Add application file

README.md

@@ -1,7 +1,7 @@
 ---
-title: Edai Langchain Chat Testing
+title: Edai Analyst Chatbot Gradio Space
 emoji: ⚡
-colorFrom: purple
+colorFrom: green
 colorTo: gray
 sdk: gradio
 sdk_version: 3.32.0

app.py

@@ -1,8 +1,11 @@
 # from langchain.agents import create_sql_agent
 from langchain.agents.agent_toolkits import SQLDatabaseToolkit
 from langchain.sql_database import SQLDatabase
-# from langchain.llms.openai import OpenAI
+from langchain.llms.openai import OpenAI
+from langchain import OpenAI, SQLDatabase, SQLDatabaseChain
+from langchain.prompts.prompt import PromptTemplate
 # from langchain.agents import AgentExecutor
+
 import openai
 import json
 import gradio as gr
@@ -12,6 +15,8 @@ import os
 import gradio
 import datetime
 
+
+
 postgres_connection_str = os.environ['POSTGRES_CONNECTION_STR']
 access_key = os.environ['AWS_ACCESS_KEY_ID']
 secret_key = os.environ['AWS_SECRET_ACCESS_KEY']
@@ -23,10 +28,10 @@ s3_client = boto3.client('s3', aws_access_key_id=access_key, aws_secret_access_k
 llm = OpenAI(temperature=0, verbose=True, openai_api_key=openai_api_key)
 toolkit = SQLDatabaseToolkit(db=db, llm=llm)
 
-agent_executor = create_sql_agent(
-    llm=llm,
-    toolkit=toolkit
-)
+# agent_executor = create_sql_agent(
+#     llm=llm,
+#     toolkit=toolkit
+# )
 
 # Function to insert the question and response into the table
 def insert_question_response(question, response):
@@ -45,21 +50,91 @@ def insert_question_response(question, response):
 
 def generate_response(question,question2):
 
-    prompt_template = """
-     Keep in mind that any site, building, or property related question should be routed to the real estate portal. 
-     Any Local or city or city-sector incentives programs are asking about the local incentives program table. 
-     Any State incentives programs are asking about the state incentives program table.
-     If you can't find the answer, make sure to look up the program field in the local and state incentives catalogs.
-     If your final answer is "I don't know", then respond with "Please adjust your question and try asking again."
-     """
-
-    chain_response = agent_executor.run(question + prompt_template)
-    # Insert the question and response into the PostgreSQL table
-    insert_question_response(question, chain_response)
-    bucket_name = 'edai-analytics'
+    _DEFAULT_TEMPLATE = """Given an input question, first create a syntactically correct {dialect} query to run, then look at the results of the query and return the answer.
+                            Use the following format:
+
+                            Question: "Question here"
+                            SQLQuery: "SQL Query to run"
+                            SQLResult: "Result of the SQLQuery"
+                            Answer: "Final answer here"
+
+                            Only use the following tables:
+
+                            {table_info}
+
+                            If someone asks about real estate or available properties, they mean the properties_us_sites_and_buildings table. To filter between a site and a building, use the type field.
+
+                            Question: {input}"""
+
+    PROMPT = PromptTemplate(
+        input_variables=["input", "table_info", "dialect"], template=_DEFAULT_TEMPLATE
+    )
+
+
+
+    custom_table_info = {
+        "properties_us_sites_and_buildings": """This table is the real estate table."""
+    }
+
+    db = SQLDatabase.from_uri(
+        postgres_connection_str,
+        schema='langchain_testing',
+        include_tables=[
+                    'properties_us_sites_and_buildings',
+    #                  'state_incentives_catalog_dfd',
+    #                  'local_incentives_sources_dfd',
+    #                  'local_incentives_catalog_dfd',
+    #                  'located_localities_dfd',
+    #                  'located_metros_dfd',
+    #                  'located_states_dfd',
+                    'incentives_benefits_output',
+    #                  'incentives_benefits_program_det',
+            #                  'employment_industry_qcew_all_di',
+    #         'located_scores_states',
+    #          'located_scores_localities',
+    #                  'located_scores_metros',
+                    'incentives_benefits_rollups',
+                    'incentives_model_math_summary',
+                    'metro_codb_all_sectors_pnl',
+                    
+    #                  'fiscal_benefits_local',
+    #                  'fiscal_benefits_state',
+    #                  'fiscal_costs_local',
+    #                  'economic_impact_rollup',
+    #                  'economic_impact_sector_details',
+
+    #                  'industry_clusters_locality',
+    #                  'industry_clusters_metro',
+    #                  'occupation_oews',
+    #                  'occupation_oews_state'
+                    ],
+        sample_rows_in_table_info=1,
+    #     custom_table_info=custom_table_info
+    )
+    # db.table_info
+    print(db.table_info)
+
+    db_chain = SQLDatabaseChain.from_llm(llm, db, prompt=PROMPT, verbose=True)#, return_intermediate_steps=True)
+
+    
+
+    # db_chain.
+
+    # prompt_template = """
+    #  Keep in mind that any site, building, or property related question should be routed to the real estate portal. 
+    #  Any Local or city or city-sector incentives programs are asking about the local incentives program table. 
+    #  Any State incentives programs are asking about the state incentives program table.
+    #  If you can't find the answer, make sure to look up the program field in the local and state incentives catalogs.
+    #  If your final answer is "I don't know", then respond with "Please adjust your question and try asking again."
+    #  """
+
+    # chain_response = agent_executor.run(question + prompt_template)
+    # # Insert the question and response into the PostgreSQL table
+    # insert_question_response(question, chain_response)
+    # bucket_name = 'edai-analytics'
     # file_name = 'flagged_text.txt'
     # s3_client.put_object(Body=input_text, Bucket=bucket_name, Key=file_name)
-    return chain_response
+    return db_chain.run("How many sites are available in VA?")
 
 
 # with gr.Blocks(theme='snehilsanyal/scikit-learn',title='EDai Chatbot') as demo: