added advanced_rag

core/ai.py

@@ -13,30 +13,6 @@
 )
 
 from llama_index.core import PromptTemplate
-
-# Templates
-QA_TEMPLATE = PromptTemplate(
-    "Context information is below.\n"
-    "---------------------\n"
-    "{context_str}\n"
-    "---------------------\n"
-    "Given this information, please answer the question: {query_str}\n"
-    "If you don't know the answer, just say that you don't know. Don't try to make up an answer.\n"
-    "Provide a detailed response and explain your reasoning step by step."
-)
-
-REFINE_TEMPLATE = PromptTemplate(
-    "The original question is as follows: {query_str}\n"
-    "We have provided an existing answer: {existing_answer}\n"
-    "We have the opportunity to refine the existing answer "
-    "(only if needed) with some more context below.\n"
-    "------------\n"
-    "{context_msg}\n"
-    "------------\n"
-    "Given the new context, refine the original answer to better "
-    "answer the question. If the context isn't useful, return the original answer."
-)
-
 # llama index imports 
 # from llama_index.legacy.text_splitter import SentenceSplitter
 from llama_index.core import (

core/ai_1.py

@@ -57,7 +57,7 @@
 }
 
 # Initialize settings
-Settings.chunk_size = 128
+Settings.chunk_size = 512
 Settings.llm = OpenAI(
     model=os.getenv("MODEL_NAME"),
     temperature=0.1,

data/Haq_data_v4.txt

@@ -837,7 +837,6 @@ How can I ensure all relevant schemes show up for an eligible citizen?
 
 English Answer:
 If the questionnaire is completed correctly, the relevant schemes will be shown as per the eligibility criteria. To address this issue:
-
 Ensure that the scheme you are looking for is mapped under your organization.
 Double-check that all questionnaire answers are accurate and complete.
 Verify that you have the latest app version and data update.

main.py

@@ -1,9 +1,10 @@
-import redis
 import logging
-from core.ai import ragindex
+# from core.ai import ragindex
+from core.ai_1 import ragindex
 from telegram import Update
 import os
 import dotenv
+from functools import lru_cache
 
 from telegram import (
     Update, 
@@ -60,6 +61,15 @@ def __new__(cls):
             cls.run_once = True
         return cls._instance
 
+@lru_cache(maxsize=128)
+def get_language_message(lang):
+    messages = {
+        "en": "You have chosen English. \nPlease tell your problem",
+        "hi": "आपने हिंदी चुनी है. \nकृपया मुझे बताएं कि आपको क्या समस्या आ रही है।",
+        "mr": "तुम्ही मराठीची निवड केली आहे. \कृपया मला तुमची समस्या सांगा"
+    }
+    return messages.get(lang, "Language not supported")
+
 async def start(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
     BotInitializer()  # To initialize only once
 
@@ -88,36 +98,18 @@ async def language_handler(update: Update, context: CallbackContext):
     )
 
 async def preferred_language_callback(update: Update, context: CallbackContext):
-
     callback_query = update.callback_query
     languages = {"1": "en", "2": "hi", "3": "mr"}
-    try:
-        preferred_language = callback_query.data
-        lang = languages.get(preferred_language)
-        context.user_data['lang'] = lang
-    except (AttributeError, ValueError):
-        lang = 'en'
-        await context.bot.send_message(
-            chat_id=update.effective_chat.id,
-            text="Error getting language! Setting default to English."
-        )
-
-    text_message = ""
-    if lang == "en":
-        text_message = "You have chosen English. \nPlease tell your problem"
-    elif lang == "hi":
-        text_message = "आपने हिंदी चुनी है. \nकृपया मुझे बताएं कि आपको क्या समस्या आ रही है।"
-    elif lang == "mr":
-        text_message = "तुम्ही मराठीची निवड केली आहे. \कृपया मला तुमची समस्या सांगा"
-
-    await context.bot.send_message(
-        chat_id=update.effective_chat.id,
-        text=text_message
-    )
+    lang = languages.get(callback_query.data, 'en')
+    context.user_data['lang'] = lang
+
+    text_message = get_language_message(lang)
+
+    await callback_query.answer()
+    await callback_query.edit_message_text(text=text_message)
 
     set_redis('lang', lang)
 
-
 async def response_handler(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
     await query_handler(update, context)
 
@@ -126,41 +118,26 @@ async def response_handler(update: Update, context: ContextTypes.DEFAULT_TYPE) -
 
 
 async def query_handler(update: Update, context: CallbackContext):
-
     lang = context.user_data.get('lang')
     if not lang:
         await language_handler(update, context)
         return
 
-    if update.message.text:
+    if update.message and update.message.text:
         text = update.message.text
         print(f"text is {text}")
-        # if check_change_language_query(text):
-        #     await language_handler(update, context)
-        #     return
         await flow(update, context, text)
-    # else:
-    #     if update.message.voice:
-    #         voice = await context.bot.get_file(update.message.voice.file_id)
-    #         await flow_voice(update, context, voice)
-
 
 async def flow(update: Update, context: ContextTypes.DEFAULT_TYPE, text):
-    response = ""
     chat_id = update.effective_chat.id
-    lang = context.user_data.get('lang')
+    lang = context.user_data.get('lang', 'en')
 
     if lang == 'en':
         response_en = ragindex(chat_id, text)
         await context.bot.send_message(chat_id=chat_id, text=response_en)
     else:
-        # response_en = ragindex(chat_id, text)
         response, response_en = bhashini_text_chat(chat_id, text, lang)
-        if response:
-            await context.bot.send_message(chat_id=chat_id, text=response)
-        else:
-            await context.bot.send_message(chat_id=chat_id, text="Sorry, I didn't get that. Reverting to english")
-            await context.bot.send_message(chat_id=chat_id, text=response_en)
+        await context.bot.send_message(chat_id=chat_id, text=response or f"Sorry, I didn't get that. Reverting to English:\n\n{response_en}")
 
 if __name__ == '__main__':
     application = ApplicationBuilder().token(token).read_timeout(30).write_timeout(30).build()
@@ -172,167 +149,4 @@ async def flow(update: Update, context: ContextTypes.DEFAULT_TYPE, text):
     application.add_handler(language_handler_)
     application.add_handler(chosen_language)
     application.add_handler(response_handler)
-    application.run_polling()
-
-
-'''
-"""
-Basic example for a bot that uses inline keyboards. For an in-depth explanation, check out
- https://github.com/python-telegram-bot/python-telegram-bot/wiki/InlineKeyboard-Example.
-"""
-import logging
-
-from telegram import InlineKeyboardButton, InlineKeyboardMarkup, Update
-from telegram.ext import Application, CallbackQueryHandler, CommandHandler, ContextTypes
-
-# Enable logging
-logging.basicConfig(
-    format="%(asctime)s - %(name)s - %(levelname)s - %(message)s", level=logging.INFO
-)
-# set higher logging level for httpx to avoid all GET and POST requests being logged
-logging.getLogger("httpx").setLevel(logging.WARNING)
-
-logger = logging.getLogger(__name__)
-
-
-async def start(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
-    """Sends a message with three inline buttons attached."""
-    keyboard = [
-        [
-            InlineKeyboardButton("Option 1", callback_data="1"),
-            InlineKeyboardButton("Option 2", callback_data="2"),
-        ],
-        [InlineKeyboardButton("Option 3", callback_data="3")],
-    ]
-
-    reply_markup = InlineKeyboardMarkup(keyboard)
-
-    await update.message.reply_text("Please choose:", reply_markup=reply_markup)
-
-
-async def button(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
-    """Parses the CallbackQuery and updates the message text."""
-    query = update.callback_query
-
-    # CallbackQueries need to be answered, even if no notification to the user is needed
-    # Some clients may have trouble otherwise. See https://core.telegram.org/bots/api#callbackquery
-    await query.answer()
-
-    await query.edit_message_text(text=f"Selected option: {query.data}")
-
-
-async def help_command(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
-    """Displays info on how to use the bot."""
-    await update.message.reply_text("Use /start to test this bot.")
-
-
-def main() -> None:
-    """Run the bot."""
-    # Create the Application and pass it your bot's token.
-    application = Application.builder().token("TOKEN").build()
-
-    application.add_handler(CommandHandler("start", start))
-    application.add_handler(CallbackQueryHandler(button))
-    application.add_handler(CommandHandler("help", help_command))
-
-    # Run the bot until the user presses Ctrl-C
-    application.run_polling(allowed_updates=Update.ALL_TYPES)
-
-
-if __name__ == "__main__":
-    main()
-'''
-
-# async def chat_handler(update: Update, context: ContextTypes.DEFAULT_TYPE, text: str):
-#     response = ""
-#     chat_id = update.effective_chat.id
-#     lang = context.user_data.get('lang')
-#     wait_message = get_random_wait_messages(
-#         not_always=True,
-#         lang=lang
-#     )
-#     if wait_message:
-#         await context.bot.send_message(chat_id=chat_id, text=wait_message)
-#     if lang == 'en':
-#         response_en, history = chat(chat_id, text)
-#     else:
-#         response, response_en, history = bhashini_text_chat(chat_id,text, lang)
-#     if response:
-#         await context.bot.send_message(chat_id=chat_id, text=response)
-#     await context.bot.send_message(chat_id=chat_id, text=response_en)
-
-# async def talk_handler(update: Update, context: ContextTypes.DEFAULT_TYPE, voice):    
-#     lang = context.user_data.get('lang')
-#     # getting audio file
-#     audio_file = voice
-#     # audio_file = await context.bot.get_file(update.message.voice.file_id)
-
-#     if lang == 'en':
-#         with tempfile.NamedTemporaryFile(suffix='.wav', delete=True) as temp_audio_file:
-#             await audio_file.download_to_drive(custom_path=temp_audio_file.name)
-#             chat_id = update.effective_chat.id
-
-#             wait_message = get_random_wait_messages(
-#                 not_always=True,
-#                 lang=lang
-#         )
-#             if wait_message:
-#                 await context.bot.send_message(chat_id=chat_id, text=wait_message)
-
-#             with open(temp_audio_file.name, "rb") as file:
-#                 audio_data = file.read()
-#                 audio_base64 = base64.b64encode(audio_data).decode('utf-8')
-
-#                 response_audio, assistant_message, history = audio_chat(
-#                     chat_id, audio_file=open(temp_audio_file.name, "rb")
-#                 )
-#                 response_audio.stream_to_file(temp_audio_file.name)
-#                 # fix this error "raise JSONDecodeError("Expecting value", s, err.value) from None" here
-#                 # duration = get_duration_pydub(temp_audio_file.name)
-#                 await context.bot.send_audio(
-#                     chat_id=chat_id, 
-#                     audio=open(temp_audio_file.name, "rb"), 
-#                     #duration=duration, 
-#                     filename="response.wav",
-#                     performer="Yojana Didi",
-#                 )
-#                 await context.bot.send_message(
-#                     chat_id=chat_id, text=assistant_message
-#                 )
-#                 file.close()
-#     else:
-#         with tempfile.NamedTemporaryFile(suffix='.mp3', delete=True) as temp_audio_file: # suffix='.wav'
-#             await audio_file.download_to_drive(custom_path=temp_audio_file.name)
-#             chat_id = update.effective_chat.id
-
-#             wait_message = get_random_wait_messages(
-#                     not_always=True,
-#                     lang=lang
-#             )
-#             if wait_message:
-#                 await context.bot.send_message(chat_id=chat_id, text=wait_message)
-
-#             with open(temp_audio_file.name, "rb") as file:
-#                 audio_data = file.read()
-#                 audio_base64 = base64.b64encode(audio_data).decode('utf-8')
-#                 response_audio, response, history = bhashini_audio_chat(
-#                     chat_id, 
-#                     audio_file=audio_base64, 
-#                     lang=lang
-#                 )
-#                 file_ = open(temp_audio_file.name, "wb")
-#                 file_.write(response_audio.content)
-#                 file_.close()
-#                 with open(temp_audio_file.name, "rb") as file:
-#                     duration = get_duration_pydub(temp_audio_file.name)
-#                     await context.bot.send_audio(
-#                         chat_id=chat_id, 
-#                         audio=open(temp_audio_file.name, "rb"), 
-#                         duration=duration, 
-#                         filename="response.mp3",
-#                         performer="Yojana Didi",
-#                     )
-#                 await context.bot.send_message(
-#                     chat_id=chat_id, text=response
-#                 )
-#                 file_.close()
+    application.run_polling()