Major feature addition 'llamaindex react agent for rag'

bootstraprag/templates/llamaindex/rag_with_react/main.py

@@ -0,0 +1,16 @@
+# driver code
+from react_agent_with_query_engine import ReActWithQueryEngine
+
+react_with_engine = ReActWithQueryEngine(input_dir='<YOUR_DATA_FOLDER>')
+
+# Start a loop to continually get input from the user
+while True:
+    # Get a query from the user
+    user_query = input("Enter your query [type 'bye' to 'exit']: ")
+
+    # Check if the user wants to terminate the loop
+    if user_query.lower() == "bye" or user_query.lower() == "exit":
+        break
+
+    response = react_with_engine.query(user_query=user_query)
+    print(response)

bootstraprag/templates/llamaindex/rag_with_react/react_agent_with_query_engine.py

@@ -0,0 +1,110 @@
+from llama_index.core import (
+    SimpleDirectoryReader,
+    VectorStoreIndex,
+    StorageContext,
+    load_index_from_storage,
+    Settings
+)
+from llama_index.core.tools import QueryEngineTool, ToolMetadata
+from llama_index.embeddings.ollama import OllamaEmbedding
+from llama_index.vector_stores.qdrant import QdrantVectorStore
+from llama_index.core.agent import ReActAgent
+from llama_index.llms.ollama import Ollama
+from llama_index.core.base.response.schema import Response, StreamingResponse, AsyncStreamingResponse, PydanticResponse
+from dotenv import load_dotenv, find_dotenv
+from typing import Union
+import qdrant_client
+import logging
+import os
+
+_ = load_dotenv(find_dotenv())
+
+logging.basicConfig(level=int(os.environ['INFO']))
+logger = logging.getLogger(__name__)
+
+
+class ReActWithQueryEngine:
+
+    RESPONSE_TYPE = Union[
+        Response, StreamingResponse, AsyncStreamingResponse, PydanticResponse
+    ]
+
+    def __init__(self, input_dir: str, similarity_top_k: int = 3):
+        self.index_loaded = False
+        self.similarity_top_k = similarity_top_k
+        self.input_dir = input_dir
+        self._index = None
+        self._engine = None
+        self.agent: ReActAgent = None
+        self.query_engine_tools = []
+
+        # use your prefered vector embeddings model
+        logger.info("initializing the OllamaEmbedding")
+        embed_model = OllamaEmbedding(model_name=os.environ['OLLMA_EMBED_MODEL'],
+                                      base_url=os.environ['OLLAMA_BASE_URL'])
+        # openai embeddings, embedding_model_name="text-embedding-3-large"
+        # embed_model = OpenAIEmbedding(embed_batch_size=10, model=embedding_model_name)
+
+        # use your prefered llm
+        llm = Ollama(model=os.environ['OLLMA_LLM_MODEL'], base_url=os.environ['OLLAMA_BASE_URL'], request_timeout=600)
+        # llm = OpenAI(model="gpt-4o")
+
+        logger.info("initializing the global settings")
+        Settings.embed_model = embed_model
+        Settings.llm = llm
+
+        # Create a local Qdrant vector store
+        logger.info("initializing the vector store related objects")
+        client = qdrant_client.QdrantClient(url=os.environ['DB_URL'], api_key=os.environ['DB_API_KEY'])
+        self.vector_store = QdrantVectorStore(client=client, collection_name=os.environ['COLLECTION_NAME'])
+        self._load_data_and_create_engine()
+
+    def _load_data_and_create_engine(self):
+        try:
+            self._index = VectorStoreIndex.from_vector_store(vector_store=self.vector_store)
+            self.index_loaded = True
+        except Exception as e:
+            self.index_loaded = False
+
+        if not self.index_loaded:
+            # load data
+            _docs = SimpleDirectoryReader(input_dir=self.input_dir).load_data()
+
+            # build and persist index
+            self._index = VectorStoreIndex.from_documents(documents=_docs)
+
+        self._engine = self._index.as_query_engine(similarity_top_k=self.similarity_top_k)
+        self._create_query_engine_tools()
+
+    def _create_query_engine_tools(self):
+        # can have more than one as per the requirement
+        self.query_engine_tools.append(
+            QueryEngineTool(
+                query_engine=self._engine,
+                metadata=ToolMetadata(
+                    name="test_tool_engine",  # change this accordingly
+                    description=(
+                        "Provides information about user query based on the information that you have. "
+                        "Use a detailed plain text question as input to the tool."
+                    ),
+                ),
+            )
+        )
+        self._create_react_agent()
+
+    def _create_react_agent(self):
+        # [Optional] Add Context
+        # context = """\
+        # You are a stock market sorcerer who is an expert on the companies Lyft and Uber.\
+        #     You will answer questions about Uber and Lyft as in the persona of a sorcerer \
+        #     and veteran stock market investor.
+        # """
+        self.agent = ReActAgent.from_tools(
+            self.query_engine_tools,
+            llm=Settings.llm,
+            verbose=True,
+            # context=context
+        )
+
+    def query(self, user_query: str) -> RESPONSE_TYPE:
+        return self.agent.query(str_or_query_bundle=user_query)

bootstraprag/templates/llamaindex/rag_with_react/readme.md

@@ -1 +1,8 @@
-## Under development, coming soon !
+## Instructions to run the code
+
+- Navigate to the root of the project and run the below command
+- `pip install -r requirements.txt`
+- In the data folder place your data preferably any ".pdf"
+- change the data directory in `main.py`
+#### Note: ensure your qdrant and ollama (if LLM models are pointing to local) are running
+- run `python main.py`

bootstraprag/templates/llamaindex/rag_with_react/requirements.txt

@@ -0,0 +1,9 @@
+qdrant-client==1.10.1
+python-dotenv==1.0.1
+llama-index==0.10.58
+llama-index-llms-openai==0.1.27
+llama-index-llms-ollama==0.2.0
+llama-index-embeddings-openai==0.1.11
+llama-index-embeddings-ollama==0.1.2
+llama-index-vector-stores-qdrant==0.2.14
+qdrant-client==1.10.1