Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+faiss_index_medical_MedEmbed/index.faiss filter=lfs diff=lfs merge=lfs -text
+faiss_index_medical_OpenAI/index.faiss filter=lfs diff=lfs merge=lfs -text
+medical_documents/14.Medicine[[:space:]](1).pdf filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,117 @@
+import os
+import streamlit as st
+from PIL import Image, ImageOps
+from langchain_openai import ChatOpenAI
+from langchain.embeddings import HuggingFaceEmbeddings, OpenAIEmbeddings
+from langchain.vectorstores import FAISS
+from langchain.chains import RetrievalQA
+from langchain import PromptTemplate
+from langchain.retrievers import ContextualCompressionRetriever
+from langchain.retrievers.document_compressors import FlashrankRerank
+from dotenv import load_dotenv
+from langchain_community.embeddings.bedrock import BedrockEmbeddings
+load_dotenv()
+# Hyperparameters
+PDF_CHUNK_SIZE = 1024
+PDF_CHUNK_OVERLAP = 256
+k = 3
+
+# Load favicon image
+def load_and_pad_image(image_path, size=(64, 64)):
+    img = Image.open(image_path)
+    return ImageOps.pad(img, size)
+
+favicon_path = "medical.png"
+favicon_image = load_and_pad_image(favicon_path)
+
+# Streamlit Page Config
+st.set_page_config(
+    page_title="Chatbot",
+    page_icon=favicon_image,
+)
+
+# Set up logo and title
+col1, col2 = st.columns([1, 8])
+with col1:
+    st.image(favicon_image)
+with col2:
+    st.markdown(
+        """
+        <h1 style='text-align: left; margin-top: -12px;'>Chatbot</h1>
+        """, unsafe_allow_html=True
+    )
+
+# Model and Embedding Selection
+model_options = ["gpt-4o", "gpt-4o-mini", "deepseek-chat"]
+selected_model = st.selectbox("Choose a GPT model", model_options)
+
+embedding_model_options = ["OpenAI", "Huggingface MedEmbed"]
+selected_embedding_model = st.selectbox("Choose an Embedding model", embedding_model_options)
+
+# Load the model
+def get_llm(selected_model):
+    api_key = os.getenv("DeepSeek_API_KEY") if selected_model == "deepseek-chat" else os.getenv("OPENAI_API_KEY")
+    return ChatOpenAI(
+        model=selected_model,
+        temperature=0,
+        max_tokens=None,
+        api_key=api_key,
+    )
+
+# Cache the vector store loading
+@st.cache_resource
+def load_vector_store(selected_embedding_model):
+    if selected_embedding_model == "OpenAI":
+        embeddings = OpenAIEmbeddings(model="text-embedding-3-large", api_key=os.getenv("OPENAI_API_KEY"))
+        return FAISS.load_local("faiss_index_medical_OpenAI", embeddings, allow_dangerous_deserialization=True)
+    else:
+        embeddings = HuggingFaceEmbeddings(model_name="abhinand/MedEmbed-large-v0.1")
+        return FAISS.load_local("faiss_index_medical_MedEmbed", embeddings, allow_dangerous_deserialization=True)
+
+# Load the selected vector store
+vector_store = load_vector_store(selected_embedding_model)
+llm = get_llm(selected_model)
+
+# Main App Logic
+def main():
+    st.session_state['knowledge_base'] = vector_store
+    st.header("Ask a Question")
+
+    question = st.text_input("Enter your question")
+    if st.button("Get Answer"):
+        knowledge_base = st.session_state['knowledge_base']
+        retriever = knowledge_base.as_retriever(search_kwargs={"k": k})
+        compressor = FlashrankRerank()
+        compression_retriever = ContextualCompressionRetriever(
+            base_compressor=compressor, base_retriever=retriever
+        )
+
+        system_prompt = """
+        You are a friendly and knowledgeable assistant who is an expert in medical education...
+        """
+
+        template = f"""
+        {system_prompt}
+        -------------------------------
+        Context: {{context}}
+        Question: {{question}}
+        Answer:
+        """
+
+        prompt = PromptTemplate(
+            template=template,
+            input_variables=['context', 'question']
+        )
+
+        qa_chain = RetrievalQA.from_chain_type(
+            llm,
+            retriever=compression_retriever,
+            return_source_documents=True,
+            chain_type_kwargs={"prompt": prompt}
+        )
+
+        response = qa_chain.invoke({"query": question})
+        st.write(f"**Answer:** {response['result']}")
+
+if __name__ == "__main__":
+    main()

faiss_index_medical_MedEmbed/index.faiss

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55c4ea45d2aa23d75f67e252d8f5e02e7b2b3c55324cc247e622752677b0ae68
+size 5873709

faiss_index_medical_MedEmbed/index.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be1ef50261c1eb4e4526269e9839f84000a4fb76640e063be2406979aef5d4b2
+size 2787575

faiss_index_medical_OpenAI/index.faiss

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a7de27af2193ddaa323c93cc83e4e5ec41d45776e1f9832954ab7bf92101532
+size 17621037

faiss_index_medical_OpenAI/index.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3899b2bffd0764a966d8f25aabccd32a99b3cacecfcfd2f69e3b4cf3487e6dd3
+size 2787575

medical_documents/14.Medicine (1).pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18cebaeb63c2b575edae04918575433efc9a9e3ed6d62c60a9164218a0d46d6e
+size 8776919

requirements.txt

@@ -0,0 +1,35 @@
+boto3
+awscli
+chromadb==0.4.14
+rank-bm25
+python-docx
+langchain
+langchain-community
+sentence-transformers
+pypdf
+rapidocr-onnxruntime
+pymupdf
+llama-index-core
+streamlit
+llama_index
+llama-index-llms-bedrock
+faiss-cpu
+langchain_openai
+python-dotenv
+transformers 
+sentence-transformers 
+unstructured 
+unstructured[pdf]
+pymupdf4llm
+requests
+beautifulsoup4
+selenium
+PyPDF2
+playwright
+#!playwright install
+nest_asyncio
+firecrawl
+langchain-cohere
+cohere-aws
+flashrank
+langchain-openai