File size: 2,567 Bytes
1915306 0e9eccb 1915306 777bb52 1915306 0e9eccb 4d0e77c 0e9eccb 4d0e77c 0e9eccb 4d0e77c 0e9eccb 8b5f3bd 0e9eccb 8b5f3bd 4a65c44 8b5f3bd 1915306 8b5f3bd e8b0ec8 0e9eccb 4d0e77c 1915306 0c08de9 987b836 1915306 987b836 1915306 4d0e77c 0c08de9 4d0e77c 0e9eccb 0c08de9 1915306 8b5f3bd 1915306 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 |
import gradio as gr
from huggingface_hub import InferenceClient
import requests
client = InferenceClient("mistralai/Mistral-Nemo-Instruct-2407")
def get_internet_data(query: str) -> str:
"""
Uses DuckDuckGo's Instant Answer API to fetch search data.
Returns a short summary of the query result.
"""
url = "https://api.duckduckgo.com"
params = {
"q": query,
"format": "json",
"no_redirect": 1,
"skip_disambig": 1,
}
try:
response = requests.get(url, params=params, timeout=5)
response.raise_for_status()
data = response.json()
# Retrieve a summary from the API's result
result = data.get("AbstractText", "")
if not result:
result = "Wala koy nakuha nga impormasyon gikan sa internet."
except Exception:
result = "Naay problema sa pagkuha sa impormasyon gikan sa internet."
return result
def respond(message, history: list[tuple[str, str]]):
system_message = "Ikaw usa ka buotan nga Chatbot. Tubaga lang sa binisaya, ug ayaw gamita ang english nga pinulungan."
max_tokens = 4096
temperature = 0.6
top_p = 0.95
messages = [{"role": "system", "content": system_message}]
for user_text, assistant_text in history:
if user_text:
messages.append({"role": "user", "content": user_text})
if assistant_text:
messages.append({"role": "assistant", "content": assistant_text})
# Automatically perform an internet search for the user's query.
search_result = get_internet_data(message)
messages.append({
"role": "assistant",
"content": f"Mga resulta gikan sa internet para sa '{message}': {search_result}"
})
# Include the original query as part of the conversation context.
messages.append({"role": "user", "content": message})
response = ""
previous_response = ""
for token_message in client.chat_completion(
messages,
max_tokens=max_tokens,
stream=True,
temperature=temperature,
top_p=top_p,
):
token = token_message.choices[0].delta.get("content", "")
if not token:
break
response += token
# Yield only if new content was added
if response != previous_response:
yield response
previous_response = response
# Optional: break if the response is too long
if len(response) > 3000:
break
demo = gr.ChatInterface(respond)
if __name__ == "__main__":
demo.launch()
|