initial

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz 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

.gitignore

@@ -0,0 +1,13 @@
+/libpff-20231205
+/libpff-alpha-20231205.tar.gz
+/venv
+/.chainlit
+/model
+/__pycache__
+.files
+*.gguf
+*.pst
+/data
+*.jpg
+/*.env
+/*.log

README.md

@@ -0,0 +1,13 @@
+---
+title: RAG Retrieve Ingest Cz Eng
+emoji: 😻
+colorFrom: purple
+colorTo: yellow
+sdk: gradio
+sdk_version: 4.19.1
+app_file: app.py
+pinned: false
+license: mit
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,4 @@
+import subprocess
+
+print("Starting the FastAPI server...")
+subprocess.run("uvicorn fast_app:app --host 0.0.0.0 --port 7860", shell=True)

fast_app.py

@@ -0,0 +1,146 @@
+from dotenv import load_dotenv
+import os
+import json
+from fastapi import FastAPI, Request, Form, Response
+from fastapi.responses import HTMLResponse
+from fastapi.templating import Jinja2Templates
+from fastapi.staticfiles import StaticFiles
+from fastapi.encoders import jsonable_encoder
+
+from langchain.vectorstores import Chroma
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+from langchain.chains import RetrievalQA
+from langchain.document_loaders import (
+    TextLoader,
+    PyPDFLoader,
+    DirectoryLoader,
+    UnstructuredFileLoader,
+)
+from langchain.document_loaders.csv_loader import CSVLoader
+from langchain.llms import OpenAI
+from langchain import PromptTemplate
+from langchain.embeddings import OpenAIEmbeddings, HuggingFaceEmbeddings
+
+from ingest import Ingest
+
+# setx OPENAI_API_KEY "your_openai_api_key_here"
+
+# Access the Hugging Face API token from an environment variable
+# huggingface_token = os.getenv("HUGGINGFACE_TOKEN")
+# if huggingface_token is None:
+#    raise ValueError("Hugging Face token is not set in environment variables.")
+
+openai_api_key = os.getenv("OPENAI_API_KEY")
+if openai_api_key is None:
+    raise ValueError("OAI token is not set in environment variables.")
+
+
+app = FastAPI()
+templates = Jinja2Templates(directory="templates")
+app.mount("/static", StaticFiles(directory="static"), name="static")
+
+czech_store = "stores/czech_512"
+english_store = "stores/english_256"
+
+ingestor = Ingest(
+    openai_api_key=openai_api_key,
+    chunk=256,
+    overlap=128,
+    czech_store=czech_store,
+    english_store=english_store,
+)
+
+load_dotenv()
+
+prompt_template = """You are a electrical engineer focused on lighting and chandeliers. Provide helpful answer to the user question.
+If you don't know the answer, just say that you don't know, don't try to make up an answer.
+
+Context: {context}
+Question: {question}
+
+Only return the helpful answer below and nothing else.
+Helpful answer:
+"""
+
+prompt = PromptTemplate(
+    template=prompt_template, input_variables=["context", "question"]
+)
+
+print("\n Prompt ready... \n\n")
+
+
+@app.get("/", response_class=HTMLResponse)
+def read_item(request: Request):
+    return templates.TemplateResponse("index.html", {"request": request})
+
+
+@app.post("/ingest_data")
+async def ingest_data(folderPath: str = Form(...), language: str = Form(...)):
+    # Determine the correct data path and store based on the language
+    if language == "czech":
+        print("\n Czech language selected....\n\n")
+        ingestor.data_czech = folderPath
+        ingestor.ingest_czech()
+        message = "Czech data ingestion complete."
+    else:
+        print("\n English language selected....\n\n")
+        ingestor.data_english = folderPath
+        ingestor.ingest_english()
+        message = "English data ingestion complete."
+
+    return {"message": message}
+
+
+@app.post("/get_response")
+async def get_response(query: str = Form(...), language: str = Form(...)):
+    print(language)
+    if language == "czech":
+        print("\n Czech language selected....\n\n")
+        embedding_model = "Seznam/simcse-dist-mpnet-paracrawl-cs-en"
+        persist_directory = czech_store
+        model_name = embedding_model
+        model_kwargs = {"device": "cpu"}
+        encode_kwargs = {"normalize_embeddings": False}
+        embedding = HuggingFaceEmbeddings(
+            model_name=model_name,
+            model_kwargs=model_kwargs,
+            encode_kwargs=encode_kwargs,
+        )
+    else:
+        print("\n English language selected....\n\n")
+        embedding_model = "text-embedding-3-large"  # Default to English
+        persist_directory = english_store
+        embedding = OpenAIEmbeddings(
+            openai_api_key=openai_api_key,
+            model=embedding_model,
+        )
+
+    vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding)
+    retriever = vectordb.as_retriever(search_kwargs={"k": 10})
+
+    chain_type_kwargs = {"prompt": prompt}
+    qa_chain = RetrievalQA.from_chain_type(
+        llm=OpenAI(openai_api_key=openai_api_key),
+        chain_type="stuff",
+        retriever=retriever,
+        return_source_documents=True,
+        chain_type_kwargs=chain_type_kwargs,
+        verbose=True,
+    )
+    response = qa_chain(query)
+
+    for i in response["source_documents"]:
+        print(f"\n{i}\n\n")
+
+    print(response)
+
+    answer = response["result"]
+    source_document = response["source_documents"][0].page_content
+    doc = response["source_documents"][0].metadata["source"]
+    response_data = jsonable_encoder(
+        json.dumps({"answer": answer, "source_document": source_document, "doc": doc})
+    )
+
+    res = Response(response_data)
+    return res

fast_app_cz(obsolete).py

@@ -0,0 +1,110 @@
+from dotenv import load_dotenv
+import os
+import json
+from fastapi import FastAPI, Request, Form, Response
+from fastapi.responses import HTMLResponse
+from fastapi.templating import Jinja2Templates
+from fastapi.staticfiles import StaticFiles
+from fastapi.encoders import jsonable_encoder
+from langchain.llms import CTransformers
+
+from langchain.vectorstores import Chroma
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+from langchain.chains import RetrievalQA
+from langchain.document_loaders import TextLoader, PyPDFLoader, DirectoryLoader
+from langchain.llms import OpenAI
+from langchain import PromptTemplate
+from langchain.embeddings import OpenAIEmbeddings, HuggingFaceEmbeddings
+
+app = FastAPI()
+load_dotenv()  
+openai_api_key = os.environ.get("OPENAI_API_KEY")
+templates = Jinja2Templates(directory="templates")
+app.mount("/static", StaticFiles(directory="static"), name="static")
+# embedding_model = "Seznam/simcse-dist-mpnet-czeng-cs-en"
+embedding_model = "Seznam/simcse-dist-mpnet-paracrawl-cs-en"
+persist_directory = "stores/seznampara_ul_512"
+
+llm = OpenAI(openai_api_key=openai_api_key)
+# llm = "model\dolphin-2.6-mistral-7b.Q4_K_S.gguf"
+# llm = "neural-chat-7b-v3-1.Q4_K_M.gguf"
+
+
+"""
+### - Local LLM settings - ###
+
+config = {
+    "max_new_tokens": 1024,
+    "repetition_penalty": 1.1,
+    "temperature": 0.1,
+    "top_k": 50,
+    "top_p": 0.9,
+    "stream": True,
+    "threads": int(os.cpu_count() / 2),
+}
+
+llm = CTransformers(
+    model=llm, model_type="mistral", lib="avx2", **config  # for CPU use
+)
+
+### - Local LLM settings end - ###
+"""
+
+prompt_template = """Use the following pieces of information to answer the user's question.
+If you don't know the answer, just say that you don't know, don't try to make up an answer.
+
+Context: {context}
+Question: {question}
+
+Only return the helpful answer below and nothing else.
+Helpful answer:
+"""
+
+prompt = PromptTemplate(
+    template=prompt_template, input_variables=["context", "question"]
+)
+
+print("\n Prompt ready... \n\n")
+
+
+model_name = embedding_model
+model_kwargs = {"device": "cpu"}
+encode_kwargs = {"normalize_embeddings": False}
+embedding = HuggingFaceEmbeddings(
+    model_name=model_name, model_kwargs=model_kwargs, encode_kwargs=encode_kwargs
+)
+vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding)
+retriever = vectordb.as_retriever(search_kwargs={"k": 3})
+
+print("\n Retrieval Ready....\n\n")
+
+
+@app.get("/", response_class=HTMLResponse)
+def read_item(request: Request):
+    return templates.TemplateResponse("index.html", {"request": request})
+
+
+@app.post("/get_response")
+async def get_response(query: str = Form(...)):
+
+    chain_type_kwargs = {"prompt": prompt}
+    qa_chain = RetrievalQA.from_chain_type(
+        llm=llm,
+        chain_type="stuff",
+        retriever=retriever,
+        return_source_documents=True,
+        chain_type_kwargs=chain_type_kwargs,
+        verbose=True,
+    )
+    response = qa_chain(query)
+    print(response)
+    answer = response["result"]
+    source_document = response["source_documents"][0].page_content
+    doc = response["source_documents"][0].metadata["source"]
+    response_data = jsonable_encoder(
+        json.dumps({"answer": answer, "source_document": source_document, "doc": doc})
+    )
+
+    res = Response(response_data)
+    return res

ingest(obsolete).py

@@ -0,0 +1,59 @@
+from langchain.vectorstores import Chroma
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+from langchain.document_loaders import (
+    PyPDFLoader,
+    DirectoryLoader,
+    UnstructuredFileLoader,
+)
+from langchain.document_loaders.csv_loader import CSVLoader
+from langchain.embeddings import (
+    OpenAIEmbeddings,
+    HuggingFaceBgeEmbeddings,
+    HuggingFaceEmbeddings,
+    HuggingFaceInstructEmbeddings,
+)
+
+
+persist_directory = "stores/test_512"
+data = "data\czech"
+chunk = 512
+overlap = 128
+# embedding_model = "Seznam/simcse-dist-mpnet-czeng-cs-en"
+embedding_model = "Seznam/simcse-dist-mpnet-paracrawl-cs-en"
+
+model_name = embedding_model
+model_kwargs = {"device": "cpu"}
+encode_kwargs = {"normalize_embeddings": False}
+embedding = HuggingFaceEmbeddings(
+    model_name=model_name, model_kwargs=model_kwargs, encode_kwargs=encode_kwargs
+)
+
+"""
+loader = CSVLoader(
+    file_path="data/emails.csv",
+    encoding="utf-8",
+    csv_args={
+        "delimiter": ";",
+    },
+)
+
+"""
+
+loader = DirectoryLoader(data, show_progress=True)
+
+
+documents = loader.load()
+text_splitter = RecursiveCharacterTextSplitter(
+    chunk_size=chunk,
+    chunk_overlap=overlap,
+)
+texts = text_splitter.split_documents(documents)
+vectordb = Chroma.from_documents(
+    documents=texts,
+    embedding=embedding,
+    persist_directory=persist_directory,
+    collection_metadata={"hnsw:space": "cosine"},
+)
+
+print("\n Vector Store Created.......\n\n")

ingest.py

@@ -0,0 +1,127 @@
+from langchain.vectorstores import Chroma
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+from langchain.document_loaders import PyPDFLoader, DirectoryLoader
+from langchain.embeddings import (
+    OpenAIEmbeddings,
+    HuggingFaceBgeEmbeddings,
+    HuggingFaceEmbeddings,
+    HuggingFaceInstructEmbeddings,
+)
+
+
+class Ingest:
+    def __init__(
+        self,
+        openai_api_key=None,
+        chunk=512,
+        overlap=256,
+        czech_store="stores/czech_512",
+        english_store="stores/english_512",
+        data_czech="data/czech",
+        data_english="data/english",
+    ):
+        self.openai_api_key = openai_api_key
+        self.chunk = chunk
+        self.overlap = overlap
+        self.czech_store = czech_store
+        self.english_store = english_store
+        self.data_czech = data_czech
+        self.data_english = data_english
+
+    def ingest_english(self):
+
+        embedding = OpenAIEmbeddings(
+            openai_api_key=self.openai_api_key,
+            model="text-embedding-3-large",
+        )
+
+        loader = DirectoryLoader(
+            self.data_english,
+            show_progress=True,
+        )
+
+        documents = loader.load()
+        text_splitter = RecursiveCharacterTextSplitter(
+            chunk_size=self.chunk,
+            chunk_overlap=self.overlap,
+        )
+        texts = text_splitter.split_documents(documents)
+
+        vectordb = Chroma.from_documents(
+            documents=texts,
+            embedding=embedding,
+            persist_directory=self.english_store,
+            collection_metadata={"hnsw:space": "cosine"},
+        )
+
+        print("\n English vector Store Created.......\n\n")
+
+    def ingest_czech(self):
+        embedding_model = "Seznam/simcse-dist-mpnet-paracrawl-cs-en"
+        model_kwargs = {"device": "cpu"}
+        encode_kwargs = {"normalize_embeddings": False}
+        embedding = HuggingFaceEmbeddings(
+            model_name=embedding_model,
+            model_kwargs=model_kwargs,
+            encode_kwargs=encode_kwargs,
+        )
+
+        loader = DirectoryLoader(
+            self.data_czech,
+            show_progress=True,
+        )
+
+        documents = loader.load()
+        text_splitter = RecursiveCharacterTextSplitter(
+            chunk_size=self.chunk,
+            chunk_overlap=self.overlap,
+        )
+
+        texts = text_splitter.split_documents(documents)
+        vectordb = Chroma.from_documents(
+            documents=texts,
+            embedding=embedding,
+            persist_directory=self.czech_store,
+            collection_metadata={"hnsw:space": "cosine"},
+        )
+
+        print("\n Czech vector Store Created.......\n\n")
+
+
+"""       
+    
+    
+    
+openai_api_key = "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"
+persist_directory = "stores/store_512"
+data = "data/"
+chunk = 512
+overlap = 256
+
+embedding = OpenAIEmbeddings(
+    openai_api_key=openai_api_key,
+    model="text-embedding-3-large",
+    #    model_kwargs={"device": "cpu"},
+)
+
+loader = DirectoryLoader(
+    data, glob="**/*.pdf", show_progress=True, loader_cls=PyPDFLoader
+)
+documents = loader.load()
+text_splitter = RecursiveCharacterTextSplitter(
+    chunk_size=chunk,
+    chunk_overlap=overlap,
+)
+texts = text_splitter.split_documents(documents)
+
+vectordb = Chroma.from_documents(
+    documents=texts,
+    embedding=embedding,
+    persist_directory=persist_directory,
+    collection_metadata={"hnsw:space": "cosine"},
+)
+
+print("\n Vector Store Created.......\n\n")
+
+"""

requirements.txt

@@ -0,0 +1,15 @@
+langchain
+fastapi
+uvicorn
+python-multipart
+ctransformers
+qdrant-client
+torch
+sentence_transformers
+chromadb
+pytesseract
+fitz
+libpff-python
+openai
+tiktoken
+frontend

support/images-text.py

@@ -0,0 +1,60 @@
+import fitz  # PyMuPDF
+from PIL import Image
+import pytesseract
+import io
+import os
+
+def extract_images_from_pdf(pdf_path, output_folder, dpi=300):
+    doc = fitz.open(pdf_path)
+    images = []
+    for page_num in range(len(doc)):
+        page = doc.load_page(page_num)
+        pix = page.get_pixmap(matrix=fitz.Matrix(dpi / 72, dpi / 72))
+        img_bytes = pix.tobytes("ppm")
+        
+        # Convert PPM to JPEG using PIL and save
+        image = Image.open(io.BytesIO(img_bytes))
+        output_filename = f"{output_folder}/page_{page_num + 1}.jpg"
+        image.save(output_filename, "JPEG")
+        
+        # Also store the image bytes for OCR processing
+        img_converted_bytes = io.BytesIO()
+        image.save(img_converted_bytes, format='JPEG')
+        images.append(img_converted_bytes.getvalue())
+        
+    print(f"Images saved to {output_folder}")
+    return images
+
+def ocr_images(images):
+    text_from_images = []
+    for image_bytes in images:
+        image = Image.open(io.BytesIO(image_bytes))
+        text = pytesseract.image_to_string(image, lang='ces')
+        text_from_images.append(text)
+    return text_from_images
+
+def save_text_to_file(text, file_path):
+    with open(file_path, 'w', encoding='utf-8') as file:
+        file.write(text)
+
+# Path to your PDF file and output locations
+pdf_path = "norm.pdf"
+output_folder = "extracted_images"
+output_text_file = "ocr_results.txt"
+
+# Ensure the output directory exists
+os.makedirs(output_folder, exist_ok=True)
+
+# Extract images from the PDF and save them as JPEG
+images = extract_images_from_pdf(pdf_path, output_folder)
+
+# Transcribe text from the extracted images with Czech language
+transcribed_texts = ocr_images(images)
+
+# Combine all texts into a single string
+combined_text = "\n".join(transcribed_texts)
+
+# Save OCR results to a text file
+save_text_to_file(combined_text, output_text_file)
+
+print(f"OCR results saved to {output_text_file}")

support/process_archives.py

@@ -0,0 +1,87 @@
+import pypff
+import re
+from datetime import datetime
+import pandas as pd
+
+print(f"[INFO] pypff version: {pypff.get_version()}")
+
+DATA_PATH = "data\Archiv_2023_2.pst"
+
+# Patterns to match and remove
+patterns = [
+    r"\bM\s\+\d{3}\d{3}\d{3}\d{3}\b",  # Phone numbers
+    r"\bP\s\+\d{3}\d{3}\d{3}\d{3}\b",  # Phone numbers
+    r"\S+@\S+",  # Email addresses
+    r"http[s]?://\S+",  # URLs
+    r"preciosalighting\.com\s*<",
+    r"Facebook\s*<",  # Social Media links
+    r"Instagram\s*<",
+    r"Youtube\s*<",
+    r"Pinterest\s*<",
+    r"Linkedin\s*<",
+    r"_+",  # Line of underscores
+    # Czech legal disclaimer
+    r"Tento e-mail je určen pouze.*od odesílatele k adresátovi\.",
+    # English legal disclaimer
+    r"This e-mail transmission is intended solely.*from the sender to the recipient\.",
+    r"From:.*\n?",
+    r"Sent:.*\n?",
+    r"To:.*\n?",
+    r"Cc:.*\n?",
+    r"Subject:.*\n?",
+    r";",  # Semicolons
+    r"[^\w\s,.]",
+]
+
+
+def extract_emails(pst_file):
+    opened_pst = pypff.open(pst_file)
+    root = opened_pst.get_root_folder()
+
+    emails = []
+
+    def process_folder(folder):
+        for folder in folder.sub_folders:
+            process_folder(folder)
+        for message in folder.sub_messages:
+            emails.append(
+                {
+                    "subject": message.subject,
+                    "body": message.plain_text_body,
+                    "sender": message.sender_name,
+                    "date": message.delivery_time,
+                }
+            )
+
+    process_folder(root)
+    return emails
+
+
+def format_item(item, patterns):
+    date = item["date"].strftime("%Y-%m-%d")
+    body = item["body"].decode("utf-8")
+    for pattern in patterns:
+        body = re.sub(pattern, "", body)
+    body = re.sub("\s+", " ", body).strip()
+
+    return {
+        "subject": item["subject"],
+        "body": body,
+        "sender": item["sender"],
+        "date": date,
+    }
+
+
+def main():
+    dataset_list = []
+    emails = extract_emails(DATA_PATH)
+    for email in emails:
+        dataset_list.append(format_item(email, patterns))
+
+    df = pd.DataFrame(dataset_list)
+    df.head()
+    df.to_csv("data\emails.csv", index=True, header=True, sep=";")
+
+
+if __name__ == "__main__":
+    main()

templates/index.html

@@ -0,0 +1,316 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>RAG question answer bot</title>
+    <link href="https://fonts.googleapis.com/css2?family=Poppins:wght@400;600&display=swap" rel="stylesheet">
+    <link href="https://cdn.jsdelivr.net/npm/[email protected]/dist/css/bootstrap.min.css" rel="stylesheet">
+    <style>
+        body {
+            background-color: black;
+            font-family: 'Poppins', sans-serif;
+            color: white;
+        }
+        .chat-container {
+            max-width: 800px;
+            margin: 50px auto;
+            margin-top: 10%;
+            padding: 20px;
+            background-color: #333;
+            border-radius: 10px;
+        }
+        .chat-heading {
+            text-align: center;
+            font-size: 2.5em;
+            font-weight: 600;
+            margin-bottom: 30px;
+            color: #ffd700; /* Golden color for the heading */
+        }
+        .chat-input {
+            margin-top: 20px; /* Added margin */
+            margin-bottom: 20px;
+            height: 100px; /* Increased height */
+        }
+        .chat-button {
+            background-color: green;
+            color: white;
+            padding: 10px 20px;
+            font-size: 1.2em;
+        }
+        .chat-response {
+            background-color: #444;
+            padding: 15px;
+            border-radius: 5px;
+            min-height: 100px; /* Minimum height for the response box */
+            margin-top: 20px;
+        }
+        .accordion {
+
+            background-color: #444;
+            border-radius: 5px;
+        }
+        .accordion-button {
+            color: white;
+            background-color: #555;
+        }
+        .accordion-body {
+            color: white; /* Improved visibility of text */
+        }
+        pre {
+            white-space:pre-wrap;
+        }
+        #exampleQueries {
+            margin-top: 10px;
+            text-align: center;
+        }
+
+        #exampleQueries button {
+            margin-right: 10px;
+            margin-bottom: 5px;
+        }
+
+
+        /* The switch - the box around the slider */
+        
+        .language-toggle {
+           text-align: right;
+            margin: 16px;
+            position: relative; /* Added position relative to align the label */
+        }
+
+        .switch {
+        position: relative;
+        display: inline-block;
+        width: 60px;
+        height: 34px;
+        }
+
+        /* Hide default HTML checkbox */
+        .switch input {
+        opacity: 0;
+        width: 0;
+        height: 0;
+        }
+
+        /* The slider */
+        .slider {
+        position: absolute;
+        cursor: pointer;
+        top: 0;
+        left: 0;
+        right: 0;
+        bottom: 0;
+        background-color: #ccc;
+        -webkit-transition: .4s;
+        transition: .4s;
+        }
+
+        .slider:before {
+        position: absolute;
+        content: "";
+        height: 26px;
+        width: 26px;
+        left: 4px;
+        bottom: 4px;
+        background-color: white;
+        -webkit-transition: .4s;
+        transition: .4s;
+        }
+
+        input:checked + .slider {
+        background-color: #2196F3;
+        }
+
+        input:focus + .slider {
+        box-shadow: 0 0 1px #2196F3;
+        }
+
+        input:checked + .slider:before {
+        -webkit-transform: translateX(26px);
+        -ms-transform: translateX(26px);
+        transform: translateX(26px);
+        }
+
+        /* Rounded sliders */
+        .slider.round {
+        border-radius: 34px;
+        }
+
+        .slider.round:before {
+        border-radius: 50%;
+        }
+
+    </style>
+
+</head>
+<body>
+    <div class="container chat-container">
+        <h1 class="chat-heading">RAG Search</h1>
+
+
+        <!-- Tab Navigation -->
+        <ul class="nav nav-tabs" id="myTab" role="tablist">
+            <li class="nav-item" role="presentation">
+                <button class="nav-link active" id="retriever-tab" data-bs-toggle="tab" data-bs-target="#retriever" type="button" role="tab" aria-controls="retriever" aria-selected="true">Retriever</button>
+            </li>
+            <li class="nav-item" role="presentation">
+                <button class="nav-link" id="ingest-tab" data-bs-toggle="tab" data-bs-target="#ingest" type="button" role="tab" aria-controls="ingest" aria-selected="false">Ingest</button>
+            </li>
+        </ul>
+
+
+        <!-- Language Toggle Switch -->
+        <!-- Tab Content -->
+        <div class="tab-content" id="myTabContent">
+            <!-- Retriever Tab Pane -->
+            <div class="tab-pane fade show active" id="retriever" role="tabpanel" aria-labelledby="retriever-tab">
+                <!-- Language Toggle Switch -->
+                <div class="language-toggle">
+                    <label class="switch">
+                        <input type="checkbox" class="languageCheckbox"> <!-- Changed to class -->
+                        <span class="slider round"></span>
+                    </label>
+                    <span class="languageLabel">English</span> <!-- Changed to class -->
+                </div>
+                
+        <div class="accordion" id="appDescriptionAccordion">
+            <div class="accordion-item">
+                <h2 class="accordion-header" id="descriptionHeading">
+                    <button class="accordion-button collapsed" type="button" data-bs-toggle="collapse" data-bs-target="#collapseDescription" aria-expanded="true" aria-controls="collapseDescription">
+                        About This App
+                    </button>
+                </h2>
+
+                <div id="collapseDescription" class="accordion-collapse collapse" aria-labelledby="descriptionHeading" data-bs-parent="#appDescriptionAccordion">
+                    <div class="accordion-body text-dark">
+                        This is a RAG implementation using Open Source stack. Intel's Neural Chat has been used to build this app along with BGE Embeddings as an embedding model, Chroma DB as a vector store, and Langchain & CTransformers as an orchestration frameworks.
+                    </div>
+                </div>
+            </div>
+        </div>
+        <!-- Example Queries Section -->
+        <div id="exampleQueries" class="mb-3">
+            <h2 class="h5">Try Example Queries:</h2>
+            <button class="btn btn-sm btn-secondary example-query">What cable can you use to hang a pendant light on?</button>
+            <button class="btn btn-sm btn-secondary example-query">What is the minimal gauge of live wires ?</button>
+            <button class="btn btn-sm btn-secondary example-query">What flamability fequirements do plastic enclosure have to meet ?</button>
+        </div>
+
+        <div class="row">
+            <div class="col">
+                <textarea id="userInput" class="form-control chat-input" placeholder="Type your query here..."></textarea>
+                <button id="submitBtn" class="btn chat-button">Submit</button>
+                <div id="response" class="chat-response"></div>
+            </div>
+        </div>
+    </div>
+
+
+
+                <!-- Ingest Tab Pane -->
+                <div class="tab-pane fade" id="ingest" role="tabpanel" aria-labelledby="ingest-tab">
+                        <!-- Language Toggle Switch -->
+                        <div class="language-toggle">
+                            <label class="switch">
+                                <input type="checkbox" class="languageCheckbox"> <!-- Changed to class -->
+                                <span class="slider round"></span>
+                            </label>
+                            <span class="languageLabel">English</span> <!-- Changed to class -->
+                        </div>
+
+                    <!-- Ingest Content -->
+                    <h2>Document Ingestion</h2>
+                    <input type="text" id="folderPath" class="form-control my-2" placeholder="Enter the path to your data folder" />
+                    <button id="ingestBtn" class="btn chat-button">Ingest</button>
+                    <div id="ingestResponse" class="chat-response"></div>
+                </div>
+            </div>
+        </div>
+
+
+
+
+    <script src="https://cdn.jsdelivr.net/npm/[email protected]/dist/js/bootstrap.bundle.min.js"></script>
+    <script>
+        // Function to update language labels and synchronize toggle states
+        function updateLanguageTogglesAndLabels(checked) {
+            document.querySelectorAll('.languageCheckbox').forEach(function(checkbox) {
+                checkbox.checked = checked; // Synchronize toggle state
+            });
+            document.querySelectorAll('.languageLabel').forEach(function(label) {
+                label.textContent = checked ? "Czech" : "English"; // Update all labels
+            });
+        }
+
+        // Attach event listeners to all language toggle switches
+        document.querySelectorAll('.languageCheckbox').forEach(function(checkbox) {
+            checkbox.addEventListener('change', function() {
+                updateLanguageTogglesAndLabels(this.checked);
+            });
+        });
+
+        document.getElementById('submitBtn').addEventListener('click', async function() {
+            var userInput = document.getElementById('userInput').value;
+            
+            var languageCheckbox = document.querySelector('.languageCheckbox').checked;   
+            var language = languageCheckbox ? "czech" : "english";
+            document.getElementById('response').innerHTML = '<p>Processing...</p>';
+            const formData = new FormData();
+            formData.append('query', userInput);
+            formData.append('language', language);
+
+            try {
+                const response = await fetch('/get_response', {
+                    method: 'POST',
+                    body: formData
+                });
+
+                if (!response.ok) {
+                    throw new Error('Network response was not ok');
+                }
+
+                const data = await response.json();
+                document.getElementById('response').innerHTML = `<p>${data.answer}</p><br><pre><b>Context: </b> ${data.source_document}</pre><br><pre><b>Source Document: </b> ${data.doc}</pre>`;
+            } catch (error) {
+                console.error('Error:', error);
+                document.getElementById('response').innerHTML = '<p>Error processing your request</p>';
+            }
+        });
+
+        // Add event listeners to example queries
+        document.querySelectorAll('.example-query').forEach(item => {
+            item.addEventListener('click', function() {
+                document.getElementById('userInput').value = this.textContent; // Insert clicked query into textarea
+            });
+        });
+
+        // Ingest data
+        document.getElementById('ingestBtn').addEventListener('click', async function() {
+            var folderPath = document.getElementById('folderPath').value;
+            var languageCheckbox = document.querySelector('.languageCheckbox').checked;
+            var language = languageCheckbox ? "czech" : "english"; // Determine the language based on the checkbox
+            document.getElementById('ingestResponse').innerHTML = '<p>Starting ingestion...</p>';
+            try {
+                const response = await fetch('/ingest_data', {
+                    method: 'POST',
+                    headers: {
+                        'Content-Type': 'application/x-www-form-urlencoded',
+                    },
+                    body: `folderPath=${folderPath}&language=${language}` // Include the language in the request body
+                });
+        
+                if (!response.ok) {
+                    throw new Error('Network response was not ok');
+                }
+        
+                const data = await response.json();
+                document.getElementById('ingestResponse').innerHTML = `<p>${data.message}</p>`;
+            } catch (error) {
+                console.error('Error:', error);
+                document.getElementById('ingestResponse').innerHTML = '<p>Error during ingestion process</p>';
+            }
+        });
+    </script>
+</body>
+</html>