update

app.py

@@ -1,3 +1,12 @@
+import streamlit as st
+
+st.set_page_config(
+    page_title="Streamlit demos",
+)
+
+st.sidebar.success("Select a demo above.")
+
+
 import streamlit as st
 import leafmap.maplibregl as leafmap
 import ibis

chat.py → pages/chat.py

@@ -7,7 +7,12 @@ with st.sidebar:
     model = st.radio("Select an LLM:", ['llava', 'gemma2', 'phi3', 'llama3', 'embed-mistral', 'mixtral', 'gorilla', 'groq-tools'])
     st.session_state["model"] =  model
 
-
+## dockerized streamlit app wants to read from os.getenv(), otherwise use st.secrets
+import os
+api_key = os.getenv("LITELLM_KEY")
+if api_key is None:
+    api_key = st.secrets["LITELLM_KEY"]
+        
 
 if "messages" not in st.session_state:
     st.session_state.messages = []
@@ -17,7 +22,7 @@ for message in st.session_state.messages:
         st.markdown(message["content"])
 
 client = OpenAI(
-    api_key = st.secrets["LITELLM_KEY"], 
+    api_key = api_key, 
     base_url = "https://llm.nrp-nautilus.io"
 )
 

rag.py → pages/rag.py

@@ -1,6 +1,13 @@
 import streamlit as st
 from langchain_community.document_loaders import PyPDFLoader
 
+## dockerized streamlit app wants to read from os.getenv(), otherwise use st.secrets
+import os
+api_key = os.getenv("LITELLM_KEY")
+if api_key is None:
+    api_key = st.secrets["LITELLM_KEY"]
+        
+
 st.title("RAG Demo")
 
 
@@ -13,22 +20,34 @@ Once the document has been uploaded and parsed, ask your questions in the chat d
 # st.sidebar.file_uploader("Choose a PDF file", type=["pdf"])
 url = st.text_input("PDF URL", "https://www.resources.ca.gov/-/media/CNRA-Website/Files/2024_30x30_Pathways_Progress_Report.pdf")
 
+# +
+import bs4
+from langchain import hub
+from langchain_chroma import Chroma
+from langchain_community.document_loaders import WebBaseLoader
+from langchain_core.output_parsers import StrOutputParser
+from langchain_core.runnables import RunnablePassthrough
+from langchain_openai import OpenAIEmbeddings
+from langchain_text_splitters import RecursiveCharacterTextSplitter
+
+
 @st.cache_data
-def doc_loader(url):
+def pdf_loader(url):
     loader = PyPDFLoader(url)
     return loader.load()
-    
-docs = doc_loader(url)
+
+docs = pdf_loader(url)
+
 
 # Set up the language model
 from langchain_openai import ChatOpenAI
-llm = ChatOpenAI(model = "llama3", api_key=st.secrets["LITELLM_KEY"], base_url = "https://llm.nrp-nautilus.io",  temperature=0)
+llm = ChatOpenAI(model = "llama3", api_key = api_key, base_url = "https://llm.nrp-nautilus.io",  temperature=0)
 
 # Set up the embedding model
 from langchain_openai import OpenAIEmbeddings
 embedding = OpenAIEmbeddings(
     model = "embed-mistral", 
-    api_key=st.secrets["LITELLM_KEY"], 
+    api_key = api_key, 
     base_url = "https://llm.nrp-nautilus.io"
 )
 
@@ -61,19 +80,6 @@ prompt = ChatPromptTemplate.from_messages(
 question_answer_chain = create_stuff_documents_chain(llm, prompt)
 rag_chain = create_retrieval_chain(retriever, question_answer_chain)
 
-# +
-# agent is ready to test:
-
-#results = rag_chain.invoke({"input": "What is the goal of CA 30x30?"})
-#results['answer']
-#results['context'][0].page_content
-#results['context'][0].metadata
-
-# -
-
-#results['context'][0].page_content
-#results['context'][0].metadata
-
 
 # Place agent inside a streamlit application:
 

sql.py → pages/sql.py

@@ -2,6 +2,13 @@ import streamlit as st
 
 st.title("SQL demo")
 
+## dockerized streamlit app wants to read from os.getenv(), otherwise use st.secrets
+import os
+api_key = os.getenv("LITELLM_KEY")
+if api_key is None:
+    api_key = st.secrets["LITELLM_KEY"]
+        
+
 parquet = st.text_input("parquet file:", "https://espm-157-f24.github.io/spatial-carl-amanda-tyler/new_haven_stats.parquet")
 
 # create sharable low-level connection, see: https://github.com/Mause/duckdb_engine
@@ -43,7 +50,7 @@ prompt = PromptTemplate.from_template(template, partial_variables = {"dialect":
 from langchain_openai import ChatOpenAI
 llm = ChatOpenAI(model="gorilla", # Try: llama3, gorilla, or groq-tools, or other models
                  temperature=0, 
-                 api_key=st.secrets["LITELLM_KEY"], 
+                 api_key=api_key, 
                  base_url = "https://llm.nrp-nautilus.io")