Modernized LlamaIndex integration (#1613)

Updated LlamaIndex example

Author: jamesbraza

examples/llamaindex/README.md

@@ -1,30 +1,27 @@
 # LocalAI Demonstration with Embeddings
 
-This demonstration shows you how to use embeddings with existing data in LocalAI. We are using the `llama_index` library to facilitate the embedding and querying processes. The `Weaviate` client is used as the embedding source.
-
-## Prerequisites
-
-Before proceeding, make sure you have the following installed:
-- Weaviate client
-- LocalAI and its dependencies
-- llama_index and its dependencies 
+This demonstration shows you how to use embeddings with existing data in LocalAI.
+We are using the `llama-index` library to facilitate the embedding and querying processes.
+The `Weaviate` client is used as the embedding source.
 
 ## Getting Started
 
-1. Clone this repository:
-
-2. Navigate to the project directory:
+1. Clone this repository and navigate to this directory
 
-3. Run the example:
+    ```bash
+    git clone git@github.com:mudler/LocalAI.git
+    cd LocalAI/examples/llamaindex
+    ```
 
-`python main.py`
+2. pip install LlamaIndex and Weviate's client: `pip install llama-index>=0.9.9 weviate-client`
+3. Run the example: `python main.py`
 
-```
+```none
 Downloading (…)lve/main/config.json: 100%|███████████████████████████| 684/684 [00:00<00:00, 6.01MB/s]
 Downloading model.safetensors: 100%|███████████████████████████████| 133M/133M [00:03<00:00, 39.5MB/s]
 Downloading (…)okenizer_config.json: 100%|███████████████████████████| 366/366 [00:00<00:00, 2.79MB/s]
 Downloading (…)solve/main/vocab.txt: 100%|█████████████████████████| 232k/232k [00:00<00:00, 6.00MB/s]
 Downloading (…)/main/tokenizer.json: 100%|█████████████████████████| 711k/711k [00:00<00:00, 18.8MB/s]
 Downloading (…)cial_tokens_map.json: 100%|███████████████████████████| 125/125 [00:00<00:00, 1.18MB/s]
 LocalAI is a community-driven project that aims to make AI accessible to everyone. It was created by Ettore Di Giacinto and is focused on providing various AI-related features such as text generation with GPTs, text to audio, audio to text, image generation, and more. The project is constantly growing and evolving, with a roadmap for future improvements. Anyone is welcome to contribute, provide feedback, and submit pull requests to help make LocalAI better.
-```
+```

examples/llamaindex/main.py

@@ -1,38 +1,29 @@
-import os
-
 import weaviate
-
-from llama_index import ServiceContext, VectorStoreIndex, StorageContext
-from llama_index.llms import LocalAI
+from llama_index import ServiceContext, VectorStoreIndex
+from llama_index.llms import LOCALAI_DEFAULTS, OpenAILike
 from llama_index.vector_stores import WeaviateVectorStore
-from llama_index.storage.storage_context import StorageContext
-
-# Weaviate client setup
-client = weaviate.Client("http://weviate.default")
 
 # Weaviate vector store setup
-vector_store = WeaviateVectorStore(weaviate_client=client, index_name="AIChroma")
-
-# Storage context setup
-storage_context = StorageContext.from_defaults(vector_store=vector_store)
+vector_store = WeaviateVectorStore(
+    weaviate_client=weaviate.Client("http://weviate.default"), index_name="AIChroma"
+)
 
-# LocalAI setup
-llm = LocalAI(temperature=0, model_name="gpt-3.5-turbo", api_base="http://local-ai.default", api_key="stub")
-llm.globally_use_chat_completions = True;
+# LLM setup, served via LocalAI
+llm = OpenAILike(temperature=0, model="gpt-3.5-turbo", **LOCALAI_DEFAULTS)
 
 # Service context setup
 service_context = ServiceContext.from_defaults(llm=llm, embed_model="local")
 
 # Load index from stored vectors
 index = VectorStoreIndex.from_vector_store(
-    vector_store,
-    storage_context=storage_context,
-    service_context=service_context
+    vector_store, service_context=service_context
 )
 
 # Query engine setup
-query_engine = index.as_query_engine(similarity_top_k=1, vector_store_query_mode="hybrid")
+query_engine = index.as_query_engine(
+    similarity_top_k=1, vector_store_query_mode="hybrid"
+)
 
 # Query example
 response = query_engine.query("What is LocalAI?")
-print(response)
+print(response)