Using NLSQLTableQueryEngine and SQLTableRetrieverQueryEngine with chat engine

[Grkrish2002]

Pls let me know if there is a way to use chat engine with NLSQLTableQueryEngine and SQLTableRetrieverQueryEngine. My requirement is to maintain the chat memory / context for Natural Language to SQL

[dosubot[bot]]

🤖

Hey @Grkrish2002! 👋 I'm Dosu, your friendly neighborhood bot here to lend a hand with bugs, answer your questions, and guide you on becoming a contributor while you wait for a real human to step in. Just a heads-up, I’m diving into your issue about integrating chat engine with NLSQLTableQueryEngine and SQLTableRetrieverQueryEngine and will circle back soon with what I find. Hang tight! 🤖💡

[Grkrish2002]

Sure thank you

[dosubot[bot]]

🤖

To integrate a chat engine with NLSQLTableQueryEngine and SQLTableRetrieverQueryEngine to maintain chat memory/context for Natural Language to SQL, you would need to create a session or context management system in your chat engine that stores the state of the conversation. This state can include the user's queries and the system's responses.

When a new message comes in from the user, you would use the NLSQLTableQueryEngine to convert the natural language query to SQL. The SQL query is then executed using the SQLTableRetrieverQueryEngine to retrieve the relevant data. The response from the SQLTableRetrieverQueryEngine can then be used to generate the system's response to the user's query.

Here is a simplified example of how you might implement this in your chat engine:

class ChatEngine:
    def __init__(self, nlsql_engine, sql_retriever):
        self.nlsql_engine = nlsql_engine
        self.sql_retriever = sql_retriever
        self.context = {}

    def handle_message(self, user_id, message):
        # Create a context for the user if it doesn't exist
        if user_id not in self.context:
            self.context[user_id] = {}

        # Use the NLSQLTableQueryEngine to convert the message to SQL
        query_bundle = QueryBundle(query_str=message)
        response = self.nlsql_engine._query(query_bundle)

        # Execute the SQL query using the SQLTableRetrieverQueryEngine
        sql_query = response.metadata["sql_query"]
        data = self.sql_retriever.retrieve(sql_query)

        # Store the query and response in the user's context
        self.context[user_id][message] = data

        # Generate a response to the user's message
        response_message = self.generate_response(data)

        return response_message

    def generate_response(self, data):
        # This method would generate a natural language response based on the data
        pass

In this example, the ChatEngine class maintains a context for each user, storing their queries and the system's responses. When a new message is received from a user, it is converted to SQL using the NLSQLTableQueryEngine and the SQL query is executed using the SQLTableRetrieverQueryEngine. The result is then stored in the user's context and used to generate a response to the user's message.

Please note that this is a simplified example and the actual implementation would depend on the specifics of your chat engine and database. You can refer to the LlamaIndex codebase for more details on how the NLSQLTableQueryEngine and SQLTableRetrieverQueryEngine work.

Sources

• llama-index-core/llama_index/core/indices/struct_store/sql_query.py
• llama-index-legacy/llama_index/legacy/indices/struct_store/sql_retriever.py

About Dosu

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

[Grkrish2002]

Ok thanks. Let me check this

[adiba-shahana]

@Grkrish2002 did you find the solution

[Grkrish2002]

No, did not find a solution. Was looking for abstracted chat engine like implementation