File size: 8,729 Bytes
ca6ed76 257879f ca6ed76 bf3527f 257879f a85fb97 257879f f963a91 f4f4fb0 ca6ed76 9432b86 c0aba14 77d13cb ca6ed76 aa93a3c 77d13cb ca6ed76 f963a91 aa93a3c f963a91 21efff8 ca6ed76 f963a91 21efff8 ca6ed76 21efff8 ca6ed76 04d5464 ca6ed76 f963a91 f4f4fb0 0ad8cb9 ca6ed76 f963a91 ca6ed76 f963a91 ca6ed76 f963a91 ca6ed76 257879f a85fb97 257879f ca6ed76 257879f ca6ed76 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 |
#chat_service.py
from typing import List, Dict, Any, Optional, Tuple
from datetime import datetime
import logging
from config.config import settings
import asyncio
from io import StringIO
import pandas as pd
logger = logging.getLogger(__name__)
class ConversationManager:
"""Manages conversation history and context"""
def __init__(self):
self.conversations: Dict[str, List[Dict[str, Any]]] = {}
self.max_history = 1
def add_interaction(
self,
session_id: str,
user_input: str,
response: str,
context: Optional[Dict[str, Any]] = None
) -> None:
if session_id not in self.conversations:
self.conversations[session_id] = []
self.conversations[session_id].append({
'timestamp': datetime.now().isoformat(),
'user_input': user_input,
'response': response,
'context': context
})
if len(self.conversations[session_id]) > self.max_history:
self.conversations[session_id] = self.conversations[session_id][-self.max_history:]
def get_history(self, session_id: str) -> List[Dict[str, Any]]:
return self.conversations.get(session_id, [])
def clear_history(self, session_id: str) -> None:
if session_id in self.conversations:
del self.conversations[session_id]
class ChatService:
def __init__(
self,
model_service,
data_service,
pdf_service,
faq_service
):
self.model = model_service.model
self.tokenizer = model_service.tokenizer
self.data_service = data_service
self.pdf_service = pdf_service
self.faq_service = faq_service
self.conversation_manager = ConversationManager()
async def search_all_sources(
self,
query: str,
top_k: int = 3
) -> Dict[str, List[Dict[str, Any]]]:
"""Search across all available data sources"""
try:
print("-----------------------------")
print("starting searches .... ")
# Await the search calls since they're coroutines
products = await self.data_service.search(query, top_k)
pdfs = await self.pdf_service.search(query, top_k)
# faqs = await self.faq_service.search_faqs(query, top_k)
results = {
'products': products or [],
'documents': pdfs or [],
# 'faqs': faqs or []
}
print("Search results:", results)
return results
except Exception as e:
logger.error(f"Error searching sources: {e}")
return {'products': [], 'documents': [], 'faqs': []}
def construct_system_prompt(self, context: str) -> str:
"""Constructs the system message."""
return (
"You are a friendly bot named: Oma Erna, specializing in Bofrost products and content. Use only the context from this prompt. "
"Return comprehensive German answers. If possible add product IDs from context. Do not make up information. The context is is truth. "
"Use the following context (product descriptions and information) for answers:\n\n"
f"{context}\n\n"
)
def construct_prompt(
self,
user_input: str,
context: str,
chat_history: List[Tuple[str, str]],
max_history_turns: int = 1
) -> str:
"""Constructs the full prompt."""
system_message = self.construct_system_prompt(context)
prompt = f"<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n{system_message}<|eot_id|>"
for user_msg, assistant_msg in chat_history[-max_history_turns:]:
prompt += f"<|start_header_id|>user<|end_header_id|>\n\n{user_msg}<|eot_id|>"
prompt += f"<|start_header_id|>assistant<|end_header_id|>\n\n{assistant_msg}<|eot_id|>"
prompt += f"<|start_header_id|>user<|end_header_id|>\n\n{user_input}<|eot_id|>"
prompt += "<|start_header_id|>assistant<|end_header_id|>\n\n"
return prompt
def build_context(
self,
search_results: Dict[str, List[Dict[str, Any]]],
chat_history: List[Dict[str, Any]]
) -> str:
"""Build context for the model from search results and chat history"""
context_parts = []
# Add relevant products
if search_results.get('products'):
products = search_results['products'][:2] # Limit to top 2 products
for product in products:
context_parts.append(
f"Produkt: {product['Name']}\n"
f"Beschreibung: {product['Description']}\n"
f"Preis: {product['Price']}€\n"
f"Kategorie: {product['ProductCategory']}"
)
# Add relevant PDF content
if search_results.get('documents'):
docs = search_results['documents'][:2]
for doc in docs:
context_parts.append(
f"Aus Dokument '{doc['source']}' (Seite {doc['page']}):\n"
f"{doc['text']}"
)
# Add relevant FAQs
if search_results.get('faqs'):
faqs = search_results['faqs'][:2]
for faq in faqs:
context_parts.append(
f"FAQ:\n"
f"Frage: {faq['question']}\n"
f"Antwort: {faq['answer']}"
)
# Add recent chat history
if chat_history:
print("--- historiy--- ")
#recent_history = chat_history[-3:] # Last 3 interactions
#history_text = "\n".join(
# f"User: {h['user_input']}\nAssistant: {h['response']}"
# for h in recent_history
#)
#context_parts.append(f"Letzte Interaktionen:\n{history_text}")
print("\n\n".join(context_parts))
return "\n\n".join(context_parts)
async def chat(
self,
user_input: str,
session_id: Any,
max_length: int = 1000
) -> Tuple[str, List[Tuple[str, str]], Dict[str, List[Dict[str, Any]]]]:
"""Main chat method that coordinates the entire conversation flow."""
try:
if not isinstance(session_id, str):
session_id = str(session_id)
chat_history_raw = self.conversation_manager.get_history(session_id)
chat_history = [
(entry['user_input'], entry['response']) for entry in chat_history_raw
]
search_results = await self.search_all_sources(user_input)
print(search_results)
context = self.build_context(search_results, chat_history_raw)
prompt = self.construct_prompt(user_input, context, chat_history)
response = self.generate_response(prompt, max_length)
self.conversation_manager.add_interaction(
session_id,
user_input,
response,
{'search_results': search_results}
)
formatted_history = [
(entry['user_input'], entry['response'])
for entry in self.conversation_manager.get_history(session_id)
]
return response, formatted_history, search_results
except Exception as e:
logger.error(f"Error in chat: {e}")
raise
def generate_response(
self,
prompt: str,
max_length: int = 1000
) -> str:
"""Generate response using the language model"""
try:
print(prompt)
inputs = self.tokenizer(
prompt,
return_tensors="pt",
truncation=True,
max_length=4096
).to(settings.DEVICE)
outputs = self.model.generate(
**inputs,
max_length=max_length,
num_return_sequences=1,
temperature=0.7,
top_p=0.9,
do_sample=True,
no_repeat_ngram_size=3,
early_stopping=False
)
response = self.tokenizer.decode(
outputs[0],
skip_special_tokens=True
)
return response.strip()
except Exception as e:
logger.error(f"Error generating response: {e}")
raise |