Spaces:
Sleeping
Sleeping
V1.0
Browse files- app.py +94 -0
- requirements.txt +4 -0
app.py
ADDED
|
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import openai
|
| 3 |
+
import streamlit as st
|
| 4 |
+
import trafilatura
|
| 5 |
+
from langchain import PromptTemplate, OpenAI, LLMChain
|
| 6 |
+
|
| 7 |
+
# Definition of the style
|
| 8 |
+
custom_css = """
|
| 9 |
+
<style>
|
| 10 |
+
.title-text {
|
| 11 |
+
color: #0077B6; /* Cambia el color del subtítulo a un azul más claro */
|
| 12 |
+
}
|
| 13 |
+
</style>
|
| 14 |
+
"""
|
| 15 |
+
|
| 16 |
+
st.sidebar.write("This program uses the OpenAI API to function. Please enter your API key below to use the program. ")
|
| 17 |
+
OPENAI_API_KEY = st.sidebar.text_input("API KEY:")
|
| 18 |
+
|
| 19 |
+
class WebContentSummarizerApp:
|
| 20 |
+
def __init__(self):
|
| 21 |
+
st.title("WebContent Summarizer")
|
| 22 |
+
# Insert the custom CSS into the application
|
| 23 |
+
st.markdown(custom_css, unsafe_allow_html=True)
|
| 24 |
+
# Use a CSS class for the subtitle
|
| 25 |
+
st.markdown('<p class="title-text">With just a simple URL, simplify your online research and content consumption. '
|
| 26 |
+
'This AI tool extracts and condenses the essential information from web pages into a concise, easy-to-read paragraph.</p>', unsafe_allow_html=True)
|
| 27 |
+
self.url = st.text_input("Paste or type the URL you want to summarize:")
|
| 28 |
+
self.run_button = st.button("Summarize")
|
| 29 |
+
|
| 30 |
+
@staticmethod
|
| 31 |
+
def scrape_website(url):
|
| 32 |
+
downloaded = trafilatura.fetch_url(url)
|
| 33 |
+
main_text = trafilatura.extract(downloaded)
|
| 34 |
+
return main_text
|
| 35 |
+
|
| 36 |
+
def run(self):
|
| 37 |
+
if self.run_button:
|
| 38 |
+
if self.url:
|
| 39 |
+
url = self.url
|
| 40 |
+
with st.spinner("Getting information from the page..."):
|
| 41 |
+
self.text = self.scrape_website(url)
|
| 42 |
+
with st.spinner("Analyzing and organizing..."):
|
| 43 |
+
result = self.llm_text_insight()
|
| 44 |
+
st.write("Summarized content:")
|
| 45 |
+
st.write(result)
|
| 46 |
+
else:
|
| 47 |
+
st.warning("Please enter a valid URL and your OpenAI API Key.")
|
| 48 |
+
|
| 49 |
+
def llm_text_insight(self):
|
| 50 |
+
openai.api_key = OPENAI_API_KEY
|
| 51 |
+
text = self.text
|
| 52 |
+
|
| 53 |
+
# Split the text into smaller fragments
|
| 54 |
+
max_length = 4096
|
| 55 |
+
chunks = [text[i:i+max_length] for i in range(0, len(text), max_length)]
|
| 56 |
+
|
| 57 |
+
results = []
|
| 58 |
+
|
| 59 |
+
template = """
|
| 60 |
+
Please take the following text interprete and transform it into an organized paragraph for a study.
|
| 61 |
+
Ensure that the paragraph is clear and coherent. If necessary, add or rearrange ideas to achieve a logical structure.
|
| 62 |
+
The text is as follows:
|
| 63 |
+
|
| 64 |
+
text: {input}
|
| 65 |
+
"""
|
| 66 |
+
|
| 67 |
+
llm = OpenAI(temperature=0, openai_api_key=OPENAI_API_KEY)
|
| 68 |
+
prompt_template = PromptTemplate.from_template(template=template)
|
| 69 |
+
chain = LLMChain(llm=llm, prompt=prompt_template)
|
| 70 |
+
|
| 71 |
+
for chunk in chunks:
|
| 72 |
+
result = chain.predict(input=chunk)
|
| 73 |
+
results.append(result)
|
| 74 |
+
|
| 75 |
+
# Check if results contain more than one paragraph and final result
|
| 76 |
+
if len(results) > 1:
|
| 77 |
+
alternate_input = " ".join(results)
|
| 78 |
+
|
| 79 |
+
llm = OpenAI(temperature=0, openai_api_key=OPENAI_API_KEY)
|
| 80 |
+
prompt_template = PromptTemplate.from_template(template=template)
|
| 81 |
+
chain = LLMChain(llm=llm, prompt=prompt_template)
|
| 82 |
+
result = chain.predict(input=alternate_input)
|
| 83 |
+
|
| 84 |
+
return result
|
| 85 |
+
|
| 86 |
+
|
| 87 |
+
else:
|
| 88 |
+
# Concatenation of the results
|
| 89 |
+
return " ".join(results)
|
| 90 |
+
|
| 91 |
+
|
| 92 |
+
if __name__ == "__main__":
|
| 93 |
+
app = WebContentSummarizerApp()
|
| 94 |
+
app.run()
|
requirements.txt
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
openai
|
| 2 |
+
streamlit
|
| 3 |
+
trafilatura
|
| 4 |
+
langchain
|