adding llm

.gradio/flagged/dataset1.csv

@@ -0,0 +1,2 @@
+user_input,output,timestamp
+,,2026-03-04 20:40:22.416508

src/llm/client.py

@@ -0,0 +1,36 @@
+from dotenv import load_dotenv
+from pathlib import Path
+load_dotenv(override=True)
+
+import os
+from groq import Groq
+
+# Debug: print to confirm key is loaded
+api_key = os.environ.get('GROQ_API_KEY')
+
+def infer_model(msg, massages = []):
+    client = Groq(api_key=api_key)
+    
+
+    massages.append({"role":"user", "content": msg})
+
+        
+    chat_completion = client.chat.completions.create(
+        messages=massages,
+        model="groq/compound",
+    )
+
+    return chat_completion, messages 
+
+msg = ""
+messages = []
+while(True):
+    msg = input("\nescreva: \n")
+    if msg == "quit":
+        break
+    out, messages = infer_model(msg, messages)
+    print(out.choices[0].message.content)
+
+
+
+