Upload app.py with huggingface_hub

app.py

@@ -12,7 +12,6 @@ import weave
 from langchain.memory import ConversationSummaryBufferMemory
 
 from langchain.embeddings import OpenAIEmbeddings
-from langchain.chat_models import ChatOpenAI
 from langchain_openai import OpenAIEmbeddings
 from langchain_community.embeddings.sentence_transformer import SentenceTransformerEmbeddings
 from langchain_community.utilities.sql_database import SQLDatabase
@@ -20,15 +19,10 @@ from langchain_community.agent_toolkits import create_sql_agent
 from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
 from langchain_core.tools import tool
 from langchain.agents import create_tool_calling_agent, AgentExecutor
-from langchain.agents import initialize_agent, Tool, AgentType
-from langchain.llms import OpenAI
 from langchain_community.vectorstores import Chroma
 from huggingface_hub import CommitScheduler
-from pydantic import BaseModel, Field, validator
-from langchain.llms import OpenAI
 from pathlib import Path
 
-
 from langchain_core.caches import BaseCache
 from langchain_openai.chat_models import ChatOpenAI
 
@@ -39,16 +33,9 @@ ChatOpenAI.model_rebuild()
 
 model_name = "gpt-4o-mini"
 
-api_key = os.getenv("OPENAI_API_KEY")
-endpoint = os.getenv("OPENAI_API_BASE")
-
-
-
-llm = ChatOpenAI(
-    openai_api_base=endpoint,  # Fill in the endpoint
-    openai_api_key="gl-U2FsdGVkX1/DTLQlsmj+RdJjPy3igB9qINuaX940XtJ0CPnGc/5sbBkPKah/C829",  # Fill in the API key
-    model="gpt-4o-mini",  # Fill in the deployment name (e.g., gpt-4o-mini)
-    streaming=False)
+#config = read_config("config_GANLP.json")  #Copy and paste the path of the config file uploaded in Colab
+api_key = "gl-U2FsdGVkX1/DTLQlsmj+RdJjPy3igB9qINuaX940XtJ0CPnGc/5sbBkPKah/C829"
+endpoint = "https://aibe.mygreatlearning.com/openai/v1"
 
 # Define the location of the SQLite database
 db_loc = 'ecomm.db'
@@ -117,6 +104,14 @@ full_prompt = ChatPromptTemplate.from_messages(
     ]
 )
 
+#Initialize the ChatOpenAI model with the extracted configuration
+# llm = ChatOpenAI(
+#     endpoint="https://aibe.mygreatlearning.com/openai/v1",
+#     api_key="gl-U2FsdGVkX1/DTLQlsmj+RdJjPy3igB9qINuaX940XtJ0CPnGc/5sbBkPKah/C829",
+#     model="gpt-4o-mini",
+#     streaming=False  # Explicitly disabling streaming
+# )
+
 
 # Create the SQL agent using the ChatOpenAI model, database, and prompt template
 sqlite_agent = create_sql_agent(
@@ -128,63 +123,31 @@ sqlite_agent = create_sql_agent(
     max_iterations=5,
     verbose=True
 )
+#### Let's convert the sql agent into a tool that our fin agent can use.
 
-conn = sqlite3.connect('ecomm.db')
-cursor = conn.cursor()
-
-
-
-def def_tool(query: str):
+@tool
+def sql_tool(user_input):
     """
-    Executes a SQL query using the agent and returns the result based on the natural language query.
-
+    Gathers information regarding purchases, transactions, returns, refunds, etc.
+    Executes a SQL query using the sqlite_agent and returns the result.
     Args:
-        query (str): A natural language query string to be processed by the agent.
-
+        user_input (str): a natural language query string explaining what information is required while also providing the necessary details to get the information.
     Returns:
-        str: The result of the SQL query execution or any error message.
+        str: The result of the SQL query execution. If an error occurs, the exception is returned as a string.
     """
     try:
-        # Define the sql_tool function which is responsible for executing the query
-        @tool
-        def sql_tool(query: str):
-            """Executes a SQL query on the e-commerce database and returns the result."""
-            try:
-                conn = sqlite3.connect('/ecomm.db')
-                cursor = conn.cursor()
-                cursor.execute(query)
-                result = cursor.fetchall()
-                conn.close()
-                return result
-            except Exception as e:
-                return str(e)
-
-        # Define the tools array for agent
-        tools = [
-            Tool(
-                name="SQLite Tool",
-                func=sql_tool,
-                description="Execute SQL queries"
-            )
-        ]
-
-        # Initialize the agent with the tools
-        agent = initialize_agent(
-            tools,
-            agent_type=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
-            llm=llm,  # Assuming llm is predefined somewhere
-            verbose=True
-        )
-
-        # Use the agent to run the natural language query
-        response = agent.run(query)
+        # Invoke the sqlite_agent with the user input (SQL query)
+        response = sqlite_agent.invoke(user_input)
 
-        # Return the response
-        return response
+        # Extract the output from the response
+        prediction = response['output']
 
     except Exception as e:
-        # If an error occurs in the process, capture and return the exception
-        return str(e)
+        # If an exception occurs, capture the exception message
+        prediction = e
+
+    # Return the result or the exception message
+    return prediction
 
 #=================================== RAG TOOL======================================#
 qna_system_message = """
@@ -238,10 +201,10 @@ retriever = vector_store.as_retriever(
     search_kwargs={'k': 5}
 )
 
-# client = OpenAI(
-#     api_key=api_key,
-#     base_url=endpoint
-# )
+client = OpenAI(
+    api_key=api_key,
+    base_url=endpoint
+)
 
 @tool
 def rag(user_input: str) -> str:
@@ -268,7 +231,7 @@ def rag(user_input: str) -> str:
     }
     ]
     try:
-        response = llm.chat.completions.create(
+        response = client.chat.completions.create(
         model="gpt-4o-mini",
         messages=prompt
         )
@@ -527,7 +490,16 @@ def chatbot_interface():
     prompt = build_prompt(details)
     tools = [sql_tool,defer_to_human, rag, register_feedback, days_since]
 
-    agent = create_tool_calling_agent(llm, tools, prompt)
+
+    chatbot = ChatOpenAI(
+    openai_api_base=endpoint,
+    openai_api_key=api_key,
+    model="gpt-4o-mini",
+    streaming=False, # Explicitly disabling streaming
+    temperature=0
+    )
+
+    agent = create_tool_calling_agent(chatbot, tools, prompt)
     agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
 
     # Display chat messages from history on app rerun