init app

.gitignore

@@ -0,0 +1,161 @@
+# 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
+
+# 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/#use-with-ide
+.pdm.toml
+
+# 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/
+.DS_Store

Dockerfile

@@ -0,0 +1,43 @@
+# Use the official Python image as the base image
+FROM python:3.9
+
+# Set the working directory to /app
+WORKDIR /app
+
+# Create a new user and switch to this user
+RUN useradd -m -u 1000 user
+USER user
+
+# Set environment variables for the new user
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:$PATH
+
+# Set the working directory for the user
+WORKDIR $HOME/app
+
+# Copy the current directory contents into the container at /app
+COPY --chown=user . $HOME/app
+
+# Copy requirements file
+COPY ./requirements.txt ~/app/requirements.txt
+
+# Copy poetry configuration files
+COPY pyproject.toml poetry.lock /app/
+
+# Upgrade pip and install poetry
+RUN pip install --upgrade pip
+RUN pip install farm-haystack[faiss]
+RUN pip install poetry
+
+# Set environment variable to create a virtual environment within the project directory
+ENV POETRY_VIRTUALENVS_IN_PROJECT=true
+
+# Lock and install project dependencies using poetry
+RUN poetry lock
+RUN poetry install
+
+# Copy the rest of the application code
+COPY . .
+
+# Define the command to run the app
+CMD ["poetry", "run", "chainlit", "run", "app.py", "--port", "7860"]

README.md

@@ -1,11 +1,45 @@
 ---
-title: Barbie Raq Hf
-emoji: 🐠
-colorFrom: pink
-colorTo: pink
+title: Haystack-rag-about-Barbie
+emoji: 🚀
+colorFrom: indigo
+colorTo: red
 sdk: docker
+app_file: app.py
 pinned: false
-license: openrail
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+
+# What do people think about the Barbie (2023) movie?
+
+This chatbot can help you identify what people think about the Barbie (2023) movie. You can also ask it information about the movie. 
+
+### Mini demo
+
+![](demo.gif)
+
+### App
+
+- Code to do web scraping from natural language query is in [faissdenseretrieval.py](faissdenseretrieval.py)
+- Code to run the app is in [app.py](app.py)
+
+### How it is built:
+
+The application uses Haystack's WebRetriever class to scrape reviews from the internet. It uses a simple NLP query: "IMDB movie reviews for the Barbie movie (2023)" and 100 top k results were fetched.  The results were then stored into a FAISS document store. 
+
+To retrieve answers I used the DensePassageRetriever class from Haystack using the following models:
+
+```
+query_embedding_model="facebook/dpr-question_encoder-single-nq-base",
+passage_embedding_model="facebook/dpr-ctx_encoder-single-nq-base",
+```
+
+The embeddings were applied onto the documents in the document store.
+
+I then initialized a Haystack pipeline whose nodes include a prompt node that uses OpenAI's GPT-4 and the DensePassageRetriever node. Its user interface was built using Chainlit.
+
+### How does it work?
+
+1. The WebRetriever will scrape the internet for reviews of the Barbie movie (2023) based on the natural language query using the SERP API.
+2. The WebRetriever transforms the results into Document objects which can then be saved into a FAISS document store.
+3. The DensePassageRetriever` node will apply embeddings to the documents in the document store and then it will use the embeddings to retrieve the top k results for a given query.
+4. When a user asks a question, the PromptNode will use the top k results to generate an answer using OpenAI's GPT-4.

app.py

@@ -0,0 +1,31 @@
+import chainlit as cl
+from faissdenseretrieval import initialize_documents, initialize_faiss_document_store, initialize_rag_pipeline
+import os
+from dotenv import load_dotenv
+
+# Load environment variables (if any)
+load_dotenv("../.env")
+load_dotenv()
+serp = os.getenv("SERP_API_KEY")
+openai_key = os.getenv("OPENAI_API_KEY")
+
+# Initialize documents
+documents = initialize_documents(serp_key=serp, nl_query="IMDB movie reviews for the Barbie movie (2023)")
+
+# Initialize document store and retriever
+document_store, retriever = initialize_faiss_document_store(documents=documents)
+
+# Initialize pipeline
+query_pipeline = initialize_rag_pipeline(retriever=retriever, openai_key=openai_key)
+
+@cl.on_message
+async def main(message: str):
+    # Use the pipeline to get a response
+    output = query_pipeline.run(query=message)
+
+    # Create a Chainlit message with the response
+    response = output['answers'][0].answer
+    msg = cl.Message(content=response)
+
+    # Send the message to the user
+    await msg.send()

chainlit.md

@@ -0,0 +1,25 @@
+# What do people think about the Barbie (2023) movie?
+
+This chatbot can help you identify what people think about the Barbie (2023) movie. You can also ask it information about the movie. 
+
+### How it is built:
+
+The application uses Haystack's WebRetriever class to scrape reviews from the internet. It uses a simple NLP query: "IMDB movie reviews for the Barbie movie (2023)" and 100 top k results were fetched.  The results were then stored into a FAISS document store. 
+
+To retrieve answers I used the DensePassageRetriever class from Haystack using the following models:
+
+
+query_embedding_model="facebook/dpr-question_encoder-single-nq-base",
+passage_embedding_model="facebook/dpr-ctx_encoder-single-nq-base",
+
+
+The embeddings were applied onto the documents in the document store.
+
+I then initialized a Haystack pipeline whose nodes include a prompt node that uses OpenAI's GPT-4 and the DensePassageRetriever node. Its user interface was built using Chainlit.
+
+### How does it work?
+
+1. The WebRetriever will scrape the internet for reviews of the Barbie movie (2023) based on the natural language query using the SERP API.
+2. The WebRetriever transforms the results into Document objects which can then be saved into a FAISS document store.
+3. The DensePassageRetriever node will apply embeddings to the documents in the document store and then it will use the embeddings to retrieve the top k results for a given query.
+4. When a user asks a question, the PromptNode will use the top k results to generate an answer using OpenAI's GPT-4.

faissdenseretrieval.py

@@ -0,0 +1,90 @@
+from haystack.nodes import WebRetriever
+from haystack.schema import Document
+from typing import List 
+from haystack.document_stores import FAISSDocumentStore
+from haystack.nodes import AnswerParser, PromptNode, PromptTemplate
+from haystack import Pipeline
+from haystack.nodes import  DensePassageRetriever
+import os
+from dotenv import load_dotenv
+
+def initialize_documents(serp_key, nl_query):
+    """
+    Initialize documents retrieved from the SERP API.
+
+    Args:
+        serp_key (str): API key for the SERP API.
+        nl_query (str): Natural language query to retrieve documents for.
+        
+    """
+     # Initialize WebRetriever
+    retriever = WebRetriever(api_key=serp_key, 
+                            mode="preprocessed_documents",
+                            top_k=100)
+
+    # Retrieve documents based a natural language query
+    documents : List[Document] = retriever.retrieve(query=nl_query)
+
+    return documents
+
+def initialize_faiss_document_store(documents):
+    """
+    Initialize a FAISS document store and retriever.
+
+    Args:
+        documents (List[Document]): List of documents to be stored in the document store.
+
+    Returns:
+        document_store (FAISSDocumentStore): FAISS document store.
+        retriever (DensePassageRetriever): Dense passage retriever.
+    """
+    
+    # Initialize document store
+    document_store = FAISSDocumentStore(faiss_index_factory_str="Flat", return_embedding=True)
+
+    retriever = DensePassageRetriever(
+                document_store=document_store,
+                query_embedding_model="facebook/dpr-question_encoder-single-nq-base",
+                passage_embedding_model="facebook/dpr-ctx_encoder-single-nq-base",
+                use_gpu=True,
+                embed_title=True,
+    )
+
+    # Delete existing documents in document store
+    document_store.delete_documents()
+    document_store.write_documents(documents)
+
+    # Add documents embeddings to index
+    document_store.update_embeddings(retriever=retriever)
+
+    return document_store, retriever
+
+def initialize_rag_pipeline(retriever, openai_key):
+    """
+    Initialize a pipeline for RAG-based question answering.
+
+    Args:
+        retriever (DensePassageRetriever): Dense passage retriever.
+        openai_key (str): API key for OpenAI.
+
+    Returns:
+        query_pipeline (Pipeline): Pipeline for RAG-based question answering.
+    """
+    prompt_template = PromptTemplate(prompt = """"Answer the following query based on the provided context. If the context does
+                                                not include an answer, reply with 'The data does not contain information related to the question'.\n
+                                                Query: {query}\n
+                                                Documents: {join(documents)}
+                                                Answer: 
+                                            """,
+                                            output_parser=AnswerParser())
+    prompt_node = PromptNode(model_name_or_path = "gpt-4",
+                            api_key = openai_key,
+                            default_prompt_template = prompt_template,
+                            max_length = 500,
+                            model_kwargs={"stream":True})
+
+    query_pipeline = Pipeline()
+    query_pipeline.add_node(component=retriever, name="Retriever", inputs=["Query"])
+    query_pipeline.add_node(component=prompt_node, name="PromptNode", inputs=["Retriever"])
+
+    return query_pipeline

pyproject.toml

@@ -0,0 +1,26 @@
+[tool.poetry]
+name = "llmops-with-haystack"
+version = "0.1.0"
+description = ""
+authors = ["Laura Gutierrez Funderburk <lgutierrwr@gmail.com>"]
+license = "Apache 2.0"
+readme = "README.md"
+
+[tool.poetry.dependencies]
+python = "^3.9"
+torch = [
+  {url = "https://download.pytorch.org/whl/cpu/torch-1.10.0%2Bcpu-cp39-cp39-linux_x86_64.whl", markers = "sys_platform == 'linux'"},
+]
+farm-haystack = {extras = ["faiss"], version = "^1.21.2"}
+chainlit = "^0.7.0"
+openai = "^0.28.0"
+jupyter = "^1.0.0"
+ipykernel = "^6.25.2"
+python-dotenv = "^1.0.0"
+datasets = "^2.14.5"
+nltk = "^3.8.1"
+
+
+[build-system]
+requires = ["poetry-core"]
+build-backend = "poetry.core.masonry.api"