File size: 11,930 Bytes
43efcb9 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 |
"""
Main RAG (Retrieval-Augmented Generation) engine implementation.
"""
import os
import logging
from typing import List, Dict, Any, Optional, Tuple, Union
import numpy as np
# Configure logging
logger = logging.getLogger(__name__)
class RAGEngine:
"""Retrieval-Augmented Generation (RAG) engine for question answering."""
def __init__(
self,
embedder,
vector_db,
llm=None,
top_k: int = 5,
search_type: str = "hybrid",
prompt_template: Optional[str] = None
):
"""
Initialize the RAG engine.
Args:
embedder: Embedding model
vector_db: Vector database for document storage and retrieval
llm: Language model for text generation (optional)
top_k: Number of documents to retrieve
search_type: Type of search ('semantic', 'keyword', 'hybrid')
prompt_template: Optional custom prompt template
"""
self.embedder = embedder
self.vector_db = vector_db
self.llm = llm
self.top_k = top_k
self.search_type = search_type
# Set default prompt template if none provided
if prompt_template is None:
from ..config import DEFAULT_PROMPT_TEMPLATE
self.prompt_template = DEFAULT_PROMPT_TEMPLATE
else:
self.prompt_template = prompt_template
def add_documents(
self,
texts: List[str],
metadata: Optional[List[Dict[str, Any]]] = None,
batch_size: int = 32
) -> List[str]:
"""
Add documents to the database.
Args:
texts: List of text chunks
metadata: Optional list of metadata dictionaries for each text
batch_size: Batch size for embedding generation
Returns:
List of document IDs
"""
from ..storage.vector_db import Document
# Handle metadata
if metadata is None:
metadata = [{} for _ in texts]
elif len(metadata) != len(texts):
raise ValueError(f"Length mismatch: got {len(texts)} texts but {len(metadata)} metadata entries")
# Generate embeddings in batches
doc_ids = []
for i in range(0, len(texts), batch_size):
batch_texts = texts[i:i+batch_size]
batch_metadata = metadata[i:i+batch_size]
# Generate embeddings
logger.info(f"Generating embeddings for batch {i//batch_size + 1}/{(len(texts)-1)//batch_size + 1}")
batch_embeddings = self.embedder.embed(batch_texts)
# Create document objects
documents = []
for text, meta, embedding in zip(batch_texts, batch_metadata, batch_embeddings):
doc = Document(text=text, metadata=meta, embedding=embedding)
documents.append(doc)
# Add to database
batch_ids = self.vector_db.add_documents(documents)
doc_ids.extend(batch_ids)
logger.info(f"Added {len(doc_ids)} documents to database")
return doc_ids
def search(
self,
query: str,
top_k: Optional[int] = None,
search_type: Optional[str] = None,
filter_dict: Optional[Dict[str, Any]] = None
) -> List[Dict[str, Any]]:
"""
Search for relevant documents.
Args:
query: Query string
top_k: Number of results to return (defaults to self.top_k)
search_type: Type of search (defaults to self.search_type)
filter_dict: Dictionary of metadata filters
Returns:
List of document dictionaries
"""
if top_k is None:
top_k = self.top_k
if search_type is None:
search_type = self.search_type
# Create filter function if filter_dict is provided
filter_func = None
if filter_dict:
def filter_func(doc):
for key, value in filter_dict.items():
# Handle nested keys (e.g., "metadata.source")
if "." in key:
parts = key.split(".")
current = doc.metadata
for part in parts[:-1]:
if part not in current:
return False
current = current[part]
if parts[-1] not in current or current[parts[-1]] != value:
return False
elif key not in doc.metadata or doc.metadata[key] != value:
return False
return True
# Generate query embedding
query_embedding = self.embedder.embed(query)
# Perform search
results = self.vector_db.search(query_embedding, top_k, filter_func)
# Convert results to dictionaries
return [
{
"id": doc.id,
"text": doc.text,
"metadata": doc.metadata,
"score": score
}
for doc, score in results
]
def generate_response(
self,
query: str,
top_k: Optional[int] = None,
search_type: Optional[str] = None,
filter_dict: Optional[Dict[str, Any]] = None,
max_tokens: int = 512
) -> Dict[str, Any]:
"""
Generate a response to a query using RAG.
Args:
query: Query string
top_k: Number of documents to retrieve
search_type: Type of search
filter_dict: Optional filter for document retrieval
max_tokens: Maximum number of tokens in the response
Returns:
Dictionary with query, response, and retrieved documents
"""
# Retrieve relevant documents
retrieved_docs = self.search(query, top_k, search_type, filter_dict)
# If no documents were found, return a default message
if not retrieved_docs:
return {
"query": query,
"response": "I couldn't find any relevant information to answer your question.",
"retrieved_documents": [],
"search_type": search_type or self.search_type
}
# Format context from retrieved documents
context = self._format_context(retrieved_docs)
# Format prompt with context and query
prompt = self.prompt_template.format(context=context, query=query)
# Generate response using LLM
if self.llm is None:
logger.warning("No LLM provided, returning only retrieved documents")
response = "No language model available to generate a response. Here's what I found in the documents."
else:
response = self._generate_llm_response(prompt, max_tokens)
# Return the results
return {
"query": query,
"response": response,
"retrieved_documents": retrieved_docs,
"search_type": search_type or self.search_type
}
def _format_context(self, documents: List[Dict[str, Any]]) -> str:
"""
Format retrieved documents into context for the prompt.
Args:
documents: List of retrieved documents
Returns:
Formatted context string
"""
context_parts = []
for i, doc in enumerate(documents):
# Extract relevant fields
text = doc["text"]
metadata = doc["metadata"]
source = metadata.get("source", "Unknown")
# Format the document
doc_text = f"Document {i+1}: [Source: {source}]\n{text}\n"
context_parts.append(doc_text)
return "\n".join(context_parts)
def _generate_llm_response(self, prompt: str, max_tokens: int) -> str:
"""
Generate a response using the LLM.
Args:
prompt: The formatted prompt
max_tokens: Maximum number of tokens in the response
Returns:
Generated response
"""
if hasattr(self.llm, "generate_openai_response"):
# OpenAI-compatible LLM
return self.llm.generate_openai_response(prompt, max_tokens)
elif hasattr(self.llm, "generate_huggingface_response"):
# HuggingFace-compatible LLM
return self.llm.generate_huggingface_response(prompt, max_tokens)
else:
# Default implementation
try:
return self.llm.generate_response(prompt, max_tokens)
except Exception as e:
logger.error(f"Error generating response: {e}")
return "I encountered an error while generating a response."
def update_prompt_template(self, new_template: str) -> None:
"""
Update the prompt template.
Args:
new_template: New prompt template
"""
self.prompt_template = new_template
logger.info("Updated prompt template")
def count_documents(self) -> int:
"""
Get the number of documents in the database.
Returns:
Number of documents
"""
return self.vector_db.count_documents()
def clear_documents(self) -> None:
"""Clear all documents from the database."""
self.vector_db.clear()
logger.info("Cleared all documents from database")
# Factory function to create the RAG engine
def create_rag_engine(
embedder=None,
vector_db=None,
llm=None,
config=None
) -> RAGEngine:
"""
Factory function to create a RAG engine.
Args:
embedder: Embedding model (if None, created based on config)
vector_db: Vector database (if None, created based on config)
llm: Language model (if None, created based on config)
config: Configuration module or dictionary
Returns:
Configured RAGEngine instance
"""
# Load configuration if provided
if config is None:
from ..config import (
TOP_K,
SEARCH_TYPE,
DEFAULT_PROMPT_TEMPLATE
)
else:
TOP_K = config.get("TOP_K", 5)
SEARCH_TYPE = config.get("SEARCH_TYPE", "hybrid")
DEFAULT_PROMPT_TEMPLATE = config.get(
"DEFAULT_PROMPT_TEMPLATE",
"""
Answer the following question based ONLY on the provided context.
Context:
{context}
Question: {query}
Answer:
"""
)
# Create embedding model if not provided
if embedder is None:
from ..embedding.model import create_embedding_model
embedder = create_embedding_model()
# Create vector database if not provided
if vector_db is None:
from ..storage.vector_db import create_vector_database
vector_db = create_vector_database(dimension=embedder.dimension)
# Create language model if not provided and requested
if llm is None:
try:
from ..llm.model import create_llm
llm = create_llm()
except (ImportError, ModuleNotFoundError):
logger.warning("LLM module not found, proceeding without an LLM")
# Create and return the RAG engine
return RAGEngine(
embedder=embedder,
vector_db=vector_db,
llm=llm,
top_k=TOP_K,
search_type=SEARCH_TYPE,
prompt_template=DEFAULT_PROMPT_TEMPLATE
)
|