Upload 5 files

app.py

@@ -0,0 +1,44 @@
+import gradio as gr
+from config import settings
+import constants
+from llms import Agents
+import datetime
+
+agents = Agents(model="gpt-4o", model_type="openai", temperature=0.5, is_structured=False)
+
+
+def get_county_info(county_name):
+    today = datetime.datetime.today().strftime("%Y-%m-%d")
+    user_input = {"county": county_name, "today": today}
+    response = agents.generate_tavily_search(f"{user_input}")
+    return response["output"], gr.update(visible=False)
+    # return "## 👇 Please select a county and click submit."
+
+
+# A list of county names for the dropdown.
+county_choices = constants.county_list
+
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# IBHS County Profiler")
+    gr.Markdown("#### A specialized research agent that gather up-to-date, authoritative information on residential risk for a county.")
+    gr.Markdown("Select a county from the list and click 'Submit' to see the details.")
+
+    with gr.Column():
+        county_dropdown = gr.Dropdown(
+            choices=county_choices,
+            label="Select a County"
+        )
+        submit_button = gr.Button("Submit")
+        progress = gr.Textbox(label="Progress bar", value="Status: Waiting for input...", interactive=False)
+
+    output_markdown = gr.Markdown()
+
+    # The click event of the button now triggers the function
+    submit_button.click(
+        fn=get_county_info,
+        inputs=county_dropdown,
+        outputs=[output_markdown, progress]
+    )
+
+if __name__ == "__main__":
+    demo.launch()

config.py

@@ -0,0 +1,37 @@
+from pydantic_settings import BaseSettings
+
+
+class Settings(BaseSettings):
+
+    COUNTY_RISK_ANALYSIS_PROMPT: str = """
+    You are “SafeHaven Analyst,” a specialized research agent that uses the Tavily search API to gather up-to-date, authoritative information on residential risk for a single U.S. county.
+
+    ⮞ **Inputs provided by the user as a dictionary with following items**  
+    - `county`: full county name plus state abbreviation (e.g., “San Bernardino County, CA”).  
+    - `today`: ISO date string (YYYY-MM-DD) representing “today.”
+
+    ⮞ **Your task**  
+    1. Run focused Tavily searches (news, government, academic, and insurer sources) published **within the last 3 years** that mention (`county` AND key terms below).  
+       - Key terms: *storm*, *flood*, *wildfire*, *hail*, *tornado*, *hurricane*, *drought*, *extreme heat*, *seismic*, *landslide*, *insurance claims*, *roof damage cost*, *window replacement cost*, *door replacement cost*, *home repair cost*.  
+    2. Analyse the top credible sources to extract:  
+       **a. Weather & Natural Hazard Landscape** – List the *current* primary hazards (likelihood + seasonal timing).  
+       **b. Damage-Cost Benchmarks** – Typical 2024-25 cost ranges (USD) in this county for repairing/replacing: roof, exterior doors, windows. Cite the source or methodology briefly in parentheses.  
+       **c. Overall Residential Risk Profile** – Concise narrative (≈120 words) rating risk *Low / Moderate / High* and explaining drivers (age of local housing stock, hazard frequency, rebuild costs, insurance availability, etc.).  
+    3. Prioritize clarity: synthesize, do not paste source snippets. If data is scarce, state that and default to statewide or regional averages, labelling them clearly.  
+    4. **Inline-cite every numeric or factual claim** with a markdown link:  
+       - Format: `**[$12 k–$18 k, [HomeAdvisor 05/2025](https://source-url.com)]**` – one citation per sentence.  
+       - Use the article/site name and month/year; the link must point to the live URL returned by Tavily.  
+    5. Total report length ≤ 300 words, using these headings exactly:  
+       **Weather & Hazard Snapshot**  
+       **Damage-Cost Benchmarks (USD)**  
+       **County Home Risk Profile**
+
+    ⮞ **Style guidelines**  
+    - Formal but readable business prose.  
+    - Use bullet lists only inside the two first sections; profile section is paragraph form.  
+    - No markdown code fences or JSON—return plain text with the three headings only.  
+    """
+
+
+settings = Settings()
+

constants.py

@@ -0,0 +1,60 @@
+county_list =[
+    "Alameda County, CA",
+    "Alpine County, CA",
+    "Amador County, CA",
+    "Butte County, CA",
+    "Calaveras County, CA",
+    "Colusa County, CA",
+    "Contra Costa County, CA",
+    "Del Norte County, CA",
+    "El Dorado County, CA",
+    "Fresno County, CA",
+    "Glenn County, CA",
+    "Humboldt County, CA",
+    "Imperial County, CA",
+    "Inyo County, CA",
+    "Kern County, CA",
+    "Kings County, CA",
+    "Lake County, CA",
+    "Lassen County, CA",
+    "Los Angeles County, CA",
+    "Madera County, CA",
+    "Marin County, CA",
+    "Mariposa County, CA",
+    "Mendocino County, CA",
+    "Merced County, CA",
+    "Modoc County, CA",
+    "Mono County, CA",
+    "Monterey County, CA",
+    "Napa County, CA",
+    "Nevada County, CA",
+    "Orange County, CA",
+    "Placer County, CA",
+    "Plumas County, CA",
+    "Riverside County, CA",
+    "Sacramento County, CA",
+    "San Benito County, CA",
+    "San Bernardino County, CA",
+    "San Diego County, CA",
+    "San Francisco County, CA",
+    "San Joaquin County, CA",
+    "San Luis Obispo County, CA",
+    "San Mateo County, CA",
+    "Santa Barbara County, CA",
+    "Santa Clara County, CA",
+    "Santa Cruz County, CA",
+    "Shasta County, CA",
+    "Sierra County, CA",
+    "Siskiyou County, CA",
+    "Solano County, CA",
+    "Sonoma County, CA",
+    "Stanislaus County, CA",
+    "Sutter County, CA",
+    "Tehama County, CA",
+    "Trinity County, CA",
+    "Tulare County, CA",
+    "Tuolumne County, CA",
+    "Ventura County, CA",
+    "Yolo County, CA",
+    "Yuba County, CA"
+]

llms.py

@@ -0,0 +1,55 @@
+import datetime
+
+from langchain.agents import create_openai_tools_agent, AgentExecutor
+from langchain.chat_models import init_chat_model
+from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
+from langchain_openai import ChatOpenAI
+from langchain_tavily import TavilySearch
+from langchain.schema import HumanMessage, SystemMessage
+from langchain_core.prompts import PromptTemplate
+from langchain_openai import ChatOpenAI
+from langchain_anthropic import ChatAnthropic
+from pydantic import BaseModel, Field
+from typing import List
+from enum import Enum
+from langchain_core.output_parsers import PydanticOutputParser
+from langchain_core.output_parsers import JsonOutputParser
+import os
+from dotenv import load_dotenv
+load_dotenv()
+
+from config import settings
+
+
+class Agents:
+    def __init__(self, temperature: float = 0.3, model: str = "gpt-4.1", model_type: str = "openai",
+                 is_structured: bool = True):
+        """
+        Initializes the UserSummaryGenerator with necessary settings and model configuration.
+        """
+
+        self.settings = settings
+        self.model_type = model_type
+
+        self.llm_tavily = init_chat_model(model="gpt-4o", model_provider="openai", temperature=0)
+
+        self.tavily_search_tool = TavilySearch(max_results=5, topic="general")
+
+        self.tavily_prompt = ChatPromptTemplate.from_messages([
+            ("system", self.settings.COUNTY_RISK_ANALYSIS_PROMPT),
+            MessagesPlaceholder(variable_name="messages"),
+            MessagesPlaceholder(variable_name="agent_scratchpad"),  # Required for tool calls
+        ])
+
+        self.tavily_agent = create_openai_tools_agent(
+            llm=self.llm_tavily,
+            tools=[self.tavily_search_tool],
+            prompt=self.tavily_prompt
+        )
+
+        self.agent_executor = AgentExecutor(agent=self.tavily_agent, tools=[self.tavily_search_tool], verbose=False)
+
+
+    def generate_tavily_search(self, user_input: str) -> str:
+        response = self.agent_executor.invoke({"messages": [HumanMessage(content=user_input)]},)
+        return response

requirements.txt

@@ -0,0 +1,18 @@
+pandas
+openpyxl
+pandas
+pyarrow
+langchain
+langchain_core
+python-dotenv
+langchain-openai
+gradio
+pydantic-settings
+pyreadstat
+plotly
+kaleido
+chardet
+charset-normalizer
+pandas-stubs==2.2.3.250527
+langchain_anthropic
+langchain-tavily