Rag-node folder name change

RAG-pipelines/RAG-Node/DOCKER_TAG

+cicd.ai4eu-dev.eu:7444/tutorials/llms/test-ragpipeline	
\ No newline at end of file

RAG-pipelines/RAG-Node/Dockerfile

+FROM python:3.10-slim
+
+RUN apt-get update -y 
+RUN apt-get install -y python3-pip python3-dev 
+RUN pip3 install --upgrade pip
+COPY requirements.txt .
+RUN pip3 install -r requirements.txt
+
+ARG TIMEZONE=Europe/Berlin
+RUN ln -sf /usr/share/zoneinfo/$TIMEZONE /etc/localtime && \
+    echo "$TIMEZONE" > /etc/timezone
+
+RUN mkdir /ragp
+COPY . /ragp
+WORKDIR /ragp
+RUN python3 -m grpc_tools.protoc --python_out=. --proto_path=. --grpc_python_out=. rag.proto
+
+ENTRYPOINT python3 -u server.py
\ No newline at end of file

RAG-pipelines/RAG-Node/README.md

+### Current pipeline architecture:
+
+The refactored RAG pipeline (single-node) is designed to handle the connection to the unified LLM interface.
+
+![alt text](image.png)
+
+Key features
+
+- Bi-directional streaming is enabled
+- Loading and ingestion - dual possibility (PDF Upload and FAISS embeddings)
+- Save workspace for future use
+
+
+This pipeline is currently in developmental stages and will undergo frequent changes. Contributions, suggestions, and feedback are welcome to help improve this project.
+
+Please refer to the following ticket to better understand the pipeline structure, eclipse/graphene/tutorials#45.

RAG-pipelines/RAG-Node/app.py

+from flask import Flask, request, jsonify, render_template, send_file
+import os
+from rag_utils import process_pdf, start_df, update_df
+import pandas as pd
+from sklearn.cluster import KMeans
+import seaborn as sns
+import matplotlib.pyplot as plt
+from io import BytesIO
+from visualization import plot_embeddings
+from langchain.embeddings import HuggingFaceEmbeddings
+import json
+from datetime import datetime
+import logging
+
+app = Flask(__name__)
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.DEBUG)
+
+user_question = []
+user_answer = []
+readme_ratings = {}
+ratings_list = []
+faithfulness = None
+relevancy= None
+local_directory = os.getenv("SHARED_FOLDER_PATH")
+
+
+def save_uploaded_file(file):
+    file_path = os.path.join("/tmp", file.filename)
+    with open(file_path, "wb") as f:
+        f.write(file.read())
+    return file_path
+
+
+@app.route("/")
+def index():
+    global faithfulness
+    global relevancy
+
+    with open("metrics.txt", mode="r") as f:        
+        content = ""
+        for line in f.readlines():
+            content += line.strip('')  # Concatenate lines with hyphen
+            
+        # Split the content into two items based on '|'
+        metrics_list = content.split('|')
+        metrics_list = [item.strip('') for item in metrics_list]  # Remove trailing hyphen        
+    try:
+        faithfulness=metrics_list[-2]
+        relevancy=metrics_list[-1].rstrip('\n')
+    
+    except:
+        faithfulness=metrics_list[:-1]
+        relevancy=metrics_list[-1].rstrip('\n')
+
+
+    return render_template("index.html", faithfulness= faithfulness , relevancy = relevancy)
+
+
+@app.route("/handle_request", methods=["POST"])
+def handle_request():
+    action = request.args.get("action")
+
+    # Upload the PDF
+    if action == "upload":
+        file = request.files.get("file")
+        if not file:
+            return jsonify({"response": "No file provided"}), 400
+
+        chunk_size = int(request.form.get("chunk_size", 500))
+        chunk_overlap = int(request.form.get("chunk_overlap", 50))
+
+        file_path = save_uploaded_file(file)
+        return jsonify(
+            {
+                "response": f"File uploaded successfully as {file_path}. Ready to process."
+            }
+        )
+
+    # Other PDF parameters such as chunk size and chunk overlap and create embeddings
+    elif action == "process":
+        file_name = request.form.get("file_name")
+        if not file_name:
+            return jsonify({"response": "No file specified for processing"}), 400
+
+        chunk_size = int(request.form.get("chunk_size", 500))
+        chunk_overlap = int(request.form.get("chunk_overlap", 50))
+
+        file_path = os.path.join("/tmp", file_name)
+        if not os.path.exists(file_path):
+            return jsonify({"response": f"File {file_name} does not exist"}), 400
+
+        response = process_pdf(
+            file_path, chunk_size, chunk_overlap
+        )  # Create embeddings here
+        return jsonify({"response": response})
+
+    # If user would like to upload existing embeddings, it can be done here
+    elif action == "upload_index":
+        file = request.files.get("file")
+        if not file:
+            return jsonify({"response": "No file provided"}), 400
+
+        response = "Should be implemented"
+        return jsonify({"response": response})
+
+    # History of Q&A
+    elif action == "chat":
+        request_data = request.get_json()
+        question = request_data.get("question", "")
+        if not question:
+            return jsonify({"response": "No question provided"}), 400
+
+        user_question.clear()
+        user_question.append(question)
+
+        # Check results file for the answer
+        question, answer = check_results_file()
+
+        if answer:
+            return jsonify({"response": [{"question": question, "answer": answer}]})
+        else:
+            return jsonify({"response": []})
+
+    else:
+        return jsonify({"response": "Invalid action"}), 400
+
+
+def check_results_file():
+    # Read the stored contents (q_id, question and answer) created by the server
+    file_path = os.path.join(local_directory, "results.txt")
+
+    if not os.path.exists(file_path):
+        print("Results file not found.")
+        return "", ""
+
+    with open(file_path, "r") as file:
+        lines = file.readlines()
+
+    question, answer = None, None
+
+    for line in lines:
+        line = line.strip()
+        if line.startswith("Question:"):
+            question = line.split(":", 1)[1].strip()
+        elif line.startswith("Answer:"):
+            answer = line.split(":", 1)[1].strip()
+
+    if not question or not answer:
+        print("No question or answer found in the results file.")
+        return "", ""
+
+    return question, answer
+
+
+def get_user_question():
+    # Send the user question to the server for answer retrieval
+    return user_question
+
+
+# Plotting begins here.............
+
+
+@app.route("/umap_plot")
+def umap_plot():
+    embeddings_model = HuggingFaceEmbeddings(model_name="BAAI/bge-base-en-v1.5")
+    df = start_df()
+    r_question, r_answer = check_results_file()
+
+    question_df = pd.DataFrame(
+        {
+            "id": "question",
+            "question": r_question,
+            "embedding": [embeddings_model.embed_query(r_question)],
+        }
+    )
+    answer_df = pd.DataFrame(
+        {
+            "id": "answer",
+            "answer": r_answer,
+            "embedding": [embeddings_model.embed_query(r_answer)],
+        }
+    )
+    df = pd.concat([question_df, answer_df, df], ignore_index=True)
+    df = update_df(df)
+    fig = plot_embeddings(df)
+    return send_plot_as_image(fig)
+
+
+@app.route("/cluster_plot")
+def cluster_plot():
+    n_clusters = 10
+    df = start_df()
+    df = update_df(df)
+    X = df[["UMAP_x", "UMAP_y"]]
+    kmeans = KMeans(n_clusters=n_clusters, random_state=42)
+    df["cluster"] = kmeans.fit_predict(X)
+
+    plt.figure(figsize=(10, 8))
+    sns.scatterplot(
+        x="UMAP_x",
+        y="UMAP_y",
+        hue="cluster",
+        data=df,
+        palette="viridis",
+        s=100,
+        alpha=0.8,
+    )
+    plt.title("Clusters based on UMAP embeddings")
+
+    img_bytes = BytesIO()
+    plt.savefig(img_bytes, format="png")
+    img_bytes.seek(0)
+    plt.close()
+
+    return send_file(img_bytes, mimetype="image/png", as_attachment=False)
+
+
+def send_plot_as_image(fig):
+    img_bytes = fig.to_image(format="png")
+    return send_file(BytesIO(img_bytes), mimetype="image/png")
+
+@app.route('/rate_readme', methods=['POST'])
+def rate_readme():
+    global faithfulness
+    global relevancy
+    faithfulness_score = float(faithfulness)
+    relevancy_score = float(relevancy)
+    
+    try:
+        data = request.json
+        rating = data['rating']
+        feedback = data.get('feedback', '')  # Get the feedback string, default to empty string if not provided
+        readme_ratings.setdefault(local_directory, []).append({'rating': rating, 'feedback': feedback, 'faithfulness' : faithfulness_score, 'answer_Relevancy' : relevancy_score})
+        logger.info(readme_ratings)
+        timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+        try:
+            ratings_filename = os.path.join(local_directory, "ratings.json")
+            rating_dict={"Rating": rating, "Feedback": feedback, "Faithfulness": faithfulness_score, "Answer_Relevancy": relevancy_score, "Timestamp": timestamp}
+            ratings_list.append(rating_dict)
+            
+            with open(ratings_filename, 'w') as file:                
+                json.dump(ratings_list, file, indent=4)
+                logger.info(f"Rating: {rating}, Feedback: {feedback}, Faithfulness: {faithfulness_score}, Answer_Relevancy: {relevancy_score}, Timestamp: {timestamp}\n")
+                
+        except Exception as e:
+            logger.exception("An error occurred")
+        return jsonify({'success': True})
+    except Exception as e:
+        logger.info("Exception:", e)
+        return jsonify({'success': False, 'error': str(e)})
+
+def app_run():
+    app.run(host="0.0.0.0", port=8062, debug=False)

RAG-pipelines/RAG-Node/client.py

+import grpc
+import rag_pb2
+import rag_pb2_grpc
+
+
+def run():
+    # Dummy implementation
+    with grpc.insecure_channel("localhost:8061") as channel:
+        stub = rag_pb2_grpc.RAGServiceStub(channel)
+        responses = stub.instruct_llm_query(rag_pb2.Empty())
+        for response in responses:
+            print(f"Received LLM Query: {response.input.user}")
+            print(f"Question: {response.qa.question}")
+
+
+if __name__ == "__main__":
+    run()

RAG-pipelines/RAG-Node/rag.proto

+syntax = "proto3";
+
+message Empty {
+
+}
+
+message LLMConfig {
+    double temp = 1;
+}
+
+message PromptInput {
+    string system = 1;
+    string user = 2;
+    string context = 3;
+    string prompt = 4;
+}
+
+message UserQuestion {
+    string question = 1;
+}
+
+message ConvoID {
+    string q_id = 1;
+}
+
+message LLMQuery {
+    LLMConfig config = 1;
+    PromptInput input = 2;
+    UserQuestion qa = 3;
+    ConvoID id = 4;
+}
+
+message LLMAnswer {
+    string text = 1;
+    ConvoID id = 2;  
+    string relevant_context = 3;
+}
+message Status {
+    string message = 1;
+}
+
+service RAGService {
+    rpc instruct_llm_query(Empty) returns( stream LLMQuery);
+    rpc retrieve_llm_response( stream LLMAnswer) returns(Status);
+}

RAG-pipelines/RAG-Node/rag_pb2.py

+# -*- coding: utf-8 -*-
+# Generated by the protocol buffer compiler.  DO NOT EDIT!
+# source: rag.proto
+# Protobuf Python Version: 5.26.1
+"""Generated protocol buffer code."""
+
+from google.protobuf import descriptor as _descriptor
+from google.protobuf import descriptor_pool as _descriptor_pool
+from google.protobuf import symbol_database as _symbol_database
+from google.protobuf.internal import builder as _builder
+# @@protoc_insertion_point(imports)
+
+_sym_db = _symbol_database.Default()
+
+
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(
+    b'\n\trag.proto"\x07\n\x05\x45mpty"\x19\n\tLLMConfig\x12\x0c\n\x04temp\x18\x01 \x01(\x01"L\n\x0bPromptInput\x12\x0e\n\x06system\x18\x01 \x01(\t\x12\x0c\n\x04user\x18\x02 \x01(\t\x12\x0f\n\x07\x63ontext\x18\x03 \x01(\t\x12\x0e\n\x06prompt\x18\x04 \x01(\t" \n\x0cUserQuestion\x12\x10\n\x08question\x18\x01 \x01(\t"\x17\n\x07\x43onvoID\x12\x0c\n\x04q_id\x18\x01 \x01(\t"t\n\x08LLMQuery\x12\x1a\n\x06\x63onfig\x18\x01 \x01(\x0b\x32\n.LLMConfig\x12\x1b\n\x05input\x18\x02 \x01(\x0b\x32\x0c.PromptInput\x12\x19\n\x02qa\x18\x03 \x01(\x0b\x32\r.UserQuestion\x12\x14\n\x02id\x18\x04 \x01(\x0b\x32\x08.ConvoID"/\n\tLLMAnswer\x12\x0c\n\x04text\x18\x01 \x01(\t\x12\x14\n\x02id\x18\x02 \x01(\x0b\x32\x08.ConvoID"\x19\n\x06Status\x12\x0f\n\x07message\x18\x01 \x01(\t2g\n\nRAGService\x12)\n\x12instruct_llm_query\x12\x06.Empty\x1a\t.LLMQuery0\x01\x12.\n\x15retrieve_llm_response\x12\n.LLMAnswer\x1a\x07.Status(\x01\x62\x06proto3'
+)
+
+_globals = globals()
+_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
+_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, "rag_pb2", _globals)
+if not _descriptor._USE_C_DESCRIPTORS:
+    DESCRIPTOR._loaded_options = None
+    _globals["_EMPTY"]._serialized_start = 13
+    _globals["_EMPTY"]._serialized_end = 20
+    _globals["_LLMCONFIG"]._serialized_start = 22
+    _globals["_LLMCONFIG"]._serialized_end = 47
+    _globals["_PROMPTINPUT"]._serialized_start = 49
+    _globals["_PROMPTINPUT"]._serialized_end = 125
+    _globals["_USERQUESTION"]._serialized_start = 127
+    _globals["_USERQUESTION"]._serialized_end = 159
+    _globals["_CONVOID"]._serialized_start = 161
+    _globals["_CONVOID"]._serialized_end = 184
+    _globals["_LLMQUERY"]._serialized_start = 186
+    _globals["_LLMQUERY"]._serialized_end = 302
+    _globals["_LLMANSWER"]._serialized_start = 304
+    _globals["_LLMANSWER"]._serialized_end = 351
+    _globals["_STATUS"]._serialized_start = 353
+    _globals["_STATUS"]._serialized_end = 378
+    _globals["_RAGSERVICE"]._serialized_start = 380
+    _globals["_RAGSERVICE"]._serialized_end = 483
+# @@protoc_insertion_point(module_scope)

RAG-pipelines/RAG-Node/rag_pb2_grpc.py

+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+
+import grpc
+import warnings
+
+import rag_pb2 as rag__pb2
+
+GRPC_GENERATED_VERSION = "1.65.5"
+GRPC_VERSION = grpc.__version__
+EXPECTED_ERROR_RELEASE = "1.66.0"
+SCHEDULED_RELEASE_DATE = "August 6, 2024"
+_version_not_supported = False
+
+try:
+    from grpc._utilities import first_version_is_lower
+
+    _version_not_supported = first_version_is_lower(
+        GRPC_VERSION, GRPC_GENERATED_VERSION
+    )
+except ImportError:
+    _version_not_supported = True
+
+if _version_not_supported:
+    warnings.warn(
+        f"The grpc package installed is at version {GRPC_VERSION},"
+        + f" but the generated code in rag_pb2_grpc.py depends on"
+        + f" grpcio>={GRPC_GENERATED_VERSION}."
+        + f" Please upgrade your grpc module to grpcio>={GRPC_GENERATED_VERSION}"
+        + f" or downgrade your generated code using grpcio-tools<={GRPC_VERSION}."
+        + f" This warning will become an error in {EXPECTED_ERROR_RELEASE},"
+        + f" scheduled for release on {SCHEDULED_RELEASE_DATE}.",
+        RuntimeWarning,
+    )
+
+
+class RAGServiceStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.instruct_llm_query = channel.unary_stream(
+            "/RAGService/instruct_llm_query",
+            request_serializer=rag__pb2.Empty.SerializeToString,
+            response_deserializer=rag__pb2.LLMQuery.FromString,
+            _registered_method=True,
+        )
+        self.retrieve_llm_response = channel.stream_unary(
+            "/RAGService/retrieve_llm_response",
+            request_serializer=rag__pb2.LLMAnswer.SerializeToString,
+            response_deserializer=rag__pb2.Status.FromString,
+            _registered_method=True,
+        )
+
+
+class RAGServiceServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def instruct_llm_query(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
+
+    def retrieve_llm_response(self, request_iterator, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
+
+
+def add_RAGServiceServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+        "instruct_llm_query": grpc.unary_stream_rpc_method_handler(
+            servicer.instruct_llm_query,
+            request_deserializer=rag__pb2.Empty.FromString,
+            response_serializer=rag__pb2.LLMQuery.SerializeToString,
+        ),
+        "retrieve_llm_response": grpc.stream_unary_rpc_method_handler(
+            servicer.retrieve_llm_response,
+            request_deserializer=rag__pb2.LLMAnswer.FromString,
+            response_serializer=rag__pb2.Status.SerializeToString,
+        ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+        "RAGService", rpc_method_handlers
+    )
+    server.add_generic_rpc_handlers((generic_handler,))
+    server.add_registered_method_handlers("RAGService", rpc_method_handlers)
+
+
+# This class is part of an EXPERIMENTAL API.
+class RAGService(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def instruct_llm_query(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_stream(
+            request,
+            target,
+            "/RAGService/instruct_llm_query",
+            rag__pb2.Empty.SerializeToString,
+            rag__pb2.LLMQuery.FromString,
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+            _registered_method=True,
+        )
+
+    @staticmethod
+    def retrieve_llm_response(
+        request_iterator,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.stream_unary(
+            request_iterator,
+            target,
+            "/RAGService/retrieve_llm_response",
+            rag__pb2.LLMAnswer.SerializeToString,
+            rag__pb2.Status.FromString,
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+            _registered_method=True,
+        )

RAG-pipelines/RAG-Node/rag_utils.py

+import os
+import umap
+import numpy as np
+
+from langchain.vectorstores import FAISS
+from langchain.embeddings.huggingface import HuggingFaceEmbeddings
+from langchain.document_loaders import PyPDFLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+import pandas as pd
+
+# FAISS folder path
+local_directory = os.getenv("SHARED_FOLDER_PATH")
+faiss_folder = os.path.join(local_directory, "faiss_index")
+
+
+def process_pdf(file_name, chunk_size=500, chunk_overlap=50):
+    # Load and process the PDF document
+    loader = PyPDFLoader(file_name)
+    raw_document = loader.load()
+
+    # Split the document into chunks
+    splitter = RecursiveCharacterTextSplitter(
+        chunk_size=chunk_size, chunk_overlap=chunk_overlap
+    )
+    chunks = splitter.split_documents(raw_document)
+
+    # Create and save FAISS index
+    embedding_model = HuggingFaceEmbeddings(model_name="BAAI/bge-base-en-v1.5")
+    faiss_db = FAISS.from_documents(chunks, embedding=embedding_model)
+    faiss_db.save_local(faiss_folder)
+    return "FAISS index processed and saved successfully"
+
+
+def faiss_db():
+    # Retriever contents can be obtained here
+    embeddings = HuggingFaceEmbeddings(model_name="BAAI/bge-base-en-v1.5")
+    try:
+        saved_db = FAISS.load_local(
+            faiss_folder, embeddings, allow_dangerous_deserialization=True
+        )
+        retriever = saved_db.as_retriever(
+            search_type="similarity", search_kwargs={"k": 4}
+        )
+        print(retriever)
+        print(type(retriever))
+        return saved_db, retriever
+    except Exception as e:
+        print(f"Error loading FAISS database: {e}")
+        return None
+
+
+# Start creating the dataframe for visualization
+def start_df():
+    saved_db, retriever = faiss_db()
+    vs = saved_db.__dict__.get("docstore")
+    index_list = saved_db.__dict__.get("index_to_docstore_id").values()
+    doc_cnt = saved_db.index.ntotal
+
+    embeddings_vec = saved_db.index.reconstruct_n()
+    doc_list = list()
+    for i, docid in enumerate(index_list):
+        a_doc = vs.search(docid)
+        doc_list.append(
+            [docid, a_doc.metadata.get("source"), a_doc.page_content, embeddings_vec[i]]
+        )
+    df = pd.DataFrame(doc_list, columns=["id", "metadata", "document", "embedding"])
+    return df
+
+
+def update_df(df):
+    embeddings = np.vstack(df["embedding"].values)
+    umap_model = umap.UMAP(n_neighbors=15, min_dist=0.1, metric="cosine")
+    umap_embeddings = umap_model.fit_transform(embeddings)
+
+    df["UMAP_x"] = umap_embeddings[:, 0]
+    df["UMAP_y"] = umap_embeddings[:, 1]
+    return df

RAG-pipelines/RAG-Node/requirements.txt

+grpcio==1.38.0
+grpcio-tools==1.38.0
+grpc-interceptor
+protobuf==3.16.0
+multithreading
+
+Bootstrap-Flask
+Flask
+Flask-WTF
+flask_session
+WTForms
+requests
+python-dotenv
+
+langchain
+langchain_community 
+
+umap-learn
+pandas
+matplotlib
+sentence-transformers
+faiss-cpu
+
+pypdf
+numpy
+seaborn
+umap-learn
+scikit-learn
+plotly
+kaleido
+ragas==0.1.7
\ No newline at end of file

RAG-pipelines/RAG-Node/server.py

+import grpc
+from concurrent import futures
+import threading
+import time
+
+import rag_pb2
+import rag_pb2_grpc
+from app import app_run, get_user_question
+import json
+import os
+import uuid
+
+from datasets import Dataset 
+from ragas.metrics import faithfulness, answer_relevancy
+from ragas import evaluate
+
+
+system_instruction = """
+user
+You are an assistant for answering questions.
+You are given the extracted parts of a long document and a question. Provide a conversational answer.
+If you don't know the answer, just say "I do not know." Don't make up an answer.
+"""
+
+
+local_directory = os.getenv("SHARED_FOLDER_PATH")
+os.environ["OPENAI_API_KEY"] = "sk-ai-builder-rhzhJNo29KLDuN2gT4JsT3BlbkFJ7Vw7lOks4OqAm4wRw1YZ"
+os.environ['OPENAI_ORGANIZATION'] = "org-XSbDbl5S9lfwrFTncFOBVOMx"
+
+
+class RAGService(rag_pb2_grpc.RAGServiceServicer):
+    def __init__(self):
+        super().__init__()
+        self.responses = {}
+        self.current_question = None  # Store the last processed question
+
+    def instruct_llm_query(self, request, context):
+        while True:
+            params = get_user_question()
+            if not params:
+                time.sleep(0.5)
+                continue
+
+            self.new_question = params[0]
+            if self.new_question == self.current_question:
+                time.sleep(0.5)
+                continue
+
+            self.current_question = self.new_question
+
+            question_id = str(uuid.uuid4())
+            input_text = f"You asked: {self.new_question}"
+
+            Pinput = rag_pb2.PromptInput(
+                system="System instruction here",
+                user=input_text,
+                context="",
+                prompt=input_text,
+            )
+
+            QAinput = rag_pb2.UserQuestion(question=self.new_question)
+            llm_query = rag_pb2.LLMQuery(
+                id=rag_pb2.ConvoID(q_id=question_id), input=Pinput, qa=QAinput
+            )
+            print(llm_query)
+            yield llm_query
+
+            while question_id not in self.responses:
+                time.sleep(0.1)
+
+    def retrieve_llm_response(self, request_iterator, context):
+        qa_pairs = []
+
+        for answer in request_iterator:
+            question_id = answer.id.q_id
+            print(f"Received LLMAnswer for ID {question_id}: {answer.text}")
+
+            self.responses[question_id] = answer.text
+
+            faithfulness, relevancy = self.calculate_metrics(answer.relevant_context, answer.text) 
+            print(f'faithfulness, relevancy: {faithfulness}, {relevancy}')
+
+            qa_pairs.append(
+                {
+                    "question_id": question_id,
+                    "question": self.new_question,
+                    "answer": answer.text,
+                }
+            )
+            with open(os.path.join(local_directory, "results.txt"), "w") as f:
+                f.write(f"Question ID: {question_id}\n")
+                f.write(f"Question: {self.new_question}\n")
+                f.write(f"Answer: {answer.text}\n\n")
+
+            
+            with open("metrics.txt", mode="a+") as f:
+                # for e0, e1, e2, e3, e4, e5 in result:
+                f.write("|" + str(round(faithfulness,3)) + "|" + str(round(relevancy,3)) + "\n")
+                f.close()
+
+
+        return rag_pb2.Status(message="All answers received successfully.")
+    
+    def calculate_metrics(self, relevant_info, rag_output):
+
+        data_samples = {
+            'question': [str(self.new_question)],
+            'answer': [str(rag_output)],
+            'contexts' : [[str(relevant_info)]]
+        }
+
+        dataset = Dataset.from_dict(data_samples)
+        faithfulness_score = evaluate(dataset,metrics=[faithfulness])
+        answer_relevancy_score = evaluate(dataset,metrics=[answer_relevancy])
+        
+        return faithfulness_score["faithfulness"], answer_relevancy_score["answer_relevancy"]
+  
+
+def serve(port):
+    server = grpc.server(futures.ThreadPoolExecutor(max_workers=10))
+    rag_pb2_grpc.add_RAGServiceServicer_to_server(RAGService(), server)
+    server.add_insecure_port(f"[::]:{port}")
+    print(f"Starting server. Listening on port: {port}")
+    server.start()
+    threading.Thread(target=app_run).start()
+    server.wait_for_termination()
+
+
+if __name__ == "__main__":
+    port = 8061
+    open('metrics.txt', 'w').close()
+    serve(port)

RAG-pipelines/RAG-Node/templates/visualizations.html

+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Visualizations</title>
+    <link rel="stylesheet" href="styles.css"> <!-- Optional: Link to your CSS file -->
+</head>
+<body>
+    <div class="container">
+        <h1>RAG Visualizations</h1>
+        <div id="visualizationContent">
+            <!-- Insert your visualizations here, e.g., charts, graphs, here, the UMAP and cluster plot here -->
+            <p>Visualization content will be displayed here.</p>
+        </div>
+        <a href="index.html" class="button">Back to Main Page</a>
+    </div>
+</body>
+</html>

RAG-pipelines/RAG-Node/visualization.py

+import umap
+import plotly.express as px
+from scipy.spatial.distance import cdist, euclidean
+from sklearn.cluster import KMeans
+import matplotlib.pyplot as plt
+import seaborn as sns
+
+
+def plot_embeddings(df):
+    question_coords = df[df["id"] == "question"][["UMAP_x", "UMAP_y"]]
+    answer_coords = df[df["id"] == "answer"][["UMAP_x", "UMAP_y"]]
+
+    if not question_coords.empty and not answer_coords.empty:
+        question_point = question_coords[["UMAP_x", "UMAP_y"]].values[0]
+        answer_point = answer_coords[["UMAP_x", "UMAP_y"]].values[0]
+
+        df["distance_to_question"] = cdist(df[["UMAP_x", "UMAP_y"]], [question_point])
+        df["distance_to_answer"] = cdist(df[["UMAP_x", "UMAP_y"]], [answer_point])
+
+        max_distance = 0.5
+        df_filtered = df[
+            (df["distance_to_question"] < max_distance)
+            | (df["distance_to_answer"] < max_distance)
+        ]
+
+        df_filtered["document"] = df_filtered["document"].fillna("No content available")
+
+        df_filtered["hover_text"] = (
+            df_filtered["document"]
+            .astype(str)
+            .apply(lambda x: "<br>".join(x.split("\n")))
+        )
+
+        fig = px.scatter(
+            df_filtered,
+            x="UMAP_x",
+            y="UMAP_y",
+            color="id",
+            hover_name="id",
+            hover_data={"document": False, "hover_text": True},
+            title="UMAP Projection of Closely Related Documents, Question, and Answer",
+            labels={"UMAP_x": "UMAP Dimension 1", "UMAP_y": "UMAP Dimension 2"},
+            color_discrete_map={"question": "red", "answer": "blue"},
+        )
+
+        fig.add_scatter(
+            x=[question_point[0]],
+            y=[question_point[1]],
+            mode="markers+text",
+            text=["Question"],
+            marker=dict(color="red", size=12),
+            textposition="top center",
+            showlegend=False,
+        )
+
+        fig.add_scatter(
+            x=[answer_point[0]],
+            y=[answer_point[1]],
+            mode="markers+text",
+            text=["Answer"],
+            marker=dict(color="blue", size=12),
+            textposition="top center",
+            showlegend=False,
+        )
+
+        distance = euclidean(question_point, answer_point)
+        mid_x = (question_point[0] + answer_point[0]) / 2
+        mid_y = (question_point[1] + answer_point[1]) / 2
+        fig.add_annotation(
+            x=mid_x,
+            y=mid_y,
+            text=f"Distance: {distance:.2f}",
+            showarrow=True,
+            arrowhead=2,
+            ax=0,
+            ay=-40,
+            font=dict(size=12, color="black"),
+            align="center",
+            bgcolor="white",
+            borderpad=4,
+        )
+
+        fig.update_layout(
+            legend_title="Legend", legend=dict(x=0.8, y=0.1, traceorder="normal")
+        )
+
+        fig.update_layout(
+            legend=dict(
+                orientation="h",
+                yanchor="bottom",
+                y=1.02,
+                xanchor="right",
+                x=1,
+                title="Legend",
+                traceorder="normal",
+                font=dict(size=12),
+                bgcolor="rgba(255, 255, 255, 0)",
+                bordercolor="rgba(0, 0, 0, 0)",
+                borderwidth=0,
+                visible=False,  # Hide the legend
+            )
+        )
+
+        fig.show()
+
+    return fig
+
+
+def cluster(df):
+    n_clusters = 10
+    X = df[["UMAP_x", "UMAP_y"]]
+
+    kmeans = KMeans(n_clusters=n_clusters, random_state=42)
+    df["cluster"] = kmeans.fit_predict(X)
+
+    plt.figure(figsize=(10, 8))
+    sns.scatterplot(
+        x="UMAP_x",
+        y="UMAP_y",
+        hue="cluster",
+        data=df,
+        palette="viridis",
+        s=100,
+        alpha=0.8,
+    )
+    plt.title("Clusters based on UMAP embeddings")
+    plt.show()