diff --git a/notebooks/GenAI/.gitignore b/notebooks/GenAI/.gitignore
new file mode 100644
index 0000000..ad23308
--- /dev/null
+++ b/notebooks/GenAI/.gitignore
@@ -0,0 +1,4 @@
+__pycache__
+.venv
+.env
+microsoft-earnings_embeddings.csv
\ No newline at end of file
diff --git a/notebooks/GenAI/embedding_demos/Demo_Suite.py b/notebooks/GenAI/embedding_demos/Demo_Suite.py
new file mode 100644
index 0000000..327eb8d
--- /dev/null
+++ b/notebooks/GenAI/embedding_demos/Demo_Suite.py
@@ -0,0 +1,61 @@
+import streamlit as st
+from styling import global_page_style
+
+def main():
+ # Set page configuration
+ # st.set_page_config(page_title="Azure OpenAI RAG Demo Suite", layout="wide")
+
+ # Title and subtitle
+ # Create columns for logo and title
+
+ st.markdown(
+ f'',
+ unsafe_allow_html=True
+ )
+ st.title("Azure OpenAI RAG Demo Suite")
+ st.markdown("### Demo Overviews")
+ st.write("""
+ Welcome to the Azure OpenAI RAG Demo Suite. On the left side-panel, you will find various demonstrations that showcase the capabilities of Azure OpenAI with a Streamlit frontend. Each demonstration is described in detail below, highlighting their unique features and functionalities.
+ """)
+
+ # Horizontal divider
+ st.markdown("---")
+
+ # Chat with Your Data section
+ st.markdown("### Chat with Your Data using Azure OpenAI API and AI Search Index (AI Search Query)")
+ st.write("""
+ This demo allows users to interact with data stored in their Azure AI Search Index using a combination of semantic and vector search methods.
+ """)
+ st.write("""
+ - **Semantic Search**: Understands the meaning and context of your queries to deliver more relevant results.
+ - **Vector Search**: Utilizes numerical representations of text to find similar content based on cosine similarity.
+ """)
+ # Ensure the user has created the Azure AI search index already
+ st.write("""
+ **Note**: Users must have created the Azure AI search index already as shown here: [Upload your own data and query over it](https://github.com/STRIDES/NIHCloudLabAzure/blob/main/notebooks/GenAI/Azure_Open_AI_README.md)
+ """)
+
+ # Horizontal divider
+ st.markdown("---")
+
+ # Generate & Search with Azure OpenAI Embeddings section
+ st.markdown("### Generate & Search with Azure OpenAI Embeddings (AOAI Embeddings)")
+ st.write("""
+ This demo enables users to generate embeddings from a pre-chunked CSV file and perform searches over the content using vector search.
+ """)
+ st.write("""
+ - **Vectorize**: Creates embeddings based on the "microsoft-earnings.csv" file provided in this directory. The embeddings are generated from the "text" column. The CSV file is pre-chunked, meaning the text has already been split and prepared for embedding generation. A new CSV file will be created to store all generated embeddings, forming your vector store.
+ - **Retrieve**: Generates embeddings based on user queries. The query embedding is then used to search for the most similar document within the vector store using cosine similarity.
+ """)
+ st.write("""
+ Example questions a user can ask about the microsoft-earnings.csv:
+ - What was said about the budget?
+ - How many people utilize GitHub to build software?
+ - How many points did Microsoft Cloud gross margin percentage increase by?
+ - What are the expectations for the Q2 cash flow?
+ """)
+
+
+if __name__ == '__main__':
+ global_page_style()
+ main()
\ No newline at end of file
diff --git a/notebooks/GenAI/embedding_demos/acs_embeddings.py b/notebooks/GenAI/embedding_demos/acs_embeddings.py
deleted file mode 100644
index 8a4a68a..0000000
--- a/notebooks/GenAI/embedding_demos/acs_embeddings.py
+++ /dev/null
@@ -1,79 +0,0 @@
-from langchain.retrievers import AzureCognitiveSearchRetriever
-from langchain.embeddings import OpenAIEmbeddings
-from langchain.vectorstores import FAISS
-from langchain.chains import RetrievalQA
-from langchain.chat_models import AzureChatOpenAI
-from PIL import Image
-import os
-import streamlit as st
-from dotenv import load_dotenv
-
-# load in .env variables
-load_dotenv()
-
-def config_keys():
- # set api keys for AOAI and Azure Search
- os.environ['OPENAI_API_VERSION'] = os.getenv('AZURE_OPENAI_VERSION')
- os.environ['OPENAI_API_KEY'] = os.getenv('AZURE_OPENAI_KEY')
- os.environ['OPENAI_API_BASE'] = os.getenv('AZURE_OPENAI_ENDPOINT')
- os.environ['OPENAI_EMBEDDING_DEPLOYMENT_NAME'] = os.getenv('AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME')
- os.environ['AZURE_COGNITIVE_SEARCH_SERVICE_NAME'] = os.getenv('AZURE_COGNITIVE_SEARCH_SERVICE_NAME')
- os.environ['AZURE_COGNITIVE_SEARCH_API_KEY'] = os.getenv('AZURE_COGNITIVE_SEARCH_API_KEY')
- os.environ['AZURE_COGNITIVE_SEARCH_INDEX_NAME'] = os.getenv('AZURE_COGNITIVE_SEARCH_INDEX_NAME')
-
-
-def main():
- # Streamlit config
- st.title("Demo - Azure OpenAI & Cognitive Search Embeddings")
- image = Image.open('image_logo2.png')
- st.image(image, caption = '')
- st.write('This program is designed to chat over your files in Azure Cognitive Search. \
- Be specific and clear with the questions you ask. \
- Welcome to CHATGPT over your own data !!')
- if 'generated' not in st.session_state:
- st.session_state.generated = []
- if 'past' not in st.session_state:
- st.session_state.past = []
-
- # create your LLM and embeddings. Will be conifuring 'azure' in the openai_api_type parameter.
- llm = AzureChatOpenAI(
- deployment_name = "gpt-35-turbo",
- openai_api_type = "azure",
- model = "gpt-35-turbo",
- temperature=0.7,
- max_tokens=200
- )
-
- embeddings = OpenAIEmbeddings(chunk_size=1, openai_api_type="azure")
-
- # ask for the user query
- query = st.text_input("Enter a search query: ", key='search_term', placeholder="")
-
- if query:
- st.session_state.past.append(query)
-
- # set up Azure Cognitive Search to retrieve documents
- # top_k = 1: we only want first related doc
- retriever = AzureCognitiveSearchRetriever(content_key="content", top_k=1)
-
- # get the relevant document from Azure Cognitive Search that are only relevant to the query being asked
- docs = retriever.get_relevant_documents(query)
-
- # create embedding from the document retrieved and place in a FAISS vector database
- db = FAISS.from_documents(documents=docs, embedding=embeddings)
-
- # set up the chain that will feed the retrieved document to the LLM
- chain = RetrievalQA.from_chain_type(llm=llm, retriever = db.as_retriever(), chain_type="stuff")
-
- # run the chain on the query asked
- response = chain.run(query)
- st.session_state.generated.append(response)
-
- with st.expander('Vector Search'):
- for i in range(len(st.session_state.generated)-1, -1, -1):
- st.info(st.session_state.past[i])
- st.success(st.session_state.generated[i])
-
-if __name__ == '__main__':
- config_keys()
- main()
diff --git a/notebooks/GenAI/embedding_demos/aoai_embeddings.py b/notebooks/GenAI/embedding_demos/aoai_embeddings.py
deleted file mode 100644
index eb694c7..0000000
--- a/notebooks/GenAI/embedding_demos/aoai_embeddings.py
+++ /dev/null
@@ -1,102 +0,0 @@
-import openai
-from openai.embeddings_utils import get_embedding, cosine_similarity # must pip install openai[embeddings]
-import pandas as pd
-import numpy as np
-import os
-import streamlit as st
-import time
-from PIL import Image
-from dotenv import load_dotenv
-
-# load in .env variables
-load_dotenv()
-
-# configure azure openai keys
-openai.api_type = 'azure'
-openai.api_version = os.environ['AZURE_OPENAI_VERSION']
-openai.api_base = os.environ['AZURE_OPENAI_ENDPOINT']
-openai.api_key = os.environ['AZURE_OPENAI_KEY']
-
-def embedding_create():
- # acquire the filename to be embed
- st.subheader("Vector Creation")
- st.write('This program is designed to embed your pre-chunked .csv file. \
- By accomplishing this task, you will be able to chat over all cotent in your .csv via vector searching. \
- Just enter the file and the program will take care of the rest (specify file path if not in this directory). \
- Welcome to CHATGPT over your own data !!')
- filename = st.text_input("Enter a file: ", key='filename', value="")
-
- # start the embeddings process if filename provided
- if filename:
-
- # read the data file to be embed
- df = pd.read_csv('C:\\src\\AzureOpenAI_Gov_Workshop\\' + filename)
- st.write(df)
-
- # calculate word embeddings
- df['embedding'] = df['text'].apply(lambda x:get_embedding(x, engine='text-embedding-ada-002'))
- df.to_csv('C:\\src\\AzureOpenAI_Gov_Workshop\\microsoft-earnings_embeddings.csv')
- time.sleep(3)
- st.subheader("Post Embedding")
- st.success('Embeddings Created Sucessfully!!')
- st.write(df)
-
-
-def embeddings_search():
-
- # Streamlit configuration
- st.subheader("Vector Search")
- st.write('This program is designed to chat over your vector stored (embedding) .csv file. \
- This Chat Bot works alongside the "Embeddings Bot" Chat Bot. \
- Be specific with the information you want to obtain over your data. \
- Welcome to CHATGPT over your own data !!')
- if 'answer' not in st.session_state:
- st.session_state.answer = []
- if 'score' not in st.session_state:
- st.session_state.score = []
- if 'past' not in st.session_state:
- st.session_state.past = []
-
- # read in the embeddings .csv
- # convert elements in 'embedding' column back to numpy array
- df = pd.read_csv('C:\\src\\AzureOpenAI_Gov_Workshop\\microsoft-earnings_embeddings.csv')
- df['embedding'] = df['embedding'].apply(eval).apply(np.array)
-
- # caluculate user query embedding
- search_term = st.text_input("Enter a search query: ", key='search_term', placeholder="")
- if search_term:
- st.session_state.past.append(search_term)
- search_term_vector = get_embedding(search_term, engine='text-embedding-ada-002')
-
- # find similiarity between query and vectors
- df['similarities'] = df['embedding'].apply(lambda x:cosine_similarity(x, search_term_vector))
- df1 = df.sort_values("similarities", ascending=False).head(5)
-
- # output the response
- answer = df1['text'].loc[df1.index[0]]
- score = df1['similarities'].loc[df1.index[0]]
- st.session_state.answer.append(answer)
- st.session_state.score.append(score)
- with st.expander('Vector Search'):
- for i in range(len(st.session_state.answer)-1, -1, -1):
- st.info(st.session_state.past[i])
- st.write(st.session_state.answer[i])
- st.write('Score: ', st.session_state.score[i])
-
-
-def main():
- # Streamlit config
- st.title("Demo-Azure OpenAI Embeddings")
- image = Image.open('image_logo2.png')
- st.image(image, caption = '')
- st.sidebar.title('Chat Bot Type Selection')
- chat_style = st.sidebar.selectbox(
- 'Choose between Embeddings Bot or Search Bot', ['Embeddings Bot','Search Bot']
- )
- if chat_style == 'Embeddings Bot':
- embedding_create()
- elif chat_style == 'Search Bot':
- embeddings_search()
-
-if __name__ == '__main__':
- main()
diff --git a/notebooks/GenAI/embedding_demos/pages/AI_Search_Query.py b/notebooks/GenAI/embedding_demos/pages/AI_Search_Query.py
new file mode 100644
index 0000000..c1c18cd
--- /dev/null
+++ b/notebooks/GenAI/embedding_demos/pages/AI_Search_Query.py
@@ -0,0 +1,106 @@
+from openai import AzureOpenAI
+import os
+import streamlit as st
+from dotenv import load_dotenv
+from styling import global_page_style
+
+# load in .env variables
+load_dotenv()
+
+# Configure Azure OpenAI params, using an Azure OpenAI account with a deployment of an embedding model
+azure_endpoint: str = os.getenv('AZURE_OPENAI_BASE')
+azure_openai_api_key: str = os.getenv('AZURE_OPENAI_KEY')
+azure_openai_api_version: str = os.getenv('AZURE_OPENAI_VERSION')
+azure_ada_deployment: str = os.getenv('AZURE_EMBEDDINGS_DEPLOYMENT')
+azure_gpt_deployment: str = os.getenv('AZURE_GPT_DEPLOYMENT')
+
+# Configure Azure AI Search params
+search_endpoint: str = os.getenv('AZURE_SEARCH_ENDPOINT')
+search_key: str = os.getenv('AZURE_SEARCH_ADMIN_KEY')
+
+def chat_on_your_data(query, search_index, messages):
+ messages.append({"role": "user", "content":query})
+ with st.chat_message("user"):
+ st.markdown(query)
+ with st.spinner('Processing...'):
+ client = AzureOpenAI(
+ azure_endpoint=azure_endpoint,
+ api_key=azure_openai_api_key,
+ api_version=azure_openai_api_version,
+ )
+ completion = client.chat.completions.create(
+ model=azure_gpt_deployment,
+ messages=[
+ {"role": "system", "content": "You are an AI assistant that helps people find information. \
+ Ensure the Markdown responses are correctly formatted before responding."},
+ {"role": "user", "content": query}
+ ],
+ max_tokens=800,
+ temperature=0.7,
+ top_p=0.95,
+ frequency_penalty=0,
+ presence_penalty=0,
+ stop=None,
+ stream=False,
+ extra_body={
+ "data_sources": [{
+ "type": "azure_search",
+ "parameters": {
+ "endpoint": f"{search_endpoint}",
+ "index_name": search_index,
+ "semantic_configuration": "default",
+ "query_type": "vector_simple_hybrid",
+ "fields_mapping": {},
+ "in_scope": True,
+ "role_information": "You are an AI assistant that helps people find information.",
+ "filter": None,
+ "strictness": 3,
+ "top_n_documents": 5,
+ "authentication": {
+ "type": "api_key",
+ "key": f"{search_key}"
+ },
+ "embedding_dependency": {
+ "type": "deployment_name",
+ "deployment_name": azure_ada_deployment
+ }
+ }
+ }]
+ }
+ )
+ print(completion)
+ response_data = completion.to_dict()
+ ai_response = response_data['choices'][0]['message']['content']
+ messages.append({"role": "assistant", "content":ai_response})
+ with st.chat_message("assistant"):
+ st.markdown(ai_response)
+
+def main():
+ st.markdown(
+ f'',
+ unsafe_allow_html=True
+ )
+ st.title("Demo - Azure OpenAI & AI Search")
+ # image = Image.open('image_logo2.png')
+ # st.image(image, caption = '')
+ st.write('This demo showcases an innovative way for users to engage with data housed in their Azure AI Search Index by leveraging both \
+ semantic and vector search techniques. Semantic search enhances the querying process by comprehending the meaning and context of \
+ user queries, thereby providing more pertinent results. Vector search, on the other hand, employs numerical representations of \
+ text to identify similar content using cosine similarity. ***For users to effectively utilize this demo, it is essential that they \
+ have previously created their Azure AI Search Index, following the necessary steps to upload and query their data as outlined [here](https://github.com/STRIDES/NIHCloudLabAzure/blob/main/notebooks/GenAI/Azure_Open_AI_README.md).***')
+ if 'messages' not in st.session_state:
+ st.session_state.messages = []
+ index_name = st.text_input(label="Azure AI Search index name:", value="")
+ st.write('-'*50)
+ if index_name:
+ query = st.chat_input('Input search query here...')
+ for message in st.session_state.messages:
+ with st.chat_message(message["role"]):
+ st.markdown(message['content'])
+ if query:
+ chat_on_your_data(query, index_name, st.session_state.messages)
+
+
+if __name__ == '__main__':
+ global_page_style()
+ main()
\ No newline at end of file
diff --git a/notebooks/GenAI/embedding_demos/pages/AOAI_Embeddings.py b/notebooks/GenAI/embedding_demos/pages/AOAI_Embeddings.py
new file mode 100644
index 0000000..77702de
--- /dev/null
+++ b/notebooks/GenAI/embedding_demos/pages/AOAI_Embeddings.py
@@ -0,0 +1,126 @@
+from openai import AzureOpenAI
+# from openai.embeddings_utils import get_embedding, cosine_similarity # must pip install openai[embeddings]
+import pandas as pd
+import numpy as np
+import os
+import streamlit as st
+import time
+from PIL import Image
+from dotenv import load_dotenv
+from styling import global_page_style
+
+# load in .env variables
+load_dotenv()
+
+# configure azure openai keys
+# openai.api_type = 'azure'
+# openai.api_version = os.environ['AZURE_OPENAI_VERSION']
+# openai.api_base = os.environ['AZURE_OPENAI_ENDPOINT']
+# openai.api_key = os.environ['AZURE_OPENAI_KEY']
+
+def get_embedding(text, engine):
+ client = AzureOpenAI(
+ api_key=os.getenv("Azure_OPENAI_KEY"),
+ api_version=os.getenv('AZURE_OPENAI_VERSION'),
+ azure_endpoint=os.getenv('AZURE_OPENAI_ENDPOINT')
+ )
+
+ embeddings = client.embeddings.create(input = [text], model=engine).data[0].embedding
+ return embeddings
+
+def cosine_similarity(a, b):
+ return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))
+
+def embedding_create():
+ # acquire the filename to be embed
+ st.subheader("Vector Creation")
+ st.write('The process of vectorization involves creating embeddings from the [microsoft-earnings.csv](https://github.com/STRIDES/NIHCloudLabAzure/blob/main/notebooks/GenAI/microsoft-earnings.csv) \
+ file located in the specified directory, utilizing the data in the "text" column. These embeddings are derived from pre-chunked text, \
+ indicating that the text has already been divided and formatted for embedding generation. The resultant embeddings will be \
+ compiled into a new CSV file, which will serve as a vector store for future reference and utilization.')
+ filename = st.text_input("Enter a file: ", key='filename', value="microsoft-earnings.csv")
+
+ # start the embeddings process if filename provided
+ if filename:
+ file_path = os.path.join('..', filename)
+ # read the data file to be embed
+ df = pd.read_csv(file_path)
+ df_placeholder = st.empty()
+ df_placeholder.dataframe(df, width=2000, height=350)
+ button_placeholder = st.empty()
+ if button_placeholder.button("Generate Embeddings"):
+ # calculate word embeddings
+ df['embedding'] = df['text'].apply(lambda x:get_embedding(x, engine=os.getenv("AZURE_EMBEDDINGS_DEPLOYMENT")))
+ df.to_csv('.\\microsoft-earnings_embeddings.csv')
+ time.sleep(3)
+ button_placeholder.success('Embeddings Created Sucessfully!!')
+ df_placeholder.dataframe(df)
+
+
+def embeddings_search():
+
+ # Streamlit configuration
+ st.subheader("Vector Search")
+ st.write('This process generates embeddings based on user queries, utilizing the compiled CSV that was created, to search for the most similar\
+ documents within the vector store by employing cosine similarity.')
+ if 'answer' not in st.session_state:
+ st.session_state.answer = []
+ if 'score' not in st.session_state:
+ st.session_state.score = []
+ if 'past' not in st.session_state:
+ st.session_state.past = []
+
+ # read in the embeddings .csv
+ # convert elements in 'embedding' column back to numpy array
+ df = pd.read_csv('.\\microsoft-earnings_embeddings.csv')
+ df['embedding'] = df['embedding'].apply(eval).apply(np.array)
+
+ # caluculate user query embedding
+ search_term = st.text_area("Enter a search query: ", key='search_term', placeholder="")
+ if search_term:
+ st.session_state.past.append(search_term)
+ search_term_vector = get_embedding(search_term, engine=os.getenv("AZURE_EMBEDDINGS_DEPLOYMENT"))
+
+ # find similiarity between query and vectors
+ df['similarities'] = df['embedding'].apply(lambda x:cosine_similarity(x, search_term_vector))
+ df1 = df.sort_values("similarities", ascending=False).head(5)
+
+ # output the response
+ answer = df1['text'].loc[df1.index[0]]
+ score = df1['similarities'].loc[df1.index[0]]
+ st.session_state.answer.append(answer)
+ st.session_state.score.append(score)
+ with st.expander('Vector Search'):
+ for i in range(len(st.session_state.answer)-1, -1, -1):
+ st.info(st.session_state.past[i])
+ st.write(st.session_state.answer[i])
+ st.write('Score: ', st.session_state.score[i])
+ with st.expander('Top 5 Results'):
+ df1 = df1.reset_index(drop=True)
+ df1.index = df1.index + 1
+ df1 = df1.rename(columns={'Unnamed: 0': 'Row Number'})
+ print(df1)
+ st.dataframe(df1)
+
+
+def main():
+ st.markdown(
+ f'',
+ unsafe_allow_html=True
+ )
+ st.title("Demo-Azure OpenAI Embeddings")
+ # image = Image.open('image_logo2.png')
+ # st.image(image, caption = '')
+ st.sidebar.title('Embedding Function Selection')
+ chat_style = st.sidebar.radio(
+ 'Choose an Embedding function below:',
+ ['Vectorize', 'Retrieve']
+ )
+ if chat_style == 'Vectorize':
+ embedding_create()
+ elif chat_style == 'Retrieve':
+ embeddings_search()
+
+if __name__ == '__main__':
+ global_page_style()
+ main()
diff --git a/notebooks/GenAI/embedding_demos/style.css b/notebooks/GenAI/embedding_demos/style.css
new file mode 100644
index 0000000..e4ebec9
--- /dev/null
+++ b/notebooks/GenAI/embedding_demos/style.css
@@ -0,0 +1,116 @@
+/* Global font family */
+body, .stApp, .stApp * {
+ font-family: sans-serif;
+}
+
+.stApp > header {
+ background-color: #ffffff;
+ color: #f8f8f8;
+ padding: 2rem;
+}
+
+.stApp {
+ background-color: #ffffff;
+ color: #000000;
+ padding: 2rem;
+}
+
+/* Sidebar styling */
+[data-testid="stSidebar"] {
+ background-color: #2c5483;
+ color: #ffffff;
+}
+
+[data-testid="stSidebarNav"] {
+ background-image: url(https://cloud.nih.gov/nih_logo.png);
+ background-repeat: no-repeat;
+ padding-top: 175px;
+ background-position: center 0px;
+ background-size: 200px;
+}
+
+[data-testid="stSidebarNav"]::before {
+ content: "";
+ margin-left: 60px;
+ margin-top: 20px;
+ font-size: 30px;
+ position: relative;
+ top: 100px;
+ font-weight: bold;
+ text-align: center;
+}
+
+/* Ensure all text in sidebar is white */
+[data-testid="stSidebar"] * {
+ color: #ffffff !important;
+}
+
+/* Success message styling */
+.success-message {
+ text-align: center;
+ margin: 20px 0;
+}
+
+/* Main title styling */
+h1 {
+ color: #000400;
+ text-align: center;
+ margin-bottom: 20px;
+ font-size: 40px;
+}
+
+/* Select box styling */
+.css-2trqyj {
+ font-size: 18px;
+}
+
+/* Number input styling */
+.css-1siy2j7 {
+ font-size: 18px;
+}
+
+/* Radio button styling */
+.css-1awt6go {
+ font-size: 18px;
+}
+
+/* DataFrame styling */
+.css-1l269bu {
+ margin-top: 20px;
+}
+
+/* Spinner styling */
+.css-1f6lu8k {
+ margin-top: 20px;
+ text-align: center;
+}
+
+/* Pages Menu styling */
+.st-emotion-cache-1rtdyuf.eczjsme13 {
+ color: #ffffff; /* This sets the text color to white */
+}
+
+.st-emotion-cache-6tkfeg.eczjsme13 {
+ color: #ffffff !important;
+}
+
+.css-10trblm, .css-1r7vkyz, .css-2trqyj {
+ color: black;
+}
+
+/* Modify the streamlit menu in top right corner */
+.st-emotion-cache-w3nhqi.ef3psqc5 {
+ color: black !important;
+}
+
+.st-emotion-cache-1wbqy5l.e17vllj40 {
+ color: black !important;
+}
+
+.st-emotion-cache-jdyw56.en6cib60 {
+ color: black !important;
+}
+
+.st-emotion-cache-cgqxho.ef3psqc6 {
+ color: black !important;
+}
\ No newline at end of file
diff --git a/notebooks/GenAI/embedding_demos/styling.py b/notebooks/GenAI/embedding_demos/styling.py
new file mode 100644
index 0000000..da99bcf
--- /dev/null
+++ b/notebooks/GenAI/embedding_demos/styling.py
@@ -0,0 +1,7 @@
+import streamlit as st
+
+def global_page_style():
+ st.set_page_config(layout="centered")
+ with open('style.css') as f:
+ css = f.read()
+ st.markdown(f'', unsafe_allow_html=True)
\ No newline at end of file
diff --git a/notebooks/GenAI/requirements.txt b/notebooks/GenAI/requirements.txt
index ebf0d31..56ca87e 100644
--- a/notebooks/GenAI/requirements.txt
+++ b/notebooks/GenAI/requirements.txt
@@ -1,12 +1,7 @@
python-dotenv
openai
-openai[embeddings]
pandas
numpy
streamlit
-langchain
-langchain-openai
-langchain-community
-azure-search-documents==11.4.0b6
-tiktoken
-faiss-cpu
+azure-search-documents
+azure-identity