update: docuementation with installation instructions

README.md

@@ -1,3 +1,40 @@
 # MedRAG Multi-Modal
 
-Multi-modal RAG for medical docmain.
+Multi-modal RAG for medical docmain.
+
+## Installation
+
+### For Development
+
+For MacOS, you need to run
+
+```bash
+brew install poppler
+```
+
+For Debian/Ubuntu, you need to run
+
+```bash
+sudo apt-get install -y poppler-utils
+```
+
+Then, you can install the dependencies using uv in the virtual environment `.venv` using
+
+```bash
+git clone https://github.com/soumik12345/medrag-multi-modal
+cd medrag-multi-modal
+pip install -U pip uv
+uv sync
+```
+
+After this, you need to activate the virtual environment using
+
+```bash
+source .venv/bin/activate
+```
+
+In the activated virtual environment, you need to install Flash Attention using
+
+```bash
+uv pip install flash-attn --no-build-isolation
+```

docs/index.md

@@ -1,3 +1,40 @@
 # MedRAG Multi-Modal
 
-Multi-modal RAG for medical docmain.
+Multi-modal RAG for medical docmain.
+
+## Installation
+
+### For Development
+
+For MacOS, you need to run
+
+```bash
+brew install poppler
+```
+
+For Debian/Ubuntu, you need to run
+
+```bash
+sudo apt-get install -y poppler-utils
+```
+
+Then, you can install the dependencies using uv in the virtual environment `.venv` using
+
+```bash
+git clone https://github.com/soumik12345/medrag-multi-modal
+cd medrag-multi-modal
+pip install -U pip uv
+uv sync
+```
+
+After this, you need to activate the virtual environment using
+
+```bash
+source .venv/bin/activate
+```
+
+In the activated virtual environment, you need to install Flash Attention using
+
+```bash
+uv pip install flash-attn --no-build-isolation
+```

medrag_multi_modal/document_loader/load_image.py

@@ -13,8 +13,8 @@ from medrag_multi_modal.document_loader.load_text import TextLoader
 
 class ImageLoader(TextLoader):
     """
-    ImageLoader is a class that extends the `TextLoader` class to handle the extraction and
-    loading of images from a PDF file.
+    `ImageLoader` is a class that extends the `TextLoader` class to handle the extraction and
+    loading of pages from a PDF file as images.
 
     This class provides functionality to convert specific pages of a PDF document into images
     and optionally publish these images to a Weave dataset.

medrag_multi_modal/retrieval/__init__.py

@@ -0,0 +1,3 @@
+from .multi_modal_retrieval import MultiModalRetriever
+
+__all__ = ["MultiModalRetriever"]