app.py

import gradio as gr
import requests
from bs4 import BeautifulSoup
from gradio_client import Client
from urllib.parse import urljoin
import pandas as pd
from io import StringIO
import json
import groq
import os


google_api_key = os.getenv('google_search')
API_URL = "https://blavken-flowiseblav.hf.space/api/v1/prediction/fbc118dc-ec00-4b59-acff-600648958be3"
api_key = os.getenv('groq')
client = groq.Client(api_key=api_key)

custom_css = """
#md {
    height: 200px;  
    font-size: 30px;
    background: #121212;
    padding: 20px;
    color: white;
    border: 1 px solid white;
    font-size:10px;
}
"""

def perplexica_search(payloads):
    client = Client("mgokg/PerplexicaApi")
    result = client.predict(
        prompt=f"{payloads}",
        optimization_mode="balanced",
        api_name="/question"
    )   
    return result

def query(payload):
    response = requests.post(API_URL, json=payload)
    return response.json()

def google_search(payloads):
    output = query({
        "question": f"{payloads}",
    })
    #search_query = f"{payloads} antworte kurz und knapp. antworte auf deutsch. du findest die antwort hier:\n {output}"
    texte=""
    for o in output:
        texte +=o
    return output

scheme = """
{"name":"","email":"","website":""}
"""

def llama(messages):
    client = Client("mgokg/selenium-screenshot-gradio")
    result = client.predict(
		message=f"{messages}",
		api_name="/predict"
    )
    return result
    
    client = Client("AiActivity/AI-Assistant")
    result = client.predict(
		message={"text":f"instruction: return a valid json object only, no comments or explanaition, fill in the missing information. use this json scheme.\n {scheme}\n leave blank if information is not verfügbar. here is the information for the values:\n{message}","files":[]},
		api_name="/chat"
    )
    print(result)

def llm(message):
    message = f'return a json object with the keys: name,email,phone,website \n the values can be found here, leave blank if value is not available:\n {message} \n return a json object only. no text, no explanaition'
    try:        
        completion = client.chat.completions.create(
            model="llama3-70b-8192",
            messages=[
                {"role": "system", "content": "You are a helpful assistant."},
                {"role": "user", "content": f"{message}"}
            ],
        )       
        return completion.choices[0].message.content
    except Exception as e:
        return f"Error in response generation: {str(e)}"

def qwen(jsondata):
    client = Client("Qwen/Qwen2.5-72B-Instruct")
    result = client.predict(
		query= f'return a json object with the keys: name,email,phone,website for each verein \n the values can be found here, leave blank if value is not available:\n {jsondata} \n return a json object only. no text, no explanaition',
		history=[],
		system="You are Qwen, created by Alibaba Cloud. You are a helpful assistant.",
		api_name="/model_chat"
    )
    return result

def list_of_clubs(ort):
    base_url = "https://vereine-in-deutschland.net"
    all_links_text = []
    initial_url = f"{base_url}/vereine/Bayern/{ort}"

    try:
        response = requests.get(initial_url)
        response.raise_for_status()
        soup = BeautifulSoup(response.content, 'html.parser')
        
        # Determine the last page
        link_element = soup.select_one('li.page-item:nth-child(8) > a:nth-child(1)')
        last_page = 10
        if link_element and 'href' in link_element.attrs:
            href = link_element['href']
            last_page = int(href.split('/')[-1])

        # Loop through all pages and collect links
        for page_number in range(1, last_page + 1):
            page_url = f"{base_url}/vereine/Bayern/{ort}/p/{page_number}"
            response = requests.get(page_url)
            response.raise_for_status()
            soup = BeautifulSoup(response.content, 'html.parser')
            target_div = soup.select_one('div.row-cols-1:nth-child(4)')

            if target_div:
                texts = [a.text for a in target_div.find_all('a', href=True)]
                all_links_text.extend(texts)
            else:
                print(f"Target div not found on page {page_number}")

    except Exception as e:
        return str(e), []

    all_links_text = all_links_text[0::2]
    return all_links_text

def process_ort(ort):
    #links_text = list_of_clubs(ort)
    #return links_text
    linkstext =   [
    "Bürgergemeinschaft Uttenreuth",
    "Bürgerinitiative Umweltverträgliche Mobilität im Schwabachtal e. V.",
    "Die Unabhängigen Uttenreuth",
    "Förderkreis der Kindergärten Uttenreuth e. V.",
    "Förderkreis Spielplatz Weiher",
    "Jagdgenossenschaft Uttenreuth",
    "Kirchweihburschen Uttenreuth",
    "Kleintierzuchtverein Uttenreuth",
    "Literaturkreis Uttenreuth",
    "Musikalisches Theater Uttenreuth",
    "Pfadfinderschaft VCP - Stamm J. F. Esper",
    "Posaunenchor Uttenreuth",
    "Sportclub SC Uttenreuth",
    "Sudetendeutsche Landsmannschaft",
    "Voltigierverein Gut Eggenhof Erlangen e. V."
]

    vereine = []
    
    for verein in linkstext:       
        prompt=f"{verein}",
        result = llama(prompt)
        vereine.append(result)
        print(result)
    #data = json.loads(vereine)
    #df = pd.DataFrame(vereine)
    return vereine
        
    for verein in links_text:
        client = Client("mgokg/gemini-2.0-flash-exp")
        result = client.predict(
    		prompt=f"impressum {verein}",
    		api_name="/perform_search"
        )
        #json_object = llm(result)
        """
        headers = {
            "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
        }    
        url = f"https://www.google.com/search?q=impressum {verein}"
        response = requests.get(url, headers=headers)
        soup = BeautifulSoup(response.content, 'html.parser')
        impressum_div = soup.find('body')
        contact_detailes = impressum_div.text
        json_object = llm(contact_detailes)
        """
        vereine.append(result)
    #dicts = [json.loads(item) for item in vereine] 
    #df = pd.DataFrame(dicts)
    #return df
    return vereine

# Create the Gradio interface
with gr.Blocks(css=custom_css) as demo:
    with gr.Row():
        #details_output = gr.DataFrame(label="Ausgabe", elem_id="md")        
        details_output = gr.Textbox(label="Ausgabe")  
    with gr.Row():
        ort_input = gr.Textbox(label="Ort eingeben", placeholder="ask anything...")      
    with gr.Row():         
        button = gr.Button("Senden")    

    # Connect the button to the function
    button.click(fn=process_ort, inputs=ort_input, outputs=details_output)   

# Launch the Gradio application
demo.launch()