indoxArcg 0.0.1

Indox Retrieval Augmentation

inDox

Official Website • Documentation • Discord

NEW: Subscribe to our mailing list for updates and news!

Indox Retrieval Augmentation is an innovative application designed to streamline information extraction from a wide range of document types, including text files, PDF, HTML, Markdown, and LaTeX. Whether structured or unstructured, Indox provides users with a powerful toolset to efficiently extract relevant data.

Indox Retrieval Augmentation is an innovative application designed to streamline information extraction from a wide range of document types, including text files, PDF, HTML, Markdown, and LaTeX. Whether structured or unstructured, Indox provides users with a powerful toolset to efficiently extract relevant data. One of its key features is the ability to intelligently cluster primary chunks to form more robust groupings, enhancing the quality and relevance of the extracted information. With a focus on adaptability and user-centric design, Indox aims to deliver future-ready functionality with more features planned for upcoming releases. Join us in exploring how Indox can revolutionize your document processing workflow, bringing clarity and organization to your data retrieval needs.

Roadmap

🤖 Model Support	Implemented	Description
Ollama (e.g. Llama3)	✅	Local Embedding and LLM Models powered by Ollama
HuggingFace	✅	Local Embedding and LLM Models powered by HuggingFace
Mistral	✅	Embedding and LLM Models by Cohere
Google (e.g. Gemini)	✅	Embedding and Generation Models by Google
OpenAI (e.g. GPT4)	✅	Embedding and Generation Models by OpenAI

| Supported Model Via Indox Api | Implemented | Description | | ----------------------------- | ----------- | ------------------------------------------------ | --- | | OpenAi | ✅ | Embedding and LLm OpenAi Model From Indox Api | | Mistral | ✅ | Embedding and LLm Mistral Model From Indox Api | | Anthropic | ❌ | Embedding and LLm Anthropic Model From Indox Api | |

📁 Loader and Splitter	Implemented	Description
Simple PDF	✅	Import PDF
UnstructuredIO	✅	Import Data through Unstructured
Clustered Load And Split	✅	Load pdf and texts. add a extra clustering layer

✨ RAG Features	Implemented	Description
Hybrid Search	❌	Semantic Search combined with Keyword Search
Semantic Caching	✅	Results saved and retrieved based on semantic meaning
Clustered Prompt	✅	Retrieve smaller chunks and do clustering and summarization
Agentic Rag	✅	Generate more reliabale answer, rank context and web search if needed
Advanced Querying	❌	Task Delegation Based on LLM Evaluation
Reranking	✅	Rerank results based on context for improved results
Customizable Metadata	❌	Free control over Metadata

🆒 Cool Bonus	Implemented	Description
Docker Support	❌	Indox is deployable via Docker
Customizable Frontend	❌	Indox's frontend is fully-customizable via the frontend

Examples

☑️ Examples	Run in Colab
Indox Api (OpenAi)
Mistral (Using Unstructured)
OpenAi (Using Clustered Split)
HuggingFace Models(Mistral)
Ollama
Evaluate with IndoxJudge

Indox Workflow

Getting Started

The following command will install the latest stable inDox

pip iIndoxQuery

Setting Up the Python Environment

If you are running this project in your local IDE, please create a Python environment to ensure all dependencies are correctly managed. You can follow the steps below to set up a virtual environment named indox:

Windows

1. Create the virtual environment:

  python -IndoxQuery

2. Activate the virtual environment:

``IndoxQuery\Scripts\activate

### macOS/Linux

1. **Create the virtual environment:**

   ```bash
   python3 -IndoxQuery

2. Activate the virtual environment:

  IndoxQuery/bin/activate

Install Dependencies

Once the virtual environment is activated, install the required dependencies by running:

  pip install -r requirements.txt

Preparing Your Data

1. Define the File Path: Specify the path to your text or PDF file.
2. Load LLM And Embedding Models: Initialize your embedding modeIndoxQuery's selection of pre-trained models.

Quick Start

Install the Required Packages

pip iIndoxQuery
pip install openai
pip install chromadb

Setting Up the Python Environment

If you are running this project in your local IDE, please create a Python environment to ensure all dependencies are correctly managed. You can follow the steps below to set up a virtual environment IndoxQuery`:

Windows

1. Create the virtual environment:

python -IndoxQuery

2. Activate the virtual environment:

`IndoxQuery_judge\Scripts\activate

### macOS/Linux

1. **Create the virtual environment:**
   ```bash
   python3 -IndoxQuery

2. **Activate the virtual environment:**
    ```bash
   IndoxQuery/bin/activate

Install Dependencies

Once the virtual environment is activated, install the required dependencies by running:

pip install -r requirements.txt

Load Environment Variables

To start, you need to load your API keys from the environment.

import os
from dotenv import load_dotenv

load_dotenv()

OPENAI_API_KEY = os.environ['OPENAI_API_KEY']

IndoxQuery Package

Import the necessary classes frIndoxQuery package.

Importing LLM and Embedding Models

InitIndoxQuery

Create an instance of IndoxRetrievalAugmentation.

openai_qa = OpenAiQA(api_key=OPENAI_API_KEY,model="gpt-3.5-turbo-0125")
openai_embeddings = OpenAiEmbedding(model="text-embedding-3-small",openai_api_key=OPENAI_API_KEY)

file_path = "sample.txt"

In this section, we take advantage of the unstructured library to load documents and split them into chunks by title. This method helps in organizing the document into manageable sections for further processing.

loader_splitter = UnstructuredLoadAndSplit(file_path=file_path)
docs = loader_splitter.load_and_chunk()

Starting processing...
End Chunking process.

Storing document chunks in a vector store is crucial for enabling efficient retrieval and search operations. By converting text data into vector representations and storing them in a vector store, you can perform rapid similarity searches and other vector-based operations.

db = ChromaVectorStore(collection_name="sample",embedding=embed_oIndoxQuery.connect_to_vectorstoIndoxQuery.store_in_vectorstore(docs)

2024-05-14 15:33:04,916 - INFO - Anonymized telemetry enabled. See                     https://docs.trychroma.com/telemetry for more information.
2024-05-14 15:33:12,587 - INFO - HTTP Request: POST https://api.openai.com/v1/embeddings "HTTP/1.1 200 OK"
2024-05-14 15:33:13,574 - INFO - Document added successfully to the vector store.

Connection established successfully.

<Indox.vectorstore.ChromaVectorStore at 0x28cf9369af0>

Quering

query = "how cinderella reach her happy ending?"

retriIndoxQuery.QuestionAnswer(vector_database=db,llm=openai_qa,top_k=5)
retriever.invoke(query)

2024-05-14 15:34:55,380 - INFO - HTTP Request: POST https://api.openai.com/v1/embeddings "HTTP/1.1 200 OK"
2024-05-14 15:35:01,917 - INFO - HTTP Request: POST https://api.openai.com/v1/chat/completions "HTTP/1.1 200 OK"
'Cinderella reached her happy ending by enduring mistreatment from her step-family, finding solace and help from the hazel tree and the little white bird, attending the royal festival where the prince recognized her as the true bride, and ultimately fitting into the golden shoe that proved her identity. This led to her marrying the prince and living happily ever after.'

retriever.context

["from the hazel-bush. Cinderella thanked him, went to her mother's\n\ngrave and planted the branch on it, and wept so much that the tears\n\nfell down on it and watered it. And it grew and became a handsome\n\ntree. Thrice a day cinderella went and sat beneath it, and wept and\n\nprayed, and a little white bird always came on the tree, and if\n\ncinderella expressed a wish, the bird threw down to her what she\n\nhad wished for.\n\nIt happened, however, that the king gave orders for a festival",
 'worked till she was weary she had no bed to go to, but had to sleep\n\nby the hearth in the cinders. And as on that account she always\n\nlooked dusty and dirty, they called her cinderella.\n\nIt happened that the father was once going to the fair, and he\n\nasked his two step-daughters what he should bring back for them.\n\nBeautiful dresses, said one, pearls and jewels, said the second.\n\nAnd you, cinderella, said he, what will you have. Father',
 'face he recognized the beautiful maiden who had danced with\n\nhim and cried, that is the true bride. The step-mother and\n\nthe two sisters were horrified and became pale with rage, he,\n\nhowever, took cinderella on his horse and rode away with her. As\n\nthey passed by the hazel-tree, the two white doves cried -\n\nturn and peep, turn and peep,\n\nno blood is in the shoe,\n\nthe shoe is not too small for her,\n\nthe true bride rides with you,\n\nand when they had cried that, the two came flying down and',
 "to send her up to him, but the mother answered, oh, no, she is\n\nmuch too dirty, she cannot show herself. But he absolutely\n\ninsisted on it, and cinderella had to be called. She first\n\nwashed her hands and face clean, and then went and bowed down\n\nbefore the king's son, who gave her the golden shoe. Then she\n\nseated herself on a stool, drew her foot out of the heavy\n\nwooden shoe, and put it into the slipper, which fitted like a\n\nglove. And when she rose up and the king's son looked at her",
 'slippers embroidered with silk and silver. She put on the dress\n\nwith all speed, and went to the wedding. Her step-sisters and the\n\nstep-mother however did not know her, and thought she must be a\n\nforeign princess, for she looked so beautiful in the golden dress.\n\nThey never once thought of cinderella, and believed that she was\n\nsitting at home in the dirt, picking lentils out of the ashes. The\n\nprince approached her, took her by the hand and danced with her.']