Update run.py

run.py

@@ -3,7 +3,66 @@ import os
 import gradio as gr
 import json
 from huggingface_hub import InferenceClient
+import gspread
+from google.oauth2 import service_account
+from datetime import datetime
+import chromadb
+
+# Google Sheets setup
+scope = ["https://spreadsheets.google.com/feeds", "https://www.googleapis.com/auth/drive"]
+key1 = os.getenv("key1")
+key2 = os.getenv("key2")
+key3 = os.getenv("key3")
+key4 = os.getenv("key4")
+key5 = os.getenv("key5")
+key6 = os.getenv("key6")
+key7 = os.getenv("key7")
+key8 = os.getenv("key8")
+key9 = os.getenv("key9")
+key10 = os.getenv("key10")
+key11 = os.getenv("key11")
+key12 = os.getenv("key12")
+key13 = os.getenv("key13")
+key14 = os.getenv("key14")
+key15 = os.getenv("key15")
+key16 = os.getenv("key16")
+key17 = os.getenv("key17")
+key18 = os.getenv("key18")
+key19 = os.getenv("key19")
+key20 = os.getenv("key20")
+key21 = os.getenv("key21")
+key22 = os.getenv("key22")
+key23 = os.getenv("key23")
+key24 = os.getenv("key24")
+key25 = os.getenv("key25")
+key26 = os.getenv("key26")
+key27 = os.getenv("key27")
+key28 = os.getenv("key28")
+pkey="-----BEGIN PRIVATE KEY-----\n"+key2+"\n"+key3+"\n"+ key4+"\n"+key5+"\n"+ key6+"\n"+key7+"\n"+key8+"\n"+key9+"\n"+key10+"\n"+key11+"\n"+key12+"\n"+key13+"\n"+key14+"\n"+key15+"\n"+key16+"\n"+key17+"\n"+key18+"\n"+key19+"\n"+key20+"\n"+key21+"\n"+key22+"\n"+key24+"\n"+key25+"\n"+key26+"\n"+key27+"\n"+key28+"\n-----END PRIVATE KEY-----\n"
+json_data={
+  "type": "service_account",
+  "project_id": "nestolechatbot",
+  "private_key_id": key1,
+  "private_key": pkey,
+  "client_email": "[email protected]",
+  "client_email": "[email protected]",
+  "client_id": "107457262210035412036",
+  "auth_uri": "https://accounts.google.com/o/oauth2/auth",
+  "token_uri": "https://oauth2.googleapis.com/token",
+  "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
+  "client_x509_cert_url": "https://www.googleapis.com/robot/v1/metadata/x509/nestoleservice%40nestolechatbot.iam.gserviceaccount.com",
+  "universe_domain": "googleapis.com"
+}
+creds = service_account.Credentials.from_service_account_info(json_data, scopes=scope)
+
+client = gspread.authorize(creds)
+sheet = client.open("nestolechatbot").sheet1  # Open the sheet
 
+def save_to_sheet(date, name, message, IP, dev, header):
+    # Write user input to the Google Sheet
+    sheet.append_row([date, name, message, IP, dev, header])
+    return f"Thanks {name}, your message has been saved!"
+    
 path='/Users/thiloid/Desktop/LSKI/ole_nest/Chatbot/LLM/chroma'
 if(os.path.exists(path)==False): path="/home/user/app/chroma"
 
@@ -109,13 +168,38 @@ def parse_for_nc(text):
     return "No"
 
 
-def format_prompt(message):
-  prompt = "" #"<s>"
-  #for user_prompt, bot_response in history:
-  #  prompt += f"[INST] {user_prompt} [/INST]"
-  #  prompt += f" {bot_response}</s> "
-  prompt += f"[INST] {message} [/INST]"
-  return prompt
+def extract_ip_and_device(headers_obj):
+    ip_address = None
+    device_info = None
+    
+    # Access the raw headers list
+    headers = headers_obj.raw
+    
+    for header in headers:
+        if len(header) != 2:
+            print(f"Unexpected header format: {header}")
+            continue
+        
+        key, value = header
+        
+        if key == b'x-forwarded-for':
+            ip_address = value.decode('utf-8')
+        elif key == b'user-agent':
+            device_info = value.decode('utf-8')
+    
+    return ip_address, device_info
+
+def format_prompt(message, history):
+    print("HISTORY")
+    print(history)
+    prompt = ""
+    if history:
+        user_prompt, bot_response = history[-1]
+        prompt += f"[INST] {user_prompt} [/INST] {bot_response}</s> "
+    prompt += f"[INST] {message} [/INST]"
+    print("Final P")
+    print(prompt)
+    return prompt
 
 def responsecritical(
     prompt, temperature=0.9, max_new_tokens=200, top_p=0.95, repetition_penalty=1.0,
@@ -149,7 +233,7 @@ def responsecritical(
 
    
 def response(
-    prompt, temperature=0.9, max_new_tokens=500, top_p=0.95, repetition_penalty=1.0,
+    request: gr.Request, prompt, temperature=0.9, max_new_tokens=500, top_p=0.95, repetition_penalty=1.0,
 ):
     temperature = float(0.9)
     if temperature < 1e-2: temperature = 1e-2
@@ -163,18 +247,25 @@ def response(
         seed=42,
     )
 
-      
+    headers = request.headers
+    IP, dev = extract_ip_and_device(headers)
+    
     if responsecritical(prompt)==False:
+        now = str(datetime.now())
+        save_to_sheet(now, prompt, "Es scheint so, als sei dies keine Frage, die sich auf die Studienorientierung bezieht", IP, dev, str(headers))
         yield "Es scheint so, als sei dies keine Frage, die sich auf die Studienorientierung bezieht"
     else:
         answernc=parse_for_nc(prompt)
         if answernc!="No":
+            now = str(datetime.now())
+            save_to_sheet(now, prompt, answernc, IP, dev, str(headers))
             yield answernc
         else:
             prompt = re.sub(r'\buni\b', 'Universität', prompt, flags=re.IGNORECASE)
             addon=""
+            search_prompt = format_prompt(prompt, history)
             results=collection.query(
-          query_texts=[prompt],
+          query_texts=[search_prompt],
           n_results=60
     )
 
@@ -185,13 +276,15 @@ def response(
             if(len(results)>1):
               addon=" Bitte berücksichtige bei deiner Antwort ausschießlich folgende Auszüge aus unserer Datenbank, sofern sie für die Antwort erforderlich sind. Beantworte die Frage knapp und präzise. Ignoriere unpassende Datenbank-Auszüge OHNE sie zu kommentieren, zu erwähnen oder aufzulisten:\n"+"\n".join(results)
             system="Du bist ein deutschsprachiges KI-basiertes Studienberater Assistenzsystem, das zu jedem Anliegen möglichst geeignete Studieninformationen empfiehlt."+addon+"\n\nUser-Anliegen:"   
-            formatted_prompt = format_prompt(system+"\n"+prompt)
+            formatted_prompt = format_prompt(system+"\n"+prompt, history)
             stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
             output = ""
             for response in stream:
                 output += response.token.text
             osaanswer=simosa(prompt, osalist, osa)
             output=output[:-4]+"\n"+osaanswer
+            now = str(datetime.now())
+            save_to_sheet(now, prompt, output, IP, dev, str(headers))
             yield output
             
 gr.ChatInterface(response, chatbot=gr.Chatbot(value=[[None,"Herzlich willkommen! Ich bin Chätti ein KI-basiertes Studienassistenzsystem, das für jede Anfrage die besten Studieninformationen empfiehlt.<br>Erzähle mir was dich interessiert! Allgemein kann ich dir diesen Test (https://www.was-studiere-ich.de/) oder diese Seminare (https://www.bw-best.de) zur Studienfindung empfehlen."]],render_markdown=True),title="German Studyhelper Chätti").queue().launch(share=True) #False, server_name="0.0.0.0", server_port=7864)