Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,4 +1,6 @@
|
|
|
|
|
| 1 |
import os, re, validators, streamlit as st
|
|
|
|
| 2 |
from youtube_transcript_api import YouTubeTranscriptApi, TranscriptsDisabled, VideoUnavailable
|
| 3 |
from langchain.prompts import PromptTemplate
|
| 4 |
from langchain.chains.summarize import load_summarize_chain
|
|
@@ -7,22 +9,19 @@ from langchain.schema import Document
|
|
| 7 |
from langchain_community.document_loaders import UnstructuredURLLoader
|
| 8 |
from langchain.document_loaders import PyPDFLoader
|
| 9 |
|
| 10 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
st.set_page_config(page_title="LangChain Summarizer", page_icon="π¦")
|
| 12 |
st.title("π¦ LangChain: Summarize YT / Webpage / PDF")
|
| 13 |
|
| 14 |
-
#
|
| 15 |
-
with st.sidebar:
|
| 16 |
-
st.header("API keys")
|
| 17 |
-
groq_api_key = st.text_input("Groq API Key", type="password")
|
| 18 |
-
if groq_api_key:
|
| 19 |
-
os.environ["GROQ_API_KEY"] = groq_api_key # for libraries
|
| 20 |
-
|
| 21 |
-
# βββββββββββββββββββββ PLACEHOLDERS / FILE & URL INPUT βββββββββββββββ
|
| 22 |
generic_url = st.text_input("Paste a YouTube / web URL here:")
|
| 23 |
uploaded_file = st.file_uploader("β¦or upload a PDF", type=["pdf"])
|
| 24 |
|
| 25 |
-
# ββββββββββββββββββββββββββ UTILITY FUNCTIONS
|
| 26 |
def get_video_id(url: str) -> str | None:
|
| 27 |
m = re.search(r"(?:v=|\/)([0-9A-Za-z_-]{11})", url)
|
| 28 |
return m.group(1) if m else None
|
|
@@ -35,9 +34,14 @@ SUMMARY_PROMPT = PromptTemplate(
|
|
| 35 |
def build_llm() -> ChatGroq:
|
| 36 |
"""Instantiate ChatGroq once and cache it in session_state."""
|
| 37 |
if "llm" not in st.session_state:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
st.session_state.llm = ChatGroq(
|
| 39 |
model="llama3-70b-8192",
|
| 40 |
-
groq_api_key=
|
| 41 |
)
|
| 42 |
return st.session_state.llm
|
| 43 |
|
|
@@ -46,10 +50,10 @@ def summarize(docs):
|
|
| 46 |
chain = load_summarize_chain(llm, chain_type="stuff", prompt=SUMMARY_PROMPT)
|
| 47 |
return chain({"input_documents": docs})["output_text"]
|
| 48 |
|
| 49 |
-
# βββββββββββββββββββββββββββββ MAIN ACTION
|
| 50 |
if st.button("Summarize"):
|
| 51 |
-
if not
|
| 52 |
-
st.error("
|
| 53 |
elif not generic_url and not uploaded_file:
|
| 54 |
st.error("Provide a URL or upload a PDF, then press Summarize.")
|
| 55 |
else:
|
|
|
|
| 1 |
+
# app.py β Streamlit Summarizer (dotenv version)
|
| 2 |
import os, re, validators, streamlit as st
|
| 3 |
+
from dotenv import load_dotenv # β NEW
|
| 4 |
from youtube_transcript_api import YouTubeTranscriptApi, TranscriptsDisabled, VideoUnavailable
|
| 5 |
from langchain.prompts import PromptTemplate
|
| 6 |
from langchain.chains.summarize import load_summarize_chain
|
|
|
|
| 9 |
from langchain_community.document_loaders import UnstructuredURLLoader
|
| 10 |
from langchain.document_loaders import PyPDFLoader
|
| 11 |
|
| 12 |
+
# βββββββββββββββ load variables from .env (runs before anything else)
|
| 13 |
+
load_dotenv() # looks for .env in project root
|
| 14 |
+
GROQ_KEY = os.getenv("GROQ_API_KEY") # expect GROQ_API_KEY=xxx in .env
|
| 15 |
+
|
| 16 |
+
# ββββββββββββββββββββββββββ STREAMLIT CONFIG ββββββββββββββββββββββββ
|
| 17 |
st.set_page_config(page_title="LangChain Summarizer", page_icon="π¦")
|
| 18 |
st.title("π¦ LangChain: Summarize YT / Webpage / PDF")
|
| 19 |
|
| 20 |
+
# βββββββββββββββββββββ PLACEHOLDERS / FILE & URL INPUT ββββββββββββββ
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
generic_url = st.text_input("Paste a YouTube / web URL here:")
|
| 22 |
uploaded_file = st.file_uploader("β¦or upload a PDF", type=["pdf"])
|
| 23 |
|
| 24 |
+
# ββββββββββββββββββββββββββ UTILITY FUNCTIONS βββββββββββββββββββββββ
|
| 25 |
def get_video_id(url: str) -> str | None:
|
| 26 |
m = re.search(r"(?:v=|\/)([0-9A-Za-z_-]{11})", url)
|
| 27 |
return m.group(1) if m else None
|
|
|
|
| 34 |
def build_llm() -> ChatGroq:
|
| 35 |
"""Instantiate ChatGroq once and cache it in session_state."""
|
| 36 |
if "llm" not in st.session_state:
|
| 37 |
+
if not GROQ_KEY:
|
| 38 |
+
raise RuntimeError(
|
| 39 |
+
"βΒ Groq API key not found. "
|
| 40 |
+
"Add GROQ_API_KEY to a `.env` file or the environment."
|
| 41 |
+
)
|
| 42 |
st.session_state.llm = ChatGroq(
|
| 43 |
model="llama3-70b-8192",
|
| 44 |
+
groq_api_key=GROQ_KEY,
|
| 45 |
)
|
| 46 |
return st.session_state.llm
|
| 47 |
|
|
|
|
| 50 |
chain = load_summarize_chain(llm, chain_type="stuff", prompt=SUMMARY_PROMPT)
|
| 51 |
return chain({"input_documents": docs})["output_text"]
|
| 52 |
|
| 53 |
+
# βββββββββββββββββββββββββββββ MAIN ACTION ββββββββββββββββββββββββββ
|
| 54 |
if st.button("Summarize"):
|
| 55 |
+
if not GROQ_KEY:
|
| 56 |
+
st.error("Groq key missing. Set **GROQ_API_KEY** in your `.env`.")
|
| 57 |
elif not generic_url and not uploaded_file:
|
| 58 |
st.error("Provide a URL or upload a PDF, then press Summarize.")
|
| 59 |
else:
|