updated new version deepseek-r1

.gitignore

@@ -0,0 +1,174 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc

Dockerfile

@@ -0,0 +1,23 @@
+# Use an official Python image
+FROM python:3.10
+
+# Install dependencies
+RUN apt-get update && apt-get install -y curl
+
+# Install Ollama
+RUN curl -fsSL https://ollama.com/install.sh | sh
+
+# Install Python dependencies
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy app files
+COPY . .
+
+# Expose Ollama API port
+EXPOSE 11434
+
+# Start Ollama in the background
+RUN ollama pull deepseek-r1:1.5b
+
+CMD ollama serve & streamlit run app.py --server.port 7860 --server.enableCORS false --server.enableXsrfProtection false

app.py

@@ -1,146 +1,35 @@
 import streamlit as st
 import os
-from langchain_community.document_loaders import PDFMinerLoader
-from langchain_community.embeddings import SentenceTransformerEmbeddings
-from langchain.text_splitter import RecursiveCharacterTextSplitter
-from langchain_community.vectorstores import FAISS
-from langchain.chains import RetrievalQA
-from langchain_community.llms import HuggingFacePipeline
-from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, pipeline
-import torch
+from components.sidebar import render_sidebar
+from components.chat_ui import display_chat
+from services.llm import initialize_qa_chain, initialize_chain
+from utils.helpers import get_file_size
 
-st.title("DocChatAI | Chat over PDF Doc")
+# import subprocess
+# process = subprocess.Popen("ollama serve", shell=True) 
+# print(process)
 
-# Custom CSS for chat messages
-st.markdown("""
-    <style>
-        .user-message {
-            text-align: right;
-            background-color: #3c8ce7;
-            color: white;
-            padding: 10px;
-            border-radius: 10px;
-            margin-bottom: 10px;
-            display: inline-block;
-            width: fit-content;
-            max-width: 70%;
-            margin-left: auto;
-            box-shadow: 0px 4px 6px rgba(0, 0, 0, 0.1);
-        }
-        .assistant-message {
-            text-align: left;
-            background-color: #d16ba5;
-            color: white;
-            padding: 10px;
-            border-radius: 10px;
-            margin-bottom: 10px;
-            display: inline-block;
-            width: fit-content;
-            max-width: 70%;
-            margin-right: auto;
-            box-shadow: 0px 4px 6px rgba(0, 0, 0, 0.1);
-        }
-    </style>
-""", unsafe_allow_html=True)
+# App Title
+st.title("DocChatAI | Chat Using Documents")
 
-def get_file_size(file):
-    file.seek(0, os.SEEK_END)
-    file_size = file.tell()
-    file.seek(0)
-    return file_size
-
-# Add a sidebar for model selection and user details
-st.sidebar.write("Settings")
-st.sidebar.write("-----------")
-model_options = ["MBZUAI/LaMini-T5-738M", "google/flan-t5-base", "google/flan-t5-small"]
-selected_model = st.sidebar.radio("Choose Model", model_options)
-st.sidebar.write("-----------")
-uploaded_file = st.sidebar.file_uploader("Upload file", type=["pdf"])
-st.sidebar.write("-----------")
-st.sidebar.write("About Me")
-st.sidebar.write("Name: Deepak Yadav")
-st.sidebar.write("Bio: Passionate about AI and machine learning. Enjoys working on innovative projects and sharing knowledge with the community.")
-st.sidebar.write("[GitHub](https://github.com/deepak7376)")
-st.sidebar.write("[LinkedIn](https://www.linkedin.com/in/dky7376/)")
-st.sidebar.write("-----------")
-
-@st.cache_resource
-def initialize_qa_chain(filepath, CHECKPOINT):
-    loader = PDFMinerLoader(filepath)
-    documents = loader.load()
-    text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=500)
-    splits = text_splitter.split_documents(documents)
-
-    # Create embeddings 
-    embeddings = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
-    vectordb = FAISS.from_documents(splits, embeddings)
-
-    # Initialize model
-    TOKENIZER = AutoTokenizer.from_pretrained(CHECKPOINT)
-    BASE_MODEL = AutoModelForSeq2SeqLM.from_pretrained(CHECKPOINT, device_map=torch.device('cpu'), torch_dtype=torch.float32)
-    pipe = pipeline(
-        'text2text-generation',
-        model=BASE_MODEL,
-        tokenizer=TOKENIZER,
-        max_length=256,
-        do_sample=True,
-        temperature=0.3,
-        top_p=0.95,
-    )
-
-    llm = HuggingFacePipeline(pipeline=pipe)
-
-    # Build a QA chain
-    qa_chain = RetrievalQA.from_chain_type(
-        llm=llm,
-        chain_type="stuff",
-        retriever=vectordb.as_retriever(),
-    )
-    return qa_chain
-
-def process_answer(instruction, qa_chain):
-    generated_text = qa_chain.run(instruction)
-    return generated_text
+# Sidebar - Model Selection & File Upload
+selected_model, temperature, top_p, max_tokens, uploaded_file = render_sidebar()
 
+mode = False
+# Check if a PDF file is uploaded
 if uploaded_file is not None:
     os.makedirs("docs", exist_ok=True)
     filepath = os.path.join("docs", uploaded_file.name)
+
     with open(filepath, "wb") as temp_file:
         temp_file.write(uploaded_file.read())
-        temp_filepath = temp_file.name
 
-    with st.spinner('Embeddings are in process...'):
-        qa_chain = initialize_qa_chain(temp_filepath, selected_model)
+    with st.spinner('Please wait...'):
+        qa_chain = initialize_qa_chain(filepath, selected_model, temperature, top_p, max_tokens)
+        mode = True
 else:
-    qa_chain = None
-
-# Initialize chat history
-if "messages" not in st.session_state:
-    st.session_state.messages = []
-
-# Display chat messages from history on app rerun
-for message in st.session_state.messages:
-    if message["role"] == "user":
-        st.markdown(f"<div class='user-message'>{message['content']}</div>", unsafe_allow_html=True)
-    else:
-        st.markdown(f"<div class='assistant-message'>{message['content']}</div>", unsafe_allow_html=True)
-
-# React to user input
-if prompt := st.chat_input("What is up?"):
-    # Display user message in chat message container
-    st.markdown(f"<div class='user-message'>{prompt}</div>", unsafe_allow_html=True)
-    # Add user message to chat history
-    st.session_state.messages.append({"role": "user", "content": prompt})
 
-    if qa_chain:
-        # Generate response
-        response = process_answer({'query': prompt}, qa_chain)
-    else:
-        # Prompt to upload a file
-        response = "Please upload a PDF file to enable the chatbot."
+    qa_chain = initialize_chain(selected_model, temperature, top_p, max_tokens)
 
-    # Display assistant response in chat message container
-    st.markdown(f"<div class='assistant-message'>{response}</div>", unsafe_allow_html=True)
-    
-    # Add assistant response to chat history
-    st.session_state.messages.append({"role": "assistant", "content": response})
+# Initialize and Display Chat History
+display_chat(qa_chain, mode)

components/chat_ui.py

@@ -0,0 +1,89 @@
+import streamlit as st
+# from services.llm import process_answer
+import time
+import re
+
+# Custom CSS for chat styling
+CHAT_CSS = """
+    <style>
+        .user-message {
+            text-align: right;
+            background-color: #3c8ce7;
+            color: white;
+            padding: 10px;
+            border-radius: 10px;
+            margin-bottom: 10px;
+            display: inline-block;
+            width: fit-content;
+            max-width: 70%;
+            margin-left: auto;
+            box-shadow: 0px 4px 6px rgba(0, 0, 0, 0.1);
+        }
+        .assistant-message {
+            text-align: left;
+            background-color: #d16ba5;
+            color: white;
+            padding: 10px;
+            border-radius: 10px;
+            margin-bottom: 10px;
+            display: inline-block;
+            width: fit-content;
+            max-width: 70%;
+            margin-right: auto;
+            box-shadow: 0px 4px 6px rgba(0, 0, 0, 0.1);
+        }
+    </style>
+"""
+def extract_thoughts(response_text):
+    """Extracts <think>...</think> content and the main answer."""
+    match = re.search(r"<think>(.*?)</think>", response_text, re.DOTALL)
+    if match:
+        thinking_part = match.group(1).strip()
+        main_answer = re.sub(r"<think>.*?</think>", "", response_text, flags=re.DOTALL).strip()
+    else:
+        thinking_part = None
+        main_answer = response_text.strip()
+    
+    return thinking_part, main_answer
+
+# Streamed response emulator
+def response_generator(response):
+    for word in response.split():
+        yield word + " "
+        time.sleep(0.05)
+
+def display_chat(qa_chain, mode):
+    st.markdown(CHAT_CSS, unsafe_allow_html=True)
+
+    if "messages" not in st.session_state:
+        st.session_state.messages = []
+
+    for message in st.session_state.messages:
+        with st.chat_message(message["role"]):
+            st.markdown(message["content"])
+
+    if prompt := st.chat_input("Ask something..."):
+        st.session_state.messages.append({"role": "user", "content": prompt})
+        with st.chat_message("user"):
+            st.markdown(prompt)
+        
+        # Get chat response
+        response = qa_chain.invoke({"input": prompt}) if mode else qa_chain.invoke({'context': prompt})
+        if not response:  # Handle empty responses
+            response = {'answer': "I don't know."}
+
+        if mode is False:
+            response = {'answer': response}
+
+        # Extract <think> part and main answer
+        thinking_part, main_answer = extract_thoughts(response['answer'])
+
+        # Display assistant response
+        with st.chat_message("assistant"):
+            if thinking_part:
+                with st.expander("💭 Thought Process"):
+                    st.markdown(thinking_part)  # Hidden by default, expandable
+                
+            response = st.write_stream(response_generator(main_answer))
+        
+        st.session_state.messages.append({"role": "assistant", "content": response})

components/sidebar.py

@@ -0,0 +1,35 @@
+import streamlit as st
+
+def render_sidebar():
+    st.sidebar.title("DocChatAI")
+    st.sidebar.subheader("Chat using PDF Document")
+    st.sidebar.write("-----------")
+    
+    # Model Selection
+    model_options = ["deepseek-r1:1.5b"]
+    selected_model = st.sidebar.radio("Choose Model", model_options)
+    
+    st.sidebar.write("-----------")
+    
+    # Hyperparameters
+    temperature = st.sidebar.slider("Temperature", min_value=0.0, max_value=1.0, value=0.7, step=0.1)
+    top_p = st.sidebar.slider("Top-p (Nucleus Sampling)", min_value=0.0, max_value=1.0, value=0.9, step=0.05)
+    max_tokens = st.sidebar.number_input("Max Tokens", min_value=10, max_value=2048, value=256, step=10)
+    
+    st.sidebar.write("-----------")
+    
+    # File Upload
+    uploaded_file = st.sidebar.file_uploader("Upload Documents", type=["pdf"])
+    
+    st.sidebar.write("-----------")
+    
+    # About Section
+    st.sidebar.write("About Me")
+    st.sidebar.write("Name: Deepak Yadav")
+    st.sidebar.write("Bio: Passionate about AI and machine learning.")
+    st.sidebar.write("[GitHub](https://github.com/deepak7376)")
+    st.sidebar.write("[LinkedIn](https://www.linkedin.com/in/dky7376/)")
+    
+    st.sidebar.write("-----------")
+    
+    return selected_model, temperature, top_p, max_tokens, uploaded_file

doc.txt

@@ -0,0 +1,14 @@
+rag_chatbot/
+│── app.py                    # Main Streamlit app
+│── components/
+│   ├── sidebar.py            # Sidebar UI (model selection, upload, user info)
+│   ├── chat_ui.py            # Chat UI styling and history
+│── services/
+│   ├── pdf_processing.py      # PDF loading and text splitting
+│   ├── vector_store.py        # Vector database (FAISS) setup
+│   ├── llm.py                 # Model initialization and QA chain
+│── utils/
+│   ├── helpers.py             # Helper functions (e.g., get_file_size)
+│── docs/                      # Folder for storing uploaded PDFs (created dynamically)
+│── requirements.txt           # Dependencies
+│── README.md                  # Project documentation

docs/dummy.txt

@@ -0,0 +1 @@
+hello

requirements.txt

@@ -5,4 +5,7 @@ langchain
 langchain-community
 accelerate 
 sentence-transformers 
-faiss-cpu
+faiss-cpu
+pymupdf
+ollama
+langchain_ollama

services/llm.py

@@ -0,0 +1,86 @@
+import ollama
+from langchain.chains import RetrievalQA
+from langchain.chains import create_retrieval_chain
+from langchain_ollama import OllamaLLM
+from services.pdf_processing import load_and_split_pdf
+from services.vector_store import create_vector_store
+from langchain.chains.combine_documents import create_stuff_documents_chain
+from langchain_core.prompts import ChatPromptTemplate
+from langchain.prompts import PromptTemplate
+import streamlit as st
+
+PROMPT_TEMPLATE = """Question: {context}
+
+Answer: Let's think step by step."""
+
+@st.cache_resource
+def initialize_qa_chain(filepath, model_name, temperature, top_p, max_tokens):
+    # Load and split the PDF
+    splits = load_and_split_pdf(filepath)
+    vectordb = create_vector_store(splits)
+
+    # Use Ollama or Hugging Face LLM
+    # Configure the LLM with additional parameters
+    llm = OllamaLLM(
+        model=model_name,
+        temperature=temperature,   # Controls randomness (0 = deterministic, 1 = max randomness)
+        max_tokens=max_tokens,   # Limit the number of tokens in the output
+        top_p=top_p          # Nucleus sampling for controlling diversity
+    )
+
+
+    # # Define strict retrieval-based prompting
+    # prompt_template = PromptTemplate(
+    #     template=(
+    #         "You are an AI assistant that only answers questions based on the provided document. "
+    #         "Do not use external knowledge. If you cannot find an answer in the document, respond with: 'I don't know.'\n\n"
+    #         "Document Context:\n{context}\n\n"
+    #         "User Question: {query}\n\n"
+    #         "Assistant Answer:"
+    #     ),
+    #     input_variables=["context", "query"]
+    # )
+
+    system_prompt = (
+    "Use the given context to answer the question. "
+    "If you don't know the answer, say you don't know. "
+    "Use three sentence maximum and keep the answer concise. "
+    "Context: {context}"
+)
+    prompt = ChatPromptTemplate.from_messages(
+        [
+            ("system", system_prompt),
+            ("human", "{input}"),
+        ]
+    )
+    question_answer_chain = create_stuff_documents_chain(llm, prompt)
+    chain = create_retrieval_chain(vectordb.as_retriever(), question_answer_chain)
+
+    # return RetrievalQA.from_chain_type(
+    #     llm=llm,
+    #     chain_type="stuff",
+    #     retriever=vectordb.as_retriever(),
+    #     chain_type_kwargs={"prompt": prompt_template}
+    # )
+    return chain
+
+@st.cache_resource
+def initialize_chain(model_name, temperature, top_p, max_tokens):
+    # Use Ollama or Hugging Face LLM
+    # Configure the LLM with additional parameters
+    llm = OllamaLLM(
+        model=model_name,
+        temperature=temperature,   # Controls randomness (0 = deterministic, 1 = max randomness)
+        max_tokens=max_tokens,   # Limit the number of tokens in the output
+        top_p=top_p          # Nucleus sampling for controlling diversity
+    )
+
+
+
+    prompt = ChatPromptTemplate.from_template(PROMPT_TEMPLATE)
+
+    chain = prompt | llm
+
+    return chain
+
+

services/pdf_processing.py

@@ -0,0 +1,8 @@
+from langchain_community.document_loaders import PyMuPDFLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+def load_and_split_pdf(filepath):
+    loader = PyMuPDFLoader(filepath)  # Use PyMuPDFLoader instead
+    documents = loader.load()
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=100)
+    return text_splitter.split_documents(documents)

services/vector_store.py

@@ -0,0 +1,10 @@
+from langchain_community.vectorstores import FAISS
+# from langchain_community.embeddings import SentenceTransformerEmbeddings
+# from langchain_community.embeddings.ollama import OllamaEmbeddings
+from langchain_huggingface import HuggingFaceEmbeddings
+
+def create_vector_store(splits):
+    # embeddings = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
+    embeddings = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
+    # embeddings = OllamaEmbeddings(model="nomic-embed-text")
+    return FAISS.from_documents(splits, embeddings)

utils/helpers.py

@@ -0,0 +1,7 @@
+import os
+
+def get_file_size(file):
+    file.seek(0, os.SEEK_END)
+    size = file.tell()
+    file.seek(0)
+    return size