Update

.gitignore

@@ -46,4 +46,6 @@ lib/
 # Scrapy
 .scrapy/
 
-_*.py
+_*.py
+*.db-shm
+*.db-wal

app.py

@@ -1,35 +1,80 @@
 # Standard library imports
-import os.path
+import os
+import re
+import sqlite3
+from typing import List, Tuple, Optional
 
 # Third-party imports
 import gradio as gr
 from dotenv import load_dotenv
-from langchain.agents import AgentExecutor, create_openai_tools_agent
-from langchain.prompts import ChatPromptTemplate
-from langchain.schema import HumanMessage, AIMessage
-from langchain_community.utilities import SQLDatabase
-from langchain_community.agent_toolkits.sql.toolkit import SQLDatabaseToolkit
-from langchain_openai import ChatOpenAI
+
+# SmolaGents imports
+from smolagents import CodeAgent, LiteLLMModel, tool
+from smolagents.agent_types import AgentText
 
 # Local imports
-from init_db import Project, db_description, json_structure
+from init_db import Project
 
 # Load environment variables
 load_dotenv()
 
-# Check if database exists
-if not os.path.exists('estate.db'):
-    raise FileNotFoundError(
-        "Database file 'estate.db' not found. Please run 'uv run init_db.py' first."
-    )
+@tool
+def sql_engine(query: str) -> List[Tuple]:
+    """
+    Allows you to perform SQL queries on the table in SQLite database. Returns list of results.
+    The table is named 'project'. Its description is as follows:
+        Columns:
+        - url: VARCHAR(255) - URL of the project
+        - title: VARCHAR(255) - Title of the project
+        - deposit: INTEGER - Percentage of payment before completion
+        - min_price: INTEGER - Lowest available apartment price in CZK with VAT
+        - status: VARCHAR(255) - Status of the project (preparation, selling, sold out)
+        - city: VARCHAR(255) - City of the project
+        - lat: FLOAT - GPS latitude coordinates
+        - lng: FLOAT - GPS longitude coordinates
+        - start_year: INTEGER - Year of construction start
+        - end_year: INTEGER - Year of construction end
+        - developer: VARCHAR(255) - Name of the construction company
+        - ignore: BOOLEAN - If True, the project does not have any apartments for sale
+        
+        # Apartment prices by type (all INTEGER in CZK with VAT)
+        - price_1kk: Price of 1+kk apartment
+        - price_2kk: Price of 2+kk apartment
+        - price_3kk: Price of 3+kk apartment
+        - price_4kk: Price of 4+kk apartment
+        - price_5kk: Price of 5+kk apartment
+        - price_6kk: Price of 6+kk apartment
+        - price_7kk: Price of 7+kk apartment
+        - price_8kk: Price of 8+kk apartment
+        - price_9kk: Price of 9+kk apartment
+        - price_10kk: Price of 10+kk apartment
+        
+        - price_1_1: Price of 1+1 apartment
+        - price_2_1: Price of 2+1 apartment
+        - price_3_1: Price of 3+1 apartment
+        - price_4_1: Price of 4+1 apartment
+        - price_5_1: Price of 5+1 apartment
+        - price_6_1: Price of 6+1 apartment
+        - price_7_1: Price of 7+1 apartment
+        - price_8_1: Price of 8+1 apartment
+        - price_9_1: Price of 9+1 apartment
+        - price_10_1: Price of 10+1 apartment
+        
+        - content: TEXT - Raw content (ignore to save tokens)
+        - created_at: DATETIME - Date and time of creation
+
+    Args:
+        query: The query to perform. This should be correct SQL.
+    """
+    con = sqlite3.connect("estate.db")
+
+    result = con.execute(query)
+    all_rows = result.fetchall()
+
+    con.close()
+    return all_rows
 
-# Initialize model and database
-model = ChatOpenAI(model="o3-mini-2025-01-31", streaming=True)
-db = SQLDatabase.from_uri("sqlite:///estate.db")
 
-# Set up SQL toolkit and tools
-toolkit = SQLDatabaseToolkit(db=db, llm=model)
-tools = toolkit.get_tools()
 
 sql_distance_query = """
 SELECT 
@@ -49,40 +94,17 @@ ORDER BY distance ASC
 LIMIT 5;
 """
 
-# Get the OpenAI tools agent prompt
-prompt = ChatPromptTemplate.from_messages([
-  ("system", f"You are a helpful assistant that presents information about real estate projects in database for investors. Use 'IS NOT NULL' when necessary. Use Czech city names. Do not present SQL or JSON. Use simple language. Do not make up information or hallucinate. If final result rows are empty respond that you could not find records matching criteria. Table structure: {db_description}. This JSON structure info can help you: {json_structure}."),
-  ("placeholder", "{chat_history}"),
-  ("human", "{input}"),
-  ("placeholder", "{agent_scratchpad}"),
-])
-
-# Create the agent with OpenAI tools format
-agent = create_openai_tools_agent(
-    llm=model,
-    tools=tools,
-    prompt=prompt
-)
 
-# Create agent executor
-agent_executor = AgentExecutor(
-    agent=agent,
-    tools=tools,
-    verbose=True
-)
+litemodel = LiteLLMModel("o3-mini-2025-01-31", api_base="https://api.openai.com/v1", api_key=os.getenv("OPENAI_API_KEY"))
+codeagent = CodeAgent(tools=[sql_engine], model=litemodel)
 
 def chat_with_sql(message, history):
     try:
-      history_langchain_format = []
-      for msg in history:
-          if msg['role'] == "user":
-              history_langchain_format.append(HumanMessage(content=msg['content']))
-          elif msg['role'] == "assistant":
-              history_langchain_format.append(AIMessage(content=msg['content']))
-      history_langchain_format.append(HumanMessage(content=message))
-      response = agent_executor.invoke({"input": message, "history": history_langchain_format})
-      for i in range(len(response["output"])):
-          yield response["output"][:i+1]
+      input_message = "You are a helpful assistant that presents information about real estate projects in database for investors. Use 'IS NOT NULL' when necessary. Translate any city names from input into Czech language. Present useful fields such as url. Do not make up information or hallucinate. If final result rows are empty respond that you could not find records matching criteria. Use only Czech city names. Note that there is limit 10 records for database result. Reformat final answer in markdown. User Input:\n\n {message}"
+      prompt = input_message.format(message=message)
+      result = codeagent.run(prompt)
+      yield  gr.ChatMessage(role="assistant", content=result)
+
     except Exception as e:
         yield f"Error: {str(e)}"
 

estate.db

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d848c98537602c1b22c1464e08b2662446e9cd32c8d48324b177baca184a253c
+oid sha256:e78ec9ee34c23b783b6451dc9107cf6e1af1ec35fdee048289bbac32503682d0
 size 2560000

init_db.py

@@ -1,62 +1,80 @@
 from peewee import *
 from datetime import datetime
 from playhouse.sqlite_ext import *
+import json
 
 # Initialize database
-
-json_structure = """JSON column with hash with keys
-
-'title': str, official title of the project
-'deposit': int (in percentage), sum of all payments before the apartment is complete ("koladace" requested)
-'cheapest_available_apartments' : hash with keys:
-    '1+kk': int, lowest price for 1+kk with VAT in CZK. It should not be booked or sold
-    '2+kk': int
-    '3+kk': int
-    '4+kk': int
-    '5+kk': int
-    '7+kk': int
-    '8+kk': int
-    '9+kk': int
-    '10+kk': int
-    '1+1': int
-    '2+1': int
-    '3+1': int
-    '4+1': int
-    '5+1': int
-    '6+1': int
-    '7+1': int
-    '8+1': int
-    '9+1': int
-    '10+1': int
- 'min_price': int, lowest available apartment price in CZK with VAT. It should not be booked or sold.
- 'status': str, status of the project (preparation, selling, sold out)
- 'city': str, city of the project
- 'lat': float, gps coordinates of the project
- 'lng': float, gps coordinates of the project
- 'start_year': int, year of construction start
- 'end_year': int, estimated year of construction end
- 'developer': str, name of the contruction company/"developer"
- 'ignore': bool, if True, the project does not have any apartments for sale. Some values can be null/unknown"""
 db = SqliteExtDatabase('estate.db')
 
-db_description = f"""Table "project" - list of real estate projects (novostavby) in Czech Republic
+db_description = """Table "project" - list of real estate projects (novostavby) in Czech Republic
     url: url of the project.
-    structure: {json_structure}.
-    content: unstructured additional information about the project.
-    created_at: date and time of creation
     title: title of the project
+    deposit: percentage of payment before completion
+    min_price: lowest available apartment price in CZK with VAT
+    status: status of the project (preparation, selling, sold out)
+    city: city of the project
+    lat: GPS latitude coordinates
+    lng: GPS longitude coordinates
+    start_year: year of construction start
+    end_year: estimated year of construction end
+    developer: name of the construction company
+    ignore: if True, the project does not have any apartments for sale
+    
+    Apartment prices are stored in separate columns for each type:
+    price_1kk, price_2kk, price_3kk, etc.
 """
 
 class Project(Model):
+    # Basic information
     url = CharField(unique=True)
     title = CharField(null=True)
-    structure = JSONField(null=True)
-    content = TextField(null=True)
+    
+    # Financial information
+    deposit = IntegerField(null=True)  # percentage
+    min_price = IntegerField(null=True)  # CZK with VAT
+    
+    # Apartment prices by type
+    price_1kk = IntegerField(null=True)
+    price_2kk = IntegerField(null=True)
+    price_3kk = IntegerField(null=True)
+    price_4kk = IntegerField(null=True)
+    price_5kk = IntegerField(null=True)
+    price_6kk = IntegerField(null=True)
+    price_7kk = IntegerField(null=True)
+    price_8kk = IntegerField(null=True)
+    price_9kk = IntegerField(null=True)
+    price_10kk = IntegerField(null=True)
+    
+    price_1_1 = IntegerField(null=True)
+    price_2_1 = IntegerField(null=True)
+    price_3_1 = IntegerField(null=True)
+    price_4_1 = IntegerField(null=True)
+    price_5_1 = IntegerField(null=True)
+    price_6_1 = IntegerField(null=True)
+    price_7_1 = IntegerField(null=True)
+    price_8_1 = IntegerField(null=True)
+    price_9_1 = IntegerField(null=True)
+    price_10_1 = IntegerField(null=True)
+    
+    # Project information
+    status = CharField(null=True)
+    city = CharField(null=True)
+    lat = FloatField(null=True)
+    lng = FloatField(null=True)
+    start_year = IntegerField(null=True)
+    end_year = IntegerField(null=True)
+    developer = CharField(null=True)
+    
+    # Additional data
+    ignore = BooleanField(null=True)
+    content = TextField(null=True)  # Keep raw content for reference
+    structure = JSONField(null=True)  # Keep JSON for backward compatibility
     created_at = DateTimeField(default=datetime.now)
 
     class Meta:
         database = db
 
+
 def init_database():
     """Initialize the database and create tables"""
     print("Initializing database...")
@@ -64,17 +82,7 @@ def init_database():
     db.drop_tables([Project])
     db.create_tables([Project])
     print("Created tables successfully!")
-    
-    # Add some test data if needed
-    test_urls = [
-        'https://brnojedna.cz/'
-    ]
-    
-    for url in test_urls:
-        Project.get_or_create(url=url)
-    
-    print(f"Number of projects in database: {Project.select().count()}")
     db.close()
 
 if __name__ == "__main__":
-    init_database() 
+    init_database()

requirements.in

@@ -1,12 +1,9 @@
 gradio
-langchain
-langchain-community
-langchain-openai
 python-dotenv
 beautifulsoup4
 requests
 peewee
 openai
-crawl4ai
 datasketch
-geocoder
+geocoder
+smolagents[litellm]