Update app.py

app.py

@@ -1,331 +1,67 @@
 import streamlit as st
-import requests
-from bs4 import BeautifulSoup
-from urllib.parse import urljoin
-import json
-import csv
-import pandas as pd
 import os
-from gradio_client import Client
+from groq import Groq
+import soundfile as sf
+from tempfile import NamedTemporaryFile
 
-#api_key = os.getenv('groq')
+# Load the API key from the environment variable
+api_key = os.getenv('groq_whisper')
 
-def parse_links_and_content(ort):
-    base_url = "https://vereine-in-deutschland.net"
-    all_links = []   
-    # Konstruiere die vollständige URL
-    initial_url = f"{base_url}/vereine/Bayern/{ort}/"
-    
+if api_key is None:
+    raise ValueError("groq_whisper environment variable is not set")
+
+# Initialize the Groq client
+client = Groq(api_key=api_key)
+
+def processaudio(audio_data):
     try:
-        # Senden der Anfrage an die initiale URL
-        response = requests.get(initial_url)
-        response.raise_for_status()  # Überprüfen, ob die Anfrage erfolgreich war
-        
-        # Parse the HTML content using BeautifulSoup
-        soup = BeautifulSoup(response.content, 'html.parser')
+        # Entpacken der Audiodaten (Sample-Rate und Numpy-Array)
+        sample_rate, samples = audio_data
         
-        # Ermittle die letzte Seite
-        link_element = soup.select_one('li.page-item:nth-child(8) > a:nth-child(1)')
-        
-        if link_element and 'href' in link_element.attrs:
-            href = link_element['href']
-            # Extrahiere die letzten beiden Zeichen der URL
-            last_two_chars = href[-2:]
+        # Temporäre Audiodatei erstellen
+        with NamedTemporaryFile(suffix=".wav", delete=True) as tmpfile:
+            # Audio als WAV-Datei speichern
+            sf.write(tmpfile.name, samples, sample_rate)
             
-            # Konvertiere die letzten beiden Zeichen in einen Integer
-            last_two_chars_int = int(last_two_chars)
-        else:
-            last_two_chars_int = 1  # Falls die letzte Seite nicht gefunden wird, nimm an, dass es nur eine Seite gibt
-
-        # Schleife durch alle Seiten und sammle Links
-        for page_number in range(1, last_two_chars_int + 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)')
+            # Datei erneut öffnen und an Groq senden
+            with open(tmpfile.name, "rb") as file:
+                transcription = client.audio.transcriptions.create(
+                    file=(os.path.basename(tmpfile.name), file.read()),
+                    model="whisper-large-v3-turbo",
+                    prompt="transcribe",
+                    language="de",
+                    response_format="json",
+                    temperature=0.0
+                )
+            return transcription.text
             
-            if target_div:
-                texts = [a.text for a in target_div.find_all('a', href=True)] 
-                all_links.extend(texts)
-            else:
-                st.write(f"Target div not found on page {page_number}")
-        
     except Exception as e:
-        return str(e), []
-        
-    all_links = all_links[0::2]
-    return all_links
-
-def scrape_links(links):
-    contact_details = []
-    client = Client("mgokg/PerplexicaApi")
-    for verein in links:   
-        result = client.predict(
-            prompt=f"{verein}",
-            api_name="/parse_links"
-        )
-        contact_details.append(result)
-        
-    return contact_details
-
-# Speichere die JSON-Daten in eine CSV-Datei 
-def save_to_csv(data, filename): 
-    keys = data[0].keys() 
-    with open(filename, 'w', newline='', encoding='utf-8') as output_file: 
-        dict_writer = csv.DictWriter(output_file, fieldnames=keys) 
-        dict_writer.writeheader() 
-        dict_writer.writerows(data)
-
-# Streamlit App
-st.title("Vereinsinformationen abrufen")
-
-ort_input = st.text_input("Ort", placeholder="Gib den Namen des Ortes ein")
-
-if st.button("Senden"):
-    links = parse_links_and_content(ort_input)
-    contact_details = scrape_links(links)
-    json_data = [json.loads(item) for item in contact_details]
-    
-    # Zeige die Ergebnisse an
-    st.json(json_data)
-    
-    # Speichere die Daten in einer CSV-Datei
-    save_to_csv(json_data, 'contact_details.csv')
-    
-    # Bereitstellung des Download-Links
-    with open('contact_details.csv', 'rb') as file:
-        st.download_button(
-            label="CSV-Datei herunterladen",
-            data=file,
-            file_name='contact_details.csv',
-            mime='text/csv'
-        )
-
-
-
-
-
-
-
-
-
-
-
-
-'''
-import streamlit as st
-#import sounddevice as sd
-import numpy as np
-import wavio
-import speech_recognition as sr
-
-
-st.title("Audio Recorder und Transkription")
-
-# Aufnahmeparameter
-duration = st.slider("Aufnahmedauer (Sekunden)", 1, 10, 5)
-fs = 44100  # Abtastrate
-
-if st.button("Aufnahme starten"):
-    st.write("Aufnahme läuft...")
-    #recording = sd.rec(int(duration * fs), samplerate=fs, channels=2)
-    #sd.wait()  # Aufnahme beenden
-
-    # Speichern der Aufnahme
-    wavio.write("aufnahme.wav", recording, fs, sampwidth=2)
-    st.write("Aufnahme abgeschlossen!")
-
-    # Transkription
-    recognizer = sr.Recognizer()
-    with sr.AudioFile("aufnahme.wav") as source:
-        audio_data = recognizer.record(source)
-        try:
-            text = recognizer.recognize_google(audio_data, language="de-DE")
-            st.write("Transkribierter Text:")
-            st.write(text)
-        except sr.UnknownValueError:
-            st.write("Audio konnte nicht erkannt werden.")
-        except sr.RequestError as e:
-            st.write(f"Fehler bei der Anfrage an den Google Speech Recognition Service: {e}")
-
-# Hinweis für Benutzer
-st.write("Klicke auf 'Aufnahme starten', um die Aufnahme zu beginnen.")
-
-
-
-
-
-
-
-import streamlit as st
-import pydub
-import speech_recognition as sr
-from io import BytesIO
-
-st.title("Audio Recorder und Transkription")
-
-# Audioaufnahme
-audio_file = st.file_uploader("Lade eine Audiodatei hoch", type=["wav", "mp3"])
-
-if audio_file is not None:
-    audio_bytes = audio_file.read()
-    audio = pydub.AudioSegment.from_file(BytesIO(audio_bytes))
+        return f"Ein Fehler ist aufgetreten: {str(e)}"
 
-    # Audio in das richtige Format konvertieren
-    audio = audio.set_frame_rate(16000).set_channels(1).set_sample_width(2)
-    audio_bytes = audio.raw_data
-
-    # Audio transkribieren
-    recognizer = sr.Recognizer()
-    audio_source = sr.AudioData(audio_bytes, frame_rate=16000, sample_width=2, channels=1)
+def process_audio(file_path):
     try:
-        text = recognizer.recognize_google(audio_source)
-        st.write("Transkribierter Text:")
-        st.write(text)
-    except sr.UnknownValueError:
-        st.write("Audio konnte nicht erkannt werden.")
-    except sr.RequestError as e:
-        st.write(f"Fehler bei der Anfrage an den Google Speech Recognition Service: {e}")
-
-# Hinweis für Benutzer
-st.write("Lade eine Audiodatei hoch, um sie zu transkribieren.")
-
-
-
-
-
-
-import streamlit as st
-from streamlit_webrtc import webrtc_streamer, AudioProcessorBase, WebRtcMode
-
-class AudioProcessor(AudioProcessorBase):
-    def recv(self, frame):
-        # Hier kannst du die Audioverarbeitung hinzufügen
-        return frame
-
-st.title("Audio Recorder")
-
-webrtc_ctx = webrtc_streamer(
-    key="audio",
-    mode=WebRtcMode.SENDRECV,
-    audio_processor_factory=AudioProcessor,
-    media_stream_constraints={"audio": True},
-    async_processing=True,
-)
-
-if webrtc_ctx.state.playing:
-    st.write("Recording...")
-else:
-    st.write("Click on Start to record audio.")
-
-
-
-import streamlit as st
-import os
-import time
-import pandas as pd
-from pandasai import SmartDatalake
-from pandasai import SmartDataframe
-from pandasai.responses.streamlit_response import StreamlitResponse
-import numpy as np
-#from pandasai import Agent
-import json
-import matplotlib.pyplot as plt
-
-
-
-os.environ['PANDASAI_API_KEY'] = "$2a$10$2s0v3C29vItNS2CO4QX10OV51/OONFCUNa4e9EU90w2Gozw88f4vK"
-st.set_page_config(page_title="SAP Data Analysis", layout="wide")
-st.image('Pandas-AI-Logo.png', caption=None)
-uploaded_file = st.file_uploader("Upload CSV data for analysis", type=['csv'])
-#uploaded_file = st.file_uploader("Upload EXcel data for analysis", type=['xlsx'])
-
-df1 = ""
-sdf = ""
-data = [{"Feb 24":"","1.":"Do","2.":"Fr","3.":"Sa","4.":"So","5.":"Mo","6.":"Di","7.":"Mi","8.":"Do","9.":"Fr","10.":"Sa","11.":"So","12.":"Mo","13.":"Di","14.":"Mi","15.":"Do","16.":"Fr","17.":"Sa","18.":"So","19.":"Mo","20.":"Di","21.":"Mi","22.":"Do","23.":"Fr","24.":"Sa","25.":"So","26.":"Mo","27.":"Di","28.":"Mi","29.":"Do"},{"Feb 24":"Standke Steffen","1.":"F","2.":"F","3.":"","4.":"","5.":"","6.":"","7.":"","8.":"","9.":"","10.":"","11.":"","12.":"","13.":"","14.":"UA","15.":"UA","16.":"","17.":"SD","18.":"SD","19.":"","20.":"","21.":"","22.":"","23.":"","24.":"","25.":"","26.":"","27.":"","28.":"","29.":""},{"Feb 24":"Will Susanne","1.":"","2.":"TZ","3.":"","4.":"","5.":"UA","6.":"","7.":"","8.":"","9.":"TZ","10.":"","11.":"","12.":"","13.":"","14.":"","15.":"","16.":"TZ","17.":"","18.":"","19.":"","20.":"","21.":"","22.":"","23.":"TZ","24.":"","25.":"","26.":"","27.":"","28.":"","29.":""},{"Feb 24":"Raab Julia","1.":"TZ","2.":"TZ","3.":"","4.":"","5.":"","6.":"","7.":"","8.":"TZ","9.":"TZ","10.":"BLOCKER","11.":"","12.":"Ü","13.":"Ü","14.":"Ü","15.":"TZ","16.":"TZ","17.":"BLOCKER","18.":"","19.":"","20.":"","21.":"","22.":"TZ","23.":"TZ","24.":"","25.":"SD","26.":"","27.":"","28.":"","29.":"TZ"},{"Feb 24":"Eckert Marion","1.":"","2.":"","3.":"","4.":"","5.":"","6.":"","7.":"","8.":"","9.":"Ü","10.":"","11.":"","12.":"S","13.":"S","14.":"S","15.":"S","16.":"S","17.":"","18.":"","19.":"","20.":"","21.":"","22.":"","23.":"","24.":"","25.":"","26.":"S","27.":"S","28.":"S","29.":"S"},{"Feb 24":"Meder, Milena","1.":"","2.":"","3.":"","4.":"","5.":"","6.":"","7.":"","8.":"","9.":"","10.":"","11.":"","12.":"F","13.":"F","14.":"","15.":"F","16.":"F","17.":"","18.":"","19.":"","20.":"","21.":"","22.":"","23.":"","24.":"","25.":"","26.":"Voloreise","27.":"","28.":"","29.":""},{"Feb 24":"Despang Angelika","1.":"","2.":"","3.":"SD","4.":"","5.":"","6.":"","7.":"","8.":"","9.":"","10.":"","11.":"","12.":"UA","13.":"UA","14.":"UA","15.":"","16.":"","17.":"","18.":"","19.":"F","20.":"F","21.":"F","22.":"F","23.":"F","24.":"","25.":"","26.":"","27.":"","28.":"","29.":""},{"Feb 24":"Heike Beudert","1.":"TZ","2.":"0,5 U","3.":"","4.":"","5.":"TZ","6.":"","7.":"","8.":"","9.":"","10.":"SD","11.":"SD","12.":"UA","13.":"UA","14.":"TZ","15.":"TZ","16.":"TZ","17.":"","18.":"","19.":"TZ","20.":"TZ","21.":"TZ","22.":"TZ","23.":"TZ","24.":"","25.":"","26.":"F","27.":"F","28.":"F","29.":"F"},{"Feb 24":"Borst Benedikt","1.":"","2.":"","3.":"","4.":"SD","5.":"F","6.":"F","7.":"F","8.":"F","9.":"F","10.":"BLOCKER","11.":"","12.":"UA","13.":"UA","14.":"F","15.":"","16.":"","17.":"","18.":"","19.":"","20.":"","21.":"","22.":"","23.":"","24.":"BLOCKER","25.":"","26.":"","27.":"","28.":"","29.":""}]
-
-#df = pd.DataFrame(data)
-#sdf = SmartDataframe(df)
-#df1
-
-if uploaded_file is not None:
-    
-    #Dateien im CSV Format
-    df1 = pd.read_csv(uploaded_file)
- 
-    # Dateien im XLSX Format
-    #df1 = pd.read_excel(uploaded_file, sheet_name=NONE)
-    #st.table(df1)
-    df1 = pd.DataFrame(df1)
-    st.success("Daten erfolgreich geladen!")
-    df1
-    #sdf = SmartDataframe(df1)
-    
-
-bild = st.empty()     
-bild.subheader("Datenanalyse & Datenvisualisierung")
-c = st.container(border=True)
-prompt = st.text_area("Enter your prompt:")
-if st.button("Generate"):
-
-        if prompt:
-            #c.text("Generating response...") 
-            if os.path.isfile('./exports/charts/temp_chart.png'):
-                os.remove('./exports/charts/temp_chart.png')
-            #spin = st.spinner
-            with c:
-                with st.spinner("Generating response..."):
-                    
-                    #bar = st.progress(20)
-                    
-                    #bar = st.progress(100)
-              
-                    with bild:
-                        sdf = SmartDataframe(df1)
-                        st.write(sdf.chat(prompt))
-                    #with st.spinner("Generating response..."):
-                       
-                        if os.path.isfile('./exports/charts/temp_chart.png'):
-                            st.image('./exports/charts/temp_chart.png')
-
-                         
-                    #st.success('Done!')
-                    #bar.progress(100)
-                    
-                #c.write(bar)
-                #c.write(st.spinner)
-                #bild.empty()
-            
-                #st.write(sdf.chat(prompt))            
-            #bar.progress(100)
-            
-
-        else:
-
-            st.error("Please enter a prompt.")
-
-
-
-#with placeholder.container():
-    #st.write("This is one element")
-    #st.write("This is another")
-
-
-
-#agent = Agent(df)
-#result = agent.chat("erstelle balkendiagramm")
-#st.write(result)
-#sdf = SmartDataframe(df)
-#sdf.chat("draw chart")
-#pandas_ai = PandasAI(llm, verbose=True, save_charts=True)
-#st.write(sdf.chat("Plot a chart"))
-#st.write(st.bar_chart(data))
-'''
-
-
-
-
-
+        # Open the audio file
+        with open(file_path, "rb") as file:
+            # Create a transcription of the audio file
+            transcription = client.audio.transcriptions.create(
+                file=(os.path.basename(file_path), file.read()),  # Correct passing of filename
+                model="whisper-large-v3-turbo",  # Required model to use for transcription
+                prompt="transcribe",  # Optional
+                language="de",  # Optional
+                response_format="json",  # Optional
+                temperature=0.0  # Optional
+            )
+            # Return the transcription text
+            return transcription.text
+    except Exception as e:
+        return f"Ein Fehler ist aufgetreten: {str(e)}"
 
+# Streamlit Interface
+st.title("Audio Transkription")
+sr_outputs = st.empty()  # Platzhalter für die Transkription
+sr_inputs = st.file_uploader("Laden Sie eine Audiodatei hoch", type=["wav", "mp3"])
 
+if sr_inputs is not None:
+    audio_data = sf.read(sr_inputs)
+    transcription = processaudio(audio_data)
+    sr_outputs.text(transcription)