Upload 2 files

.env

@@ -0,0 +1,15 @@
+PERSIST_DIRECTORY=db
+SOURCE_DIRECTORY=data
+EMBEDDINGS_MODEL_NAME=sentence-transformers/msmarco-distilbert-base-tas-b
+TRANSFORMERS_CACHE=models
+SQL_PWD=1ngA2!KoBIU0R9k3
+SQL_USR_NM=riscadmin
+SQL_HOST=oigspm.database.windows.net
+SQL_TBL=oigspm
+MODEL_NAME=riscmltest-lxnyq
+ENDPOINT_API_KEY=KpAQpyQEHITtvuFi3GPd56zVHe4TPCdX
+ENDPOINT_URL=https://riscmltest-lxnyq.eastus2.inference.ml.azure.com/score
+OAUTH_AZURE_AD_CLIENT_ID=123ce628-135a-4c91-9229-bc32b24337ff
+OAUTH_AZURE_AD_CLIENT_SECRET=XjM8Q~3MDupCgwOLxk1Vhi6Ds7EhEAUiVNaTVaXQ
+OAUTH_AZURE_AD_TENANT_ID=df44fade-4360-42c6-b484-06e69bf8d3d1
+OPENAI_API_KEY=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

sqlchat.py

@@ -0,0 +1,156 @@
+            
+from langchain.embeddings import HuggingFaceEmbeddings
+#from langchain.vectorstores import FAISS
+from langchain.schema import Document
+
+#from langchain.vectorstores import Chroma 
+from langchain.llms import AzureMLOnlineEndpoint
+from langchain.chat_models.azureml_endpoint import ContentFormatterBase
+import json
+
+from langchain.chains import create_sql_query_chain
+
+import chainlit as cl
+
+
+from typing import Dict
+
+
+# Now we can create the agent, adjusting the standard SQL Agent suffix to consider our use case. 
+# Although the most straightforward way to handle this would be to include it just in the tool description, 
+# this is often not enough and we need to specify it in the agent prompt using the suffix argument in the constructor.
+
+from langchain.agents import create_sql_agent, AgentType
+from langchain.agents.agent_toolkits import SQLDatabaseToolkit
+from langchain.utilities import SQLDatabase
+from langchain.chat_models import ChatOpenAI
+
+import os
+OPENAI_API_KEY = os.environ['OPENAI_API_KEY']
+
+
+
+
+def create_agent():
+   # conn_str = "mssql+pyodbc://" + SQL_USR_NM + ":" + PWD + "@" + SQL_HOST + "/" + SQL_TBL + "?driver=ODBC+Driver+18+for+SQL+Server"
+
+    # Create the SQLDatabase object
+    db = SQLDatabase.from_uri('sqlite:///spm.db')
+    llm = ChatOpenAI(temperature=0.05, model="gpt-3.5-turbo-16k-0613")
+    db_chain = SQLDatabaseChain.from_llm(llm, db, verbose=True)
+    return db_chain
+
+    #toolkit = SQLDatabaseToolkit(db=db, llm=llm)
+
+custom_suffix = """
+Compose a query in the All_data table in the db database.
+Here is a description of each column:
+destn_area_name: The name of the destination area.
+destn_district_name: The name of the destination district.
+score: The score of the destination area.
+avg_days_todelr: The average number of days to deliver to the destination area.
+time_per: The time period of the data.
+orgn_area: The code of the origin area.
+orgn_dist: The code of the origin district.
+orgn_area_name: The name of the origin area.
+orgn_dist_name: The name of the origin district.
+destn_area: The code of the destination area.
+destn_dist: The code of the destination district.
+destn_area_name: The name of the destination area.
+destn_dist_name: The name of the destination district.
+prodt: The product type.
+rptg_start_date: The start date of the reporting period.
+rptg_end_date: The end date of the reporting period.
+mo: The month of the reporting period.
+pstl_qtr: The quarter of the Postal reporting period.
+pstl_yr: The year of the Postal reporting period.
+score: The score of the destination area.
+score_plus_1: The score of the destination area plus 1.
+"""
+
+#agent = create_sql_agent(llm=llm,
+#                         toolkit=toolkit,
+#                         verbose=False,
+#                         agent_type=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
+#                         extra_tools=custom_tool_list,
+#                         suffix=custom_suffix, 
+#                         handle_parsing_errors=True
+#                        )
+
+from langchain.prompts import PromptTemplate
+
+
+def build_sql_chain(llm, db):
+    
+    dialect = "Azure SQL"
+    table_info = "All_data"
+    few_shots = {"What are the top 10 performing areas?": "SELECT TOP 10 destn_area_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_area_name ORDER BY AvgScore DESC",
+                        "What are the worst 10 performing areas?": "SELECT TOP 10 destn_area_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_area_name ORDER BY AvgScore ASC",
+                        "What districts have the highest volume of mail?": "SELECT TOP 10 destn_district_name, COUNT(*) AS Volume FROM All_data GROUP BY destn_district_name ORDER BY Volume DESC",
+                        "What districts have the lowest volume of mail?": "SELECT TOP 10 destn_district_name, COUNT(*) AS Volume FROM All_data GROUP BY destn_district_name ORDER BY Volume ASC",
+                        "What are the top 10 performing districts?": "SELECT TOP 10 destn_district_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_district_name ORDER BY AvgScore DESC",
+                        "What are the worst 10 performing districts?": "SELECT TOP 10 destn_district_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_district_name ORDER BY AvgScore ASC",
+                        "What districts gave the fastest delivery time?": "SELECT TOP 10 destn_district_name, AVG(avg_days_todelr) AS AvgDeliveryTime FROM All_data GROUP BY destn_district_name ORDER BY AvgDeliveryTime ASC"}
+    fs = str(few_shots)
+                        
+
+    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}.
+
+    Some examples of SQL queries that correspond to questions are:
+
+    \{"What are the top 10 performing areas?": "SELECT TOP 10 destn_area_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_area_name ORDER BY AvgScore DESC",
+     "What are the worst 10 performing areas?": "SELECT TOP 10 destn_area_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_area_name ORDER BY AvgScore ASC",
+     "What districts have the highest volume of mail?": "SELECT TOP 10 destn_district_name, COUNT(*) AS Volume FROM All_data GROUP BY destn_district_name ORDER BY Volume DESC",
+     "What districts have the lowest volume of mail?": "SELECT TOP 10 destn_district_name, COUNT(*) AS Volume FROM All_data GROUP BY destn_district_name ORDER BY Volume ASC",
+     "What are the top 10 performing districts?": "SELECT TOP 10 destn_district_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_district_name ORDER BY AvgScore DESC",
+     "What are the worst 10 performing districts?": "SELECT TOP 10 destn_district_name, AVG(score) AS AvgScore FROM All_data GROUP BY destn_district_name ORDER BY AvgScore ASC",
+     "What districts gave the fastest delivery time?": "SELECT TOP 10 destn_district_name, AVG(avg_days_todelr) AS AvgDeliveryTime FROM All_data GROUP BY destn_district_name ORDER BY AvgDeliveryTime ASC"\}
+
+    Question: {input}"""
+
+    CUSTOM_PROMPT = PromptTemplate(
+        input_variables=["input", "table_info", "dialect"], template=TEMPLATE
+    )
+
+    # Set verbose=True to see the full prompt:
+    return create_sql_query_chain(llm=llm, db=db)
+
+#from langchain.llms import OpenAI
+from langchain_experimental.sql import SQLDatabaseChain
+#sql_chain = build_sql_chain(llm, db)
+
+
+@cl.on_chat_start
+async def main():
+    # Parse the command line arguments
+   # args = parse_arguments()
+    await cl.Message(content="Welcome to GeoData!").send()
+    
+    # activate/deactivate the streaming StdOut callback for LLMs
+    #callbacks = [StreamingStdOutCallbackHandler()]
+
+    #sql_chain = build_sql_chain(llm, db)
+
+
+@cl.on_message
+async def msg(message: str):
+    # Retrieve the chain from the user session
+    #sql_chain = cl.user_session.get("sql_chain")  # type: RetrievalQA
+    agent = create_agent()
+    m = message.content
+    #res = sql_chain.invoke({"question": m})
+    res = agent.run({"query": m})
+    # Call the chain asynchronously
+
+    print(res)
+    await cl.Message(content=res).send()