Spaces:
Sleeping
Sleeping
File size: 8,431 Bytes
14f13a5 |
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 |
"""
Document chunking strategies for RAG.
Implements semantic chunking with overlap, metadata enrichment,
and configurable strategies for different content types.
"""
import re
from typing import List, Dict, Any, Optional
from dataclasses import dataclass
import logging
logger = logging.getLogger(__name__)
@dataclass
class DocumentChunk:
"""Represents a single document chunk with metadata."""
content: str
metadata: Dict[str, Any]
chunk_id: str
def to_dict(self) -> Dict[str, Any]:
"""Convert to dictionary for storage."""
return {
"content": self.content,
"metadata": self.metadata,
"chunk_id": self.chunk_id
}
class SemanticChunker:
"""
Smart chunking that preserves semantic meaning.
Features:
- Splits on natural boundaries (paragraphs, sentences)
- Maintains context with overlap
- Preserves code blocks intact
- Enriches chunks with metadata
"""
def __init__(
self,
chunk_size: int = 600,
chunk_overlap: int = 100,
preserve_code_blocks: bool = True
):
self.chunk_size = chunk_size
self.chunk_overlap = chunk_overlap
self.preserve_code_blocks = preserve_code_blocks
def chunk_document(
self,
text: str,
metadata: Optional[Dict[str, Any]] = None
) -> List[DocumentChunk]:
"""
Split document into semantically meaningful chunks.
Args:
text: Document text to chunk
metadata: Optional metadata to attach to all chunks
Returns:
List of DocumentChunk objects
"""
if not text or not text.strip():
logger.warning("Empty text provided for chunking")
return []
metadata = metadata or {}
# Extract and preserve code blocks
code_blocks = []
if self.preserve_code_blocks:
text, code_blocks = self._extract_code_blocks(text)
# Split into paragraphs first
paragraphs = self._split_paragraphs(text)
# Create chunks
chunks = []
current_chunk = []
current_size = 0
for i, para in enumerate(paragraphs):
para_tokens = self._estimate_tokens(para)
# If single paragraph exceeds chunk size, split by sentences
if para_tokens > self.chunk_size:
if current_chunk:
chunks.append(self._create_chunk(
current_chunk,
metadata,
len(chunks)
))
current_chunk = []
current_size = 0
# Split long paragraph
sentence_chunks = self._split_long_paragraph(para, metadata, len(chunks))
chunks.extend(sentence_chunks)
continue
# Add paragraph to current chunk
if current_size + para_tokens <= self.chunk_size:
current_chunk.append(para)
current_size += para_tokens
else:
# Save current chunk
if current_chunk:
chunks.append(self._create_chunk(
current_chunk,
metadata,
len(chunks)
))
# Start new chunk with overlap
overlap_text = self._get_overlap_text(current_chunk)
current_chunk = [overlap_text, para] if overlap_text else [para]
current_size = self._estimate_tokens(overlap_text) + para_tokens
# Add remaining chunk
if current_chunk:
chunks.append(self._create_chunk(
current_chunk,
metadata,
len(chunks)
))
# Reinsert code blocks
if code_blocks:
chunks = self._reinsert_code_blocks(chunks, code_blocks)
logger.info(f"Created {len(chunks)} chunks from document")
return chunks
def _extract_code_blocks(self, text: str) -> tuple[str, List[Dict[str, str]]]:
"""Extract code blocks to preserve them intact."""
code_pattern = r'```[\s\S]*?```|`[^`]+`'
code_blocks = []
def replace_code(match):
placeholder = f"__CODE_BLOCK_{len(code_blocks)}__"
code_blocks.append({
"placeholder": placeholder,
"content": match.group(0)
})
return placeholder
text_without_code = re.sub(code_pattern, replace_code, text)
return text_without_code, code_blocks
def _reinsert_code_blocks(
self,
chunks: List[DocumentChunk],
code_blocks: List[Dict[str, str]]
) -> List[DocumentChunk]:
"""Reinsert code blocks into chunks."""
for chunk in chunks:
for code_block in code_blocks:
chunk.content = chunk.content.replace(
code_block["placeholder"],
code_block["content"]
)
return chunks
def _split_paragraphs(self, text: str) -> List[str]:
"""Split text into paragraphs."""
# Split on double newlines or more
paragraphs = re.split(r'\n\s*\n', text)
return [p.strip() for p in paragraphs if p.strip()]
def _split_long_paragraph(
self,
paragraph: str,
metadata: Dict[str, Any],
start_idx: int
) -> List[DocumentChunk]:
"""Split a long paragraph by sentences."""
sentences = re.split(r'(?<=[.!?])\s+', paragraph)
chunks = []
current_chunk = []
current_size = 0
for sentence in sentences:
sentence_tokens = self._estimate_tokens(sentence)
if current_size + sentence_tokens <= self.chunk_size:
current_chunk.append(sentence)
current_size += sentence_tokens
else:
if current_chunk:
chunks.append(self._create_chunk(
current_chunk,
metadata,
start_idx + len(chunks)
))
current_chunk = [sentence]
current_size = sentence_tokens
if current_chunk:
chunks.append(self._create_chunk(
current_chunk,
metadata,
start_idx + len(chunks)
))
return chunks
def _create_chunk(
self,
text_segments: List[str],
metadata: Dict[str, Any],
chunk_idx: int
) -> DocumentChunk:
"""Create a DocumentChunk from text segments."""
content = "\n\n".join(text_segments)
# Enrich metadata
enriched_metadata = {
**metadata,
"chunk_index": chunk_idx,
"chunk_size": len(content),
"has_code": "```" in content or "`" in content,
}
chunk_id = f"{metadata.get('source', 'unknown')}_{chunk_idx}"
return DocumentChunk(
content=content,
metadata=enriched_metadata,
chunk_id=chunk_id
)
def _get_overlap_text(self, chunks: List[str]) -> str:
"""Get overlap text from previous chunks."""
if not chunks:
return ""
combined = " ".join(chunks[-2:])
tokens = self._estimate_tokens(combined)
if tokens <= self.chunk_overlap:
return combined
# Truncate to overlap size
words = combined.split()
overlap_words = words[-(self.chunk_overlap // 4):]
return " ".join(overlap_words)
@staticmethod
def _estimate_tokens(text: str) -> int:
"""Rough token estimation (1 token ≈ 4 characters)."""
return len(text) // 4
def create_chunker(chunk_size: int = 600, chunk_overlap: int = 100) -> SemanticChunker:
"""Factory function to create a chunker instance."""
return SemanticChunker(
chunk_size=chunk_size,
chunk_overlap=chunk_overlap,
preserve_code_blocks=True
)
|