-
-
Notifications
You must be signed in to change notification settings - Fork 373
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #144 from 0xThresh/feat-text-to-sql-rag
feat: New text-to-SQL pipe using LlamaIndex
- Loading branch information
Showing
2 changed files
with
98 additions
and
1 deletion.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,92 @@ | ||
""" | ||
title: Llama Index DB Pipeline | ||
author: 0xThresh | ||
date: 2024-07-01 | ||
version: 1.0 | ||
license: MIT | ||
description: A pipeline for using text-to-SQL for retrieving relevant information from a database using the Llama Index library. | ||
requirements: llama_index, sqlalchemy, psycopg2-binary | ||
""" | ||
|
||
from typing import List, Union, Generator, Iterator | ||
import os | ||
from llama_index.llms.ollama import Ollama | ||
from llama_index.core.query_engine import NLSQLTableQueryEngine | ||
from llama_index.core import SQLDatabase, PromptTemplate | ||
from sqlalchemy import create_engine | ||
) | ||
|
||
|
||
class Pipeline: | ||
def __init__(self): | ||
self.PG_HOST = os.environ["PG_HOST"] | ||
self.PG_PORT = os.environ["PG_PORT"] | ||
self.PG_USER = os.environ["PG_USER"] | ||
self.PG_PASSWORD = os.environ["PG_PASSWORD"] | ||
self.PG_DB = os.environ["PG_DB"] | ||
self.ollama_host = "http://host.docker.internal:11434" # Make sure to update with the URL of your Ollama host, such at http://localhost:11434 or remote server address | ||
self.model = "phi3:medium-128k" # Model to use for text-to-SQL generation | ||
self.engine = None | ||
self.nlsql_response = "" | ||
self.tables = ["db_table"] # Update to the name of the database table you want to get data from | ||
|
||
def init_db_connection(self): | ||
self.engine = create_engine(f"postgresql+psycopg2://{self.PG_USER}:{self.PG_PASSWORD}@{self.PG_HOST}:{self.PG_PORT}/{self.PG_DB}") | ||
return self.engine | ||
|
||
|
||
async def on_startup(self): | ||
# This function is called when the server is started. | ||
self.init_db_connection() | ||
|
||
async def on_shutdown(self): | ||
# This function is called when the server is stopped. | ||
pass | ||
|
||
def pipe( | ||
self, user_message: str, model_id: str, messages: List[dict], body: dict | ||
) -> Union[str, Generator, Iterator]: | ||
# Debug logging is required to see what SQL query is generated by the LlamaIndex library; enable on Pipelines server if needed | ||
|
||
# Create database reader for Postgres | ||
sql_database = SQLDatabase(self.engine, include_tables=self.tables) | ||
|
||
# Set up LLM connection; uses phi3 model with 128k context limit since some queries have returned 20k+ tokens | ||
llm = Ollama(model=self.model, base_url=self.ollama_host, request_timeout=180.0, context_window=30000) | ||
|
||
# Set up the custom prompt used when generating SQL queries from text | ||
text_to_sql_prompt = """ | ||
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. | ||
You can order the results by a relevant column to return the most interesting examples in the database. | ||
Unless the user specifies in the question a specific number of examples to obtain, query for at most 5 results using the LIMIT clause as per Postgres. You can order the results to return the most informative data in the database. | ||
Never query for all the columns from a specific table, only ask for a few relevant columns given the question. | ||
You should use DISTINCT statements and avoid returning duplicates wherever possible. | ||
Pay attention to use only the column names that you can see in the schema description. Be careful to not query for columns that do not exist. Pay attention to which column is in which table. Also, qualify column names with the table name when needed. You are required to use the following format, each taking one line: | ||
Question: Question here | ||
SQLQuery: SQL Query to run | ||
SQLResult: Result of the SQLQuery | ||
Answer: Final answer here | ||
Only use tables listed below. | ||
{schema} | ||
Question: {query_str} | ||
SQLQuery: | ||
""" | ||
|
||
text_to_sql_template = PromptTemplate(text_to_sql_prompt) | ||
|
||
query_engine = NLSQLTableQueryEngine( | ||
sql_database=sql_database, | ||
tables=self.tables, | ||
llm=llm, | ||
embed_model="local", | ||
text_to_sql_prompt=text_to_sql_template, | ||
streaming=True | ||
) | ||
|
||
response = query_engine.query(user_message) | ||
|
||
return response.response_gen | ||
|
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters