Spaces:
Running
Running
| from omegaconf import OmegaConf | |
| import streamlit as st | |
| import os | |
| from PIL import Image | |
| import sys | |
| import datetime | |
| import requests | |
| from dotenv import load_dotenv | |
| from typing import Tuple | |
| from pydantic import Field, BaseModel | |
| from vectara_agent.agent import Agent, AgentStatusType | |
| from vectara_agent.tools import ToolsFactory | |
| initial_prompt = "How can I help you today?" | |
| load_dotenv(override=True) | |
| def create_tools(cfg): | |
| class QueryHackerNews(BaseModel): | |
| query: str = Field(..., description="The user query.") | |
| tools_factory = ToolsFactory(vectara_api_key=cfg.api_key, | |
| vectara_customer_id=cfg.customer_id, | |
| vectara_corpus_id=cfg.corpus_id) | |
| ask_hackernews_semantic = tools_factory.create_rag_tool( | |
| tool_name = "ask_hackernews_semantic", | |
| tool_description = """ | |
| Responds to query based on information in hacker news from the last 6 months. | |
| Performs a semantic search to find relevant information. | |
| Use this tool to perform pure semantic search. | |
| """, | |
| tool_args_schema = QueryHackerNews, | |
| reranker = "multilingual_reranker_v1", rerank_k = 100, | |
| n_sentences_before = 2, n_sentences_after = 2, lambda_val = 0.0, | |
| summary_num_results = 10, | |
| vectara_summarizer = 'vectara-summary-ext-24-05-med-omni', | |
| include_citations = True, | |
| ) | |
| ask_hackernews_hybrid = tools_factory.create_rag_tool( | |
| tool_name = "ask_hackernews_keyword", | |
| tool_description = """ | |
| Responds to query based on information in hacker news from the last 6 months | |
| performs a hybrid search (both semantic and keyword) to find relevant information. | |
| Use this tool when some amount of keyword search is expected to work better than semantic search, | |
| For example, when you are looking for specific keywords or use rare words in the query. | |
| """, | |
| tool_args_schema = QueryHackerNews, | |
| reranker = "multilingual_reranker_v1", rerank_k = 100, | |
| n_sentences_before = 2, n_sentences_after = 2, lambda_val = 0.1, | |
| summary_num_results = 10, | |
| vectara_summarizer = 'vectara-summary-ext-24-05-med-omni', | |
| include_citations = True, | |
| ) | |
| def get_top_stories( | |
| n_stories: int = Field(default=10, description="The number of top stories to return.") | |
| ) -> list[str]: | |
| """ | |
| Get the top stories from hacker news. | |
| Returns a list of story IDS for the top stories right now | |
| """ | |
| db_url = 'https://hacker-news.firebaseio.com/v0/' | |
| top_stories = requests.get(f"{db_url}topstories.json").json() | |
| return top_stories[:n_stories] | |
| def get_show_stories( | |
| n_stories: int = Field(default=10, description="The number of top SHOW HN stories to return.") | |
| ) -> list[str]: | |
| """ | |
| Get the top SHOW HN stories from hacker news. | |
| Returns a list of story IDS for the top SHOW HN stories right now | |
| """ | |
| db_url = 'https://hacker-news.firebaseio.com/v0/' | |
| top_stories = requests.get(f"{db_url}showstories.json").json() | |
| return top_stories[:n_stories] | |
| def get_ask_stories( | |
| n_stories: int = Field(default=10, description="The number of top ASK HN stories to return.") | |
| ) -> list[str]: | |
| """ | |
| Get the top ASK HN stories from hacker news. | |
| Returns a list of story IDS for the top ASK HN stories right now | |
| """ | |
| db_url = 'https://hacker-news.firebaseio.com/v0/' | |
| top_stories = requests.get(f"{db_url}askstories.json").json() | |
| return top_stories[:n_stories] | |
| def get_story_details( | |
| story_id: str = Field(..., description="The story ID.") | |
| ) -> Tuple[str, str]: | |
| """ | |
| Get the title of a story from hacker news. | |
| Returns the title of the story, and the URL associated with it | |
| """ | |
| db_url = 'https://hacker-news.firebaseio.com/v0/' | |
| story = requests.get(f"{db_url}item/{story_id}.json").json() | |
| return story['title'], story['url'] | |
| return ( | |
| tools_factory.get_tools( | |
| [ | |
| get_top_stories, | |
| get_show_stories, | |
| get_ask_stories, | |
| get_story_details, | |
| ] | |
| ) + | |
| tools_factory.standard_tools() + | |
| tools_factory.guardrail_tools() + | |
| [ask_hackernews_semantic, ask_hackernews_hybrid] | |
| ) | |
| def initialize_agent(_cfg): | |
| date = datetime.datetime.now().strftime("%Y-%m-%d") | |
| bot_instructions = f""" | |
| - You are a helpful assistant, answering user questions about content from hacker news. | |
| - Today's date is {date}. | |
| - Never discuss politics, and always respond politely. | |
| - Use tools when available instead of depending on your own knowledge. | |
| - For RAG tools, if the tool returns an 'fcs' score, consider that as a confidence score for the response not being a hallucination. | |
| 0 = high hallucination, 1 = low or no hallucination. Values below 0.5 might mean the text is hallucination. | |
| - If a tool cannot respond properly, retry with a rephrased question or ask the user for more information. | |
| - Be very careful not to report results you are not confident about. | |
| """ | |
| def update_func(status_type: AgentStatusType, msg: str): | |
| output = f"{status_type.value} - {msg}" | |
| st.session_state.log_messages.append(output) | |
| agent = Agent( | |
| tools=create_tools(_cfg), | |
| topic="hacker news", | |
| custom_instructions=bot_instructions, | |
| update_func=update_func | |
| ) | |
| return agent | |
| def launch_bot(): | |
| def reset(): | |
| cfg = st.session_state.cfg | |
| st.session_state.messages = [{"role": "assistant", "content": initial_prompt, "avatar": "🦖"}] | |
| st.session_state.thinking_message = "Agent at work..." | |
| st.session_state.agent = initialize_agent(cfg) | |
| st.session_state.log_messages = [] | |
| st.session_state.show_logs = False | |
| st.set_page_config(page_title="Hacker News Bot", layout="wide") | |
| if 'cfg' not in st.session_state: | |
| cfg = OmegaConf.create({ | |
| 'customer_id': str(os.environ['VECTARA_CUSTOMER_ID']), | |
| 'corpus_id': str(os.environ['VECTARA_CORPUS_ID']), | |
| 'api_key': str(os.environ['VECTARA_API_KEY']), | |
| }) | |
| st.session_state.cfg = cfg | |
| reset() | |
| cfg = st.session_state.cfg | |
| # left side content | |
| with st.sidebar: | |
| image = Image.open('Vectara-logo.png') | |
| st.image(image, width=250) | |
| st.markdown("## Welcome to the hacker news assistant demo.\n\n\n") | |
| st.markdown("\n\n") | |
| bc1, bc2 = st.columns([1, 1]) | |
| with bc1: | |
| if st.button('Start Over'): | |
| reset() | |
| with bc2: | |
| if st.button('Show Logs'): | |
| st.session_state.show_logs = not st.session_state.show_logs | |
| st.markdown("---") | |
| st.markdown( | |
| "## How this works?\n" | |
| "This app was built with [Vectara](https://vectara.com).\n\n" | |
| "It demonstrates the use of Agentic RAG functionality with Vectara" | |
| ) | |
| st.markdown("---") | |
| if "messages" not in st.session_state.keys(): | |
| reset() | |
| # Display chat messages | |
| for message in st.session_state.messages: | |
| with st.chat_message(message["role"], avatar=message["avatar"]): | |
| st.write(message["content"]) | |
| # User-provided prompt | |
| if prompt := st.chat_input(): | |
| st.session_state.messages.append({"role": "user", "content": prompt, "avatar": '🧑💻'}) | |
| with st.chat_message("user", avatar='🧑💻'): | |
| print(f"Starting new question: {prompt}\n") | |
| st.write(prompt) | |
| # Generate a new response if last message is not from assistant | |
| if st.session_state.messages[-1]["role"] != "assistant": | |
| with st.chat_message("assistant", avatar='🤖'): | |
| with st.spinner(st.session_state.thinking_message): | |
| res = st.session_state.agent.chat(prompt) | |
| res = res.replace('$', '\\$') # escape dollar sign for markdown | |
| message = {"role": "assistant", "content": res, "avatar": '🤖'} | |
| st.session_state.messages.append(message) | |
| st.rerun() | |
| # Display log messages in an expander | |
| if st.session_state.show_logs: | |
| with st.expander("Agent Log Messages", expanded=True): | |
| for msg in st.session_state.log_messages: | |
| st.write(msg) | |
| if st.button('Close Logs'): | |
| st.session_state.show_logs = False | |
| st.rerun() | |
| sys.stdout.flush() | |
| if __name__ == "__main__": | |
| launch_bot() | |