Spaces:
Build error
Build error
Moved keys and hidden variables into streamlit secrets
Browse files- .gitignore +2 -0
- app.py +5 -7
- chatbot/utils.py +10 -11
.gitignore
CHANGED
|
@@ -2,3 +2,5 @@ venv/
|
|
| 2 |
data/
|
| 3 |
.DS_Store
|
| 4 |
config.json
|
|
|
|
|
|
|
|
|
| 2 |
data/
|
| 3 |
.DS_Store
|
| 4 |
config.json
|
| 5 |
+
.streamlit/
|
| 6 |
+
secrets.toml
|
app.py
CHANGED
|
@@ -1,25 +1,23 @@
|
|
| 1 |
-
import
|
| 2 |
|
| 3 |
import streamlit as st
|
| 4 |
|
| 5 |
from chatbot.utils import download_test_data
|
| 6 |
from chatbot.utils import load_data
|
| 7 |
|
|
|
|
|
|
|
|
|
|
| 8 |
# Initialize message history
|
| 9 |
st.header("Chat with André's research 💬 📚")
|
| 10 |
|
| 11 |
if "messages" not in st.session_state.keys(): # Initialize the chat message history
|
| 12 |
st.session_state.messages = [{"role": "assistant", "content": "Ask me a question about André's research!"}]
|
| 13 |
|
| 14 |
-
# Load config values
|
| 15 |
-
with open(r"config.json") as config_file:
|
| 16 |
-
config_details = json.load(config_file)
|
| 17 |
-
|
| 18 |
-
|
| 19 |
def main():
|
| 20 |
# setup dataset
|
| 21 |
download_test_data()
|
| 22 |
-
index = load_data(
|
| 23 |
chat_engine = index.as_chat_engine(chat_mode="condense_question", verbose=True)
|
| 24 |
|
| 25 |
if prompt := st.chat_input("Your question"): # Prompt for user input and save to chat history
|
|
|
|
| 1 |
+
import os
|
| 2 |
|
| 3 |
import streamlit as st
|
| 4 |
|
| 5 |
from chatbot.utils import download_test_data
|
| 6 |
from chatbot.utils import load_data
|
| 7 |
|
| 8 |
+
# add OpenAI API key to environemntal variables
|
| 9 |
+
os.environ["OPENAI_API_KEY"] = st.secrets["OPENAI_API_KEY"]
|
| 10 |
+
|
| 11 |
# Initialize message history
|
| 12 |
st.header("Chat with André's research 💬 📚")
|
| 13 |
|
| 14 |
if "messages" not in st.session_state.keys(): # Initialize the chat message history
|
| 15 |
st.session_state.messages = [{"role": "assistant", "content": "Ask me a question about André's research!"}]
|
| 16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
def main():
|
| 18 |
# setup dataset
|
| 19 |
download_test_data()
|
| 20 |
+
index = load_data()
|
| 21 |
chat_engine = index.as_chat_engine(chat_mode="condense_question", verbose=True)
|
| 22 |
|
| 23 |
if prompt := st.chat_input("Your question"): # Prompt for user input and save to chat history
|
chatbot/utils.py
CHANGED
|
@@ -14,22 +14,21 @@ from llama_index.llms import AzureOpenAI
|
|
| 14 |
def download_test_data():
|
| 15 |
# url = f"https://drive.google.com/drive/folders/uc?export=download&confirm=pbef&id={file_id}"
|
| 16 |
url = "https://drive.google.com/drive/folders/1uDSAWtLvp1YPzfXUsK_v6DeWta16pq6y"
|
| 17 |
-
with st.spinner(text="Downloading test data.
|
| 18 |
download_folder(url=url, quiet=False, use_cookies=False, output="./data/")
|
| 19 |
|
| 20 |
-
|
| 21 |
@st.cache_resource(show_spinner=False)
|
| 22 |
-
def load_data(
|
| 23 |
with st.spinner(text="Loading and indexing the provided dataset – hang tight! This may take a few seconds."):
|
| 24 |
documents = SimpleDirectoryReader(input_dir="./data", recursive=True).load_data()
|
| 25 |
llm = AzureOpenAI(
|
| 26 |
model="gpt-3.5-turbo",
|
| 27 |
-
engine=
|
| 28 |
temperature=0.5,
|
| 29 |
-
api_key=os.
|
| 30 |
-
api_base=
|
| 31 |
api_type="azure",
|
| 32 |
-
api_version=
|
| 33 |
system_prompt="You are an expert on André's research and your job is to answer"
|
| 34 |
"technical questions. Assume that all questions are related to"
|
| 35 |
"André's research. Keep your answers technical and based on facts"
|
|
@@ -38,11 +37,11 @@ def load_data(config_details):
|
|
| 38 |
# You need to deploy your own embedding model as well as your own chat completion model
|
| 39 |
embed_model = OpenAIEmbedding(
|
| 40 |
model="text-embedding-ada-002",
|
| 41 |
-
deployment_name=
|
| 42 |
-
api_key=os.
|
| 43 |
-
api_base=
|
| 44 |
api_type="azure",
|
| 45 |
-
api_version=
|
| 46 |
)
|
| 47 |
service_context = ServiceContext.from_defaults(llm=llm, embed_model=embed_model)
|
| 48 |
set_global_service_context(service_context)
|
|
|
|
| 14 |
def download_test_data():
|
| 15 |
# url = f"https://drive.google.com/drive/folders/uc?export=download&confirm=pbef&id={file_id}"
|
| 16 |
url = "https://drive.google.com/drive/folders/1uDSAWtLvp1YPzfXUsK_v6DeWta16pq6y"
|
| 17 |
+
with st.spinner(text="Downloading test data. This might take a minute."):
|
| 18 |
download_folder(url=url, quiet=False, use_cookies=False, output="./data/")
|
| 19 |
|
|
|
|
| 20 |
@st.cache_resource(show_spinner=False)
|
| 21 |
+
def load_data():
|
| 22 |
with st.spinner(text="Loading and indexing the provided dataset – hang tight! This may take a few seconds."):
|
| 23 |
documents = SimpleDirectoryReader(input_dir="./data", recursive=True).load_data()
|
| 24 |
llm = AzureOpenAI(
|
| 25 |
model="gpt-3.5-turbo",
|
| 26 |
+
engine=st.secrets["ENGINE"],
|
| 27 |
temperature=0.5,
|
| 28 |
+
api_key=os.environ["OPENAI_API_KEY"],
|
| 29 |
+
api_base=st.secrets["OPENAI_API_BASE"],
|
| 30 |
api_type="azure",
|
| 31 |
+
api_version=st.secrets["OPENAI_API_VERSION"],
|
| 32 |
system_prompt="You are an expert on André's research and your job is to answer"
|
| 33 |
"technical questions. Assume that all questions are related to"
|
| 34 |
"André's research. Keep your answers technical and based on facts"
|
|
|
|
| 37 |
# You need to deploy your own embedding model as well as your own chat completion model
|
| 38 |
embed_model = OpenAIEmbedding(
|
| 39 |
model="text-embedding-ada-002",
|
| 40 |
+
deployment_name=st.secrets["ENGINE_EMBEDDING"],
|
| 41 |
+
api_key=os.environ["OPENAI_API_KEY"],
|
| 42 |
+
api_base=st.secrets["OPENAI_API_BASE"],
|
| 43 |
api_type="azure",
|
| 44 |
+
api_version=st.secrets["OPENAI_API_VERSION"],
|
| 45 |
)
|
| 46 |
service_context = ServiceContext.from_defaults(llm=llm, embed_model=embed_model)
|
| 47 |
set_global_service_context(service_context)
|