Update app.py

app.py

@@ -1,146 +1,184 @@
-import gradio as gr
 import os
-import spaces
-from transformers import GemmaTokenizer, AutoModelForCausalLM
-from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
-from threading import Thread
-
-# Set an environment variable
-HF_TOKEN = os.environ.get("HF_TOKEN", None)
-
-
-DESCRIPTION = '''
-<div>
-<h1 style="text-align: center;">Meta Llama3 8B</h1>
-<p>This Space demonstrates the instruction-tuned model <a href="https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct"><b>Meta Llama3 8b Chat</b></a>. Meta Llama3 is the new open LLM and comes in two sizes: 8b and 70b. Feel free to play with it, or duplicate to run privately!</p>
-<p>🔎 For more details about the Llama3 release and how to use the model with <code>transformers</code>, take a look <a href="https://huggingface.co/blog/llama3">at our blog post</a>.</p>
-<p>🦕 Looking for an even more powerful model? Check out the <a href="https://huggingface.co/chat/"><b>Hugging Chat</b></a> integration for Meta Llama 3 70b</p>
-</div>
-'''
-
-LICENSE = """
-<p/>
-
----
-Built with Meta Llama 3
-"""
-
-PLACEHOLDER = """
-<div style="padding: 30px; text-align: center; display: flex; flex-direction: column; align-items: center;">
-   <img src="https://ysharma-dummy-chat-app.hf.space/file=/tmp/gradio/8e75e61cc9bab22b7ce3dec85ab0e6db1da5d107/Meta_lockup_positive%20primary_RGB.jpg" style="width: 80%; max-width: 550px; height: auto; opacity: 0.55;  "> 
-   <h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">Meta llama3</h1>
-   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Ask me anything...</p>
-</div>
-"""
-
-
-css = """
-h1 {
-  text-align: center;
-  display: block;
-}
-
-#duplicate-button {
-  margin: auto;
-  color: white;
-  background: #1565c0;
-  border-radius: 100vh;
-}
-"""
-
-# Load the tokenizer and model
-tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct")
-model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct", device_map="auto")  # to("cuda:0") 
-terminators = [
-    tokenizer.eos_token_id,
-    tokenizer.convert_tokens_to_ids("<|eot_id|>")
-]
+import gradio as gr
+import requests
+from crewai import Agent, Task, Crew, Process
+
+from langchain_openai import ChatOpenAI
+
+from langchain_community.tools import DuckDuckGoSearchRun, DuckDuckGoSearchResults
+from crewai_tools import tool, SeleniumScrapingTool, ScrapeWebsiteTool
+from duckduckgo_search import DDGS
+
+from newspaper import Article
+
+# Ensure essential environment variables are set
+openai_api_key = os.getenv('OPENAI_API_KEY')
+if not openai_api_key:
+    raise EnvironmentError("OPENAI_API_KEY is not set in environment variables")
+
+def fetch_content(url):
+    try:
+        article = Article(url)
+        article.download()
+        article.parse()
+        return article.text
+    except Exception as e:
+        print("ERROR: " + str(e))
+        return f"Error fetching content: {e}"
+
+# Define the DuckDuckGoSearch tool
+@tool('DuckDuckGoSearchResults')
+def search_results(search_query: str) -> dict:
+    """
+    Performs a web search to gather and return a collection of search results.
+    This tool automates the retrieval of web-based information related to a specified query.
+    Args:
+    - search_query (str): The query string that specifies the information to be searched on the web. This should be a clear and concise expression of the user's information needs.
+    Returns:
+    - list: A list of dictionaries, where each dictionary represents a search result. Each dictionary includes 'snippet' of the page and the 'link' with the url linking to it.
+    """
+    results = DDGS().text(search_query, max_results=5, timelimit='m')
+    results_list = [{"title": result['title'], "snippet": result['body'], "link": result['href']} for result in results]
+    return results_list
 
-@spaces.GPU(duration=120)
-def chat_llama3_8b(message: str, 
-              history: list, 
-              temperature: float, 
-              max_new_tokens: int
-             ) -> str:
+@tool('WebScrapper')
+def web_scrapper(url: str, topic: str) -> str:
     """
-    Generate a streaming response using the llama3-8b model.
+    A tool designed to extract and read the content of a specified link and generate a summary on a specific topic.
+    It is capable of handling various types of web pages by making HTTP requests and parsing the received HTML content.
+    This tool is particularly useful for web scraping tasks, data collection, or extracting specific information from websites.
+    
     Args:
-        message (str): The input message.
-        history (list): The conversation history used by ChatInterface.
-        temperature (float): The temperature for generating the response.
-        max_new_tokens (int): The maximum number of new tokens to generate.
+    - url (str): The URL from which to scrape content.
+    - topic (str): The specific topic on which to generate a summary.
     Returns:
-        str: The generated response.
+    - summary (str): summary of the url on the topic
     """
-    conversation = []
-    for user, assistant in history:
-        conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
-    conversation.append({"role": "user", "content": message})
+    # Scrape content from the specified URL
+    content = fetch_content(url)
+    
+    # Prepare the prompt for generating the summary
+    prompt = f"Generate a summary of the following content on the topic ## {topic} ### \n\nCONTENT:\n\n" + content
+    
+    # Generate the summary using OpenAI
+    openai_llm = ChatOpenAI(temperature=0.4, model_name="gpt-3.5-turbo")
+    response = openai_llm.invoke(prompt)
 
-    input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt").to(model.device)
+    summary_response = f"""###
+    Summary:
+    {response.content}
     
-    streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
+    URL: {url}
+    ###
+    """
+    
+    return summary_response
 
-    generate_kwargs = dict(
-        input_ids= input_ids,
-        streamer=streamer,
-        max_new_tokens=max_new_tokens,
-        do_sample=True,
-        temperature=temperature,
-        eos_token_id=terminators,
-    )
-    # This will enforce greedy generation (do_sample=False) when the temperature is passed 0, avoiding the crash.             
-    if temperature == 0:
-        generate_kwargs['do_sample'] = False
-        
-    t = Thread(target=model.generate, kwargs=generate_kwargs)
-    t.start()
+def kickoff_crew(topic: str, model_choice: str) -> str:
+    try:
+        # Initialize the OpenAI language model
+        openai_llm = ChatOpenAI(temperature=0, model_name=model_choice)
+    
+        # Define Agents with OpenAI LLM
+        researcher = Agent(
+            role='Researcher',
+            goal=f'Search and Collect detailed information on topic ## {topic} ##',
+            tools=[search_results, web_scrapper],
+            llm=openai_llm,
+            backstory=(
+                "You are a meticulous researcher, skilled at navigating vast amounts of information to extract essential insights on any given topic. "
+                "Your dedication to detail ensures the reliability and thoroughness of your findings. "
+                "With a strategic approach, you carefully analyze and document data, aiming to provide accurate and trustworthy results."
+            ),
+            allow_delegation=False,
+            max_iter=15,
+            max_rpm=20,
+            memory=True,
+            verbose=True
+        )
 
-    outputs = []
-    for text in streamer:
-        outputs.append(text)
-        #print(outputs)
-        yield "".join(outputs)
         
+        editor = Agent(
+            role='Editor',
+            goal=f'Compile and refine the information into a comprehensive report on topic ## {topic} ##',
+            llm=openai_llm,
+            backstory=(
+                "As an expert editor, you specialize in transforming raw data into clear, engaging reports. "
+                "Your strong command of language and attention to detail ensure that each report not only conveys essential insights "
+                "but is also easily understandable and appealing to diverse audiences. "
+            ),
+            allow_delegation=False,
+            max_iter=5,
+            max_rpm=15,
+            memory=True,
+            verbose=True
+        )
+        
+        # Define Tasks
+        research_task = Task(
+            description=(
+                f"Use the DuckDuckGoSearchResults tool to collect initial search snippets on ## {topic} ##. "
+                f"If more detailed searches are required, generate and execute new queries related to ## {topic} ##. "
+                "Subsequently, employ the WebScrapper tool to delve deeper into significant URLs identified from the snippets, extracting further information and insights. "
+                "Compile these findings into a preliminary draft, documenting all relevant sources, titles, and links associated with the topic. "
+                "Ensure high accuracy throughout the process and avoid any fabrication or misrepresentation of information."
+            ),
+            expected_output=(
+                "A structured draft report about the topic, featuring an introduction, a detailed main body organized by different aspects of the topic, and a conclusion. "
+                "Each section should properly cite sources, providing a thorough overview of the information gathered."
+            ),
+            agent=researcher
+        )
 
-# Gradio block
-chatbot=gr.Chatbot(height=450, placeholder=PLACEHOLDER, label='Gradio ChatInterface')
-
-with gr.Blocks(fill_height=True, css=css) as demo:
-    
-    gr.Markdown(DESCRIPTION)
-    gr.DuplicateButton(value="Duplicate Space for private use", elem_id="duplicate-button")
-    gr.ChatInterface(
-        fn=chat_llama3_8b,
-        chatbot=chatbot,
-        fill_height=True,
-        additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=False, render=False),
-        additional_inputs=[
-            gr.Slider(minimum=0,
-                      maximum=1, 
-                      step=0.1,
-                      value=0.95, 
-                      label="Temperature", 
-                      render=False),
-            gr.Slider(minimum=128, 
-                      maximum=4096,
-                      step=1,
-                      value=512, 
-                      label="Max new tokens", 
-                      render=False ),
-            ],
-        examples=[
-            ['How to setup a human base on Mars? Give short answer.'],
-            ['Explain theory of relativity to me like I’m 8 years old.'],
-            ['What is 9,000 * 9,000?'],
-            ['Write a pun-filled happy birthday message to my friend Alex.'],
-            ['Justify why a penguin might make a good king of the jungle.']
-            ],
-        cache_examples=False,
-                     )
+        
+        edit_task = Task(
+            description=(
+                "Review and refine the initial draft report from the research task. Organize the content logically to enhance information flow. "
+                "Verify the accuracy of all data, correct discrepancies, and update information to ensure it reflects current knowledge and is well-supported by sources. "
+                "Improve the report's readability by enhancing language clarity, adjusting sentence structures, and maintaining a consistent tone. "
+                "Include a section listing all sources used, formatted as bullet points following this template: "
+                "- title: url'."
+            ),
+            expected_output=(
+                f"A polished, comprehensive report on topic ## {topic} ##, with a clear, professional narrative that accurately reflects the research findings. "
+                "The report should include an introduction, an extensive discussion section, a concise conclusion, and a well-organized source list. "
+                "Ensure the document is grammatically correct and ready for publication or presentation."
+            ),
+            agent=editor,
+            context=[research_task]
+        )
     
-    gr.Markdown(LICENSE)
+        # Forming the Crew
+        crew = Crew(
+            agents=[researcher, editor],
+            tasks=[research_task, edit_task],
+            process=Process.sequential,
+        )
     
+        # Kick-off the research process
+        result = crew.kickoff()
+        if not isinstance(result, str):
+            result = str(result)
+        return result
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+def main():
+    """Set up the Gradio interface for the CrewAI Research Tool."""
+    with gr.Blocks() as demo:
+        gr.Markdown("## CrewAI Research Tool")
+        topic_input = gr.Textbox(label="Enter Topic", placeholder="Type here...")
+        model_choice = gr.Radio(choices=["gpt-3.5-turbo", "gpt-4"], label="Choose Model")
+        submit_button = gr.Button("Start Research")
+        output = gr.Markdown(label="Result")
+
+        submit_button.click(
+            fn=kickoff_crew,
+            inputs=[topic_input, model_choice],
+            outputs=output
+        )
+
+
 if __name__ == "__main__":
     demo.launch()