "Rayleigh scattering causes shorter blue wavelengths to scatter more, making the sky appear blue."
+                response = reference if random.random() > 0.3 else "It is because of reflection."
+
+            data.append({
+                "prompt": prompt,
+                "response": response,
+                "task": task,
+                "agent": agent_name,
+                "reference": reference
+            })
+    return pd.DataFrame(data)