added tools

README.md

@@ -1,5 +1,5 @@
 ---
-title: Finance chat
+title: Hacker News chat
 emoji: 🐨
 colorFrom: indigo
 colorTo: indigo
@@ -8,7 +8,7 @@ app_port: 8501
 app_file: app.py
 pinned: false
 license: apache-2.0
-short_description: Finance chatbot using vectara-agent
+short_description: chatbot with HN data using vectara-agent
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -3,128 +3,127 @@ from omegaconf import OmegaConf
 import streamlit as st
 import os
 from PIL import Image
-import re
 import sys
 import datetime
-import pandas as pd
 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
 
-
-tickers = {
-    "AAPL": "Apple Computer", 
-    "GOOG": "Google", 
-    "AMZN": "Amazon",
-    "SNOW": "Snowflake",
-    "TEAM": "Atlassian",
-    "TSLA": "Tesla",
-    "NVDA": "Nvidia",
-    "MSFT": "Microsoft",
-    "AMD": "Advanced Micro Devices",
-    "INTC": "Intel",
-    "NFLX": "Netflix",
-}
-years = [2020, 2021, 2022, 2023, 2024]
 initial_prompt = "How can I help you today?"
 
 load_dotenv(override=True)
 
 def create_tools(cfg):    
-
-    def get_company_info() -> list[str]:
-        """
-        Returns a dictionary of companies you can query about. Always check this before using any other tool.
-        The output is a dictionary of valid ticker symbols mapped to company names.
-        You can use this to identify the companies you can query about, and their ticker information.
-        """
-        return tickers
-
-    def get_valid_years() -> list[str]:
-        """
-        Returns a list of the years for which financial reports are available.
-        Always check this before using any other tool.
-        """
-        return years
     
-    # Tool to get the income statement for a given company and year using the FMP API
-    def get_income_statement(
-        ticker=Field(description="the ticker symbol of the company."),
-        year=Field(description="the year for which to get the income statement."),
-    ) -> str:
-        """
-        Get the income statement for a given company and year using the FMP (https://financialmodelingprep.com) API.
-        Returns a dictionary with the income statement data. All data is in USD, but you can convert it to more compact form like K, M, B.
-        """
-        fmp_api_key = os.environ.get("FMP_API_KEY", None)
-        if fmp_api_key is None:
-            return "FMP_API_KEY environment variable not set. This tool does not work."
-        url = f"https://financialmodelingprep.com/api/v3/income-statement/{ticker}?apikey={fmp_api_key}"
-        response = requests.get(url)
-        if response.status_code == 200:
-            data = response.json()
-            income_statement = pd.DataFrame(data)
-            income_statement["date"] = pd.to_datetime(income_statement["date"])
-            income_statement_specific_year = income_statement[
-                income_statement["date"].dt.year == int(year)
-            ]
-            values_dict = income_statement_specific_year.to_dict(orient="records")[0]
-            return f"Financial results: {', '.join([f'{key}: {value}' for key, value in values_dict.items() if key not in ['date', 'cik', 'link', 'finalLink']])}"
-        else:
-            return "FMP API returned error. This tool does not work."
-
-    class QueryTranscriptsArgs(BaseModel):
+    class QueryHackerNews(BaseModel):
         query: str = Field(..., description="The user query.")
-        year: int = Field(..., description=f"The year. an integer between {min(years)} and {max(years)}.")
-        ticker: str = Field(..., description=f"The company ticker. Must be a valid ticket symbol from the list {tickers.keys()}.")
 
     tools_factory = ToolsFactory(vectara_api_key=cfg.api_key, 
                                  vectara_customer_id=cfg.customer_id, 
                                  vectara_corpus_id=cfg.corpus_id)
-    ask_transcripts = tools_factory.create_rag_tool(
-        tool_name = "ask_transcripts",
+    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 = """
-        Given a company name and year, 
-        returns a response (str) to a user question about a company, based on analyst call transcripts about the company's financial reports for that year.
-        You can ask this tool any question about the compaany including risks, opportunities, financial performance, competitors and more.
-        make sure to provide the a valid company ticker and year.
+        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 = QueryTranscriptsArgs,
-        tool_filter_template = "doc.year = {year} and doc.ticker = '{ticker}'",
+        tool_args_schema = QueryHackerNews,
         reranker = "multilingual_reranker_v1", rerank_k = 100, 
-        n_sentences_before = 2, n_sentences_after = 2, lambda_val = 0.01,
+        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,
     )
 
-    return (tools_factory.get_tools(
+    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_company_info, 
-                    get_valid_years,
-                    get_income_statement,
+                    get_top_stories, 
+                    get_show_stories,
+                    get_ask_stories,
+                    get_story_details,
                 ]
             ) +
-            tools_factory.standard_tools() + 
-            tools_factory.financial_tools() + 
-            tools_factory.guardrail_tools() +
-            [ask_transcripts]
+        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")
-    financial_bot_instructions = f"""
-    - You are a helpful financial assistant, with expertise in finanal reporting, in conversation with a user. 
+    bot_instructions = f"""
+    - You are a helpful assistant, answering user questions about content from hacker news. 
     - Today's date is {date}.
-    - Guardrails: never discuss politics, and always respond politely.
-    - Respond in a compact format by using appropriate units of measure (e.g., K for thousands, M for millions, B for billions). 
-      Do not report the same number twice (e.g. $100K and 100,000 USD).
-    - Use tools when available instead of depending on your own knowledge.
+    - 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.
-    - When querying a tool for a numeric value or KPI, use a concise and non-ambiguous description of what you are looking for. 
-    - If you calculate a metric, make sure you have all the necessary information to complete the calculation. Don't guess.
     - Be very careful not to report results you are not confident about.
     """
 
@@ -134,8 +133,8 @@ def initialize_agent(_cfg):
 
     agent = Agent(
         tools=create_tools(_cfg),
-        topic="10-K annual financial reports",
-        custom_instructions=financial_bot_instructions,
+        topic="hacker news",
+        custom_instructions=bot_instructions,
         update_func=update_func
     )
     return agent
@@ -149,7 +148,7 @@ def launch_bot():
         st.session_state.log_messages = []
         st.session_state.show_logs = False
 
-    st.set_page_config(page_title="Financial Assistant", layout="wide")
+    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']),
@@ -164,11 +163,7 @@ def launch_bot():
     with st.sidebar:
         image = Image.open('Vectara-logo.png')
         st.image(image, width=250)
-        st.markdown("## Welcome to the financial assistant demo.\n\n\n")
-        companies = ", ".join(tickers.values())
-        st.markdown(
-            f"This assistant can help you with any questions about the financials of several companies:\n\n **{companies}**.\n"
-        )
+        st.markdown("## Welcome to the hacker news assistant demo.\n\n\n")
 
         st.markdown("\n\n")
         bc1, bc2 = st.columns([1, 1])
@@ -208,8 +203,8 @@ def launch_bot():
         with st.chat_message("assistant", avatar='🤖'):
             with st.spinner(st.session_state.thinking_message):
                 res = st.session_state.agent.chat(prompt)
-                cleaned = re.sub(r'\[\d+\]', '', res).replace('$', '\\$')
-            message = {"role": "assistant", "content": cleaned, "avatar": '🤖'}
+                res = res.replace('$', '\\$')  # escape dollar sign for markdown
+            message = {"role": "assistant", "content": res, "avatar": '🤖'}
             st.session_state.messages.append(message)
             st.rerun()