Spaces:
Sleeping
Sleeping
File size: 14,046 Bytes
192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 afad319 192b2d2 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 |
#!/usr/bin/env python3
"""
# Simplified PDF Processor for Hugging Face Spaces
This module provides comprehensive PDF processing functionality for the RAG system.
## Overview
The PDF processor handles the complete pipeline from raw PDF files to structured,
searchable document chunks. It includes:
- **Text Extraction**: Robust PDF text extraction with error handling
- **Text Cleaning**: Intelligent preprocessing and normalization
- **Metadata Extraction**: Document title, author, and file information
- **Smart Chunking**: Multiple chunk sizes for optimal retrieval
- **Query Preprocessing**: Text normalization for search queries
## Key Features
- 📄 **Multi-format Support**: Handles various PDF structures and layouts
- 🧹 **Intelligent Cleaning**: Removes noise while preserving important content
- 📏 **Flexible Chunking**: Multiple chunk sizes for different use cases
- 🔍 **Search Optimization**: Preprocessing for better retrieval performance
- 🛡️ **Error Handling**: Graceful handling of corrupted or problematic files
## Architecture
The processor follows a modular design:
1. **Text Extraction**: Raw PDF to text conversion
2. **Text Cleaning**: Noise removal and normalization
3. **Metadata Extraction**: Document information extraction
4. **Chunking**: Intelligent text segmentation
5. **Query Processing**: Search query optimization
## Usage Example
```python
processor = SimplePDFProcessor()
processed_doc = processor.process_document("document.pdf", [100, 400])
print(f"Processed {len(processed_doc.chunks)} chunks")
```
"""
import os
import re
import uuid
from typing import List, Dict, Optional
from dataclasses import dataclass
from pathlib import Path
import pypdf
from loguru import logger
# =============================================================================
# DATA STRUCTURES
# =============================================================================
@dataclass
class DocumentChunk:
"""
Represents a processed document chunk with metadata
Attributes:
text: The cleaned and processed text content
doc_id: Unique identifier for the source document
filename: Name of the source PDF file
chunk_id: Unique identifier for this specific chunk
chunk_size: Target size used for chunking (in tokens)
"""
text: str
doc_id: str
filename: str
chunk_id: str
chunk_size: int
@dataclass
class ProcessedDocument:
"""
Represents a completely processed PDF document
Attributes:
filename: Name of the PDF file
title: Extracted or inferred document title
author: Extracted or inferred document author
chunks: List of processed document chunks
"""
filename: str
title: str
author: str
chunks: List[DocumentChunk]
# =============================================================================
# MAIN PDF PROCESSOR CLASS
# =============================================================================
class SimplePDFProcessor:
"""
Simplified PDF processor for Hugging Face Spaces
This class provides comprehensive PDF processing capabilities including:
- Text extraction and cleaning
- Metadata extraction
- Intelligent chunking
- Query preprocessing
- Error handling and logging
"""
def __init__(self):
"""
Initialize the PDF processor with default settings
Sets up stop words and processing parameters for optimal
document processing and search performance.
"""
# Common English stop words for query preprocessing
self.stop_words = {
"the",
"a",
"an",
"and",
"or",
"but",
"in",
"on",
"at",
"to",
"for",
"of",
"with",
"by",
"is",
"are",
"was",
"were",
"be",
"been",
"being",
"have",
"has",
"had",
"do",
"does",
"did",
"will",
"would",
"could",
"should",
"may",
"might",
"can",
"this",
"that",
"these",
"those",
}
def process_document(
self, file_path: str, chunk_sizes: List[int] = None
) -> ProcessedDocument:
"""
Process a PDF document through the complete pipeline
This method orchestrates the entire PDF processing workflow:
1. Extracts text from the PDF file
2. Cleans and normalizes the text
3. Extracts document metadata
4. Creates chunks of different sizes
5. Returns a structured document object
Args:
file_path: Path to the PDF file to process
chunk_sizes: List of chunk sizes to create (in tokens)
Returns:
ProcessedDocument object with metadata and chunks
Raises:
Exception: If document processing fails
"""
if chunk_sizes is None:
chunk_sizes = [100, 400] # Default chunk sizes
try:
# Step 1: Extract raw text from PDF
text = self._extract_text(file_path)
# Step 2: Clean and normalize the text
cleaned_text = self._clean_text(text)
# Step 3: Extract document metadata
metadata = self._extract_metadata(file_path)
# Step 4: Create chunks of different sizes
chunks = []
doc_id = str(uuid.uuid4()) # Generate unique document ID
for chunk_size in chunk_sizes:
chunk_list = self._create_chunks(
cleaned_text, chunk_size, doc_id, metadata["filename"]
)
chunks.extend(chunk_list)
# Step 5: Return processed document
return ProcessedDocument(
filename=metadata["filename"],
title=metadata["title"],
author=metadata["author"],
chunks=chunks,
)
except Exception as e:
logger.error(f"Error processing document {file_path}: {e}")
raise
def _extract_text(self, file_path: str) -> str:
"""
Extract text content from a PDF file
This method:
1. Opens the PDF file safely
2. Iterates through all pages
3. Extracts text from each page
4. Combines all text with proper spacing
5. Handles extraction errors gracefully
Args:
file_path: Path to the PDF file
Returns:
Extracted text content as a string
Raises:
Exception: If text extraction fails
"""
try:
with open(file_path, "rb") as file:
# Create PDF reader object
pdf_reader = pypdf.PdfReader(file)
text = ""
# Extract text from each page
for page in pdf_reader.pages:
page_text = page.extract_text()
if page_text:
text += page_text + "\n"
return text
except Exception as e:
logger.error(f"Error extracting text from {file_path}: {e}")
raise
def _clean_text(self, text: str) -> str:
"""
Clean and normalize extracted text
This method performs comprehensive text cleaning:
1. Removes excessive whitespace and newlines
2. Normalizes special characters while preserving punctuation
3. Removes page numbers and headers/footers
4. Ensures consistent formatting
Args:
text: Raw extracted text from PDF
Returns:
Cleaned and normalized text
"""
# Remove excessive whitespace (multiple spaces, tabs, etc.)
text = re.sub(r"\s+", " ", text)
# Remove special characters but preserve important punctuation
# This keeps: letters, numbers, spaces, and common punctuation
text = re.sub(r"[^\w\s\.\,\!\?\;\:\-\(\)\[\]\{\}]", "", text)
# Remove standalone page numbers at line ends
# These are often artifacts from PDF extraction
text = re.sub(r"\b\d+\b(?=\s*\n)", "", text)
# Normalize excessive newlines to consistent paragraph breaks
text = re.sub(r"\n\s*\n\s*\n+", "\n\n", text)
return text.strip()
def _extract_metadata(self, file_path: str) -> Dict[str, str]:
"""
Extract metadata from PDF file
This method attempts to extract:
1. Document title from PDF metadata
2. Author information from PDF metadata
3. Falls back to filename if metadata is unavailable
Args:
file_path: Path to the PDF file
Returns:
Dictionary containing filename, title, and author
"""
try:
with open(file_path, "rb") as file:
pdf_reader = pypdf.PdfReader(file)
info = pdf_reader.metadata
return {
"filename": Path(file_path).name,
"title": (
info.get("/Title", Path(file_path).stem)
if info
else Path(file_path).stem
),
"author": info.get("/Author", "Unknown") if info else "Unknown",
}
except Exception as e:
logger.warning(f"Error extracting metadata from {file_path}: {e}")
# Fallback to basic information
return {
"filename": Path(file_path).name,
"title": Path(file_path).stem,
"author": "Unknown",
}
def _create_chunks(
self, text: str, chunk_size: int, doc_id: str, filename: str
) -> List[DocumentChunk]:
"""
Create text chunks of specified size
This method implements intelligent chunking:
1. Splits text into sentences for natural boundaries
2. Groups sentences into chunks of target size
3. Ensures chunks don't exceed the specified token limit
4. Creates unique identifiers for each chunk
Args:
text: Clean text to chunk
chunk_size: Target chunk size in tokens
doc_id: Unique document identifier
filename: Source filename
Returns:
List of DocumentChunk objects
"""
chunks = []
# Split text into sentences for natural chunking
sentences = self._split_into_sentences(text)
current_chunk = ""
chunk_id = 0
for sentence in sentences:
# Estimate token count (rough approximation using word count)
estimated_tokens = len(sentence.split())
# Add sentence to current chunk if it fits
if len(current_chunk.split()) + estimated_tokens <= chunk_size:
current_chunk += sentence + " "
else:
# Save current chunk if not empty
if current_chunk.strip():
chunks.append(
DocumentChunk(
text=current_chunk.strip(),
doc_id=doc_id,
filename=filename,
chunk_id=f"{doc_id}_{chunk_id}",
chunk_size=chunk_size,
)
)
chunk_id += 1
# Start new chunk with current sentence
current_chunk = sentence + " "
# Add the last chunk if not empty
if current_chunk.strip():
chunks.append(
DocumentChunk(
text=current_chunk.strip(),
doc_id=doc_id,
filename=filename,
chunk_id=f"{doc_id}_{chunk_id}",
chunk_size=chunk_size,
)
)
return chunks
def _split_into_sentences(self, text: str) -> List[str]:
"""
Split text into sentences for intelligent chunking
This method:
1. Uses regex patterns to identify sentence boundaries
2. Filters out very short sentences (likely noise)
3. Ensures minimum sentence quality
Args:
text: Text to split into sentences
Returns:
List of sentence strings
"""
# Split on sentence-ending punctuation
sentences = re.split(r"[.!?]+", text)
# Clean and filter sentences
cleaned_sentences = []
for sentence in sentences:
sentence = sentence.strip()
# Only include sentences with meaningful content (minimum 3 words)
if sentence and len(sentence.split()) > 3:
cleaned_sentences.append(sentence)
return cleaned_sentences
def preprocess_query(self, query: str) -> str:
"""
Preprocess query text for better search performance
This method applies text normalization techniques:
1. Converts to lowercase for case-insensitive matching
2. Removes punctuation that might interfere with search
3. Filters out common stop words
4. Returns normalized query string
Args:
query: Raw query string from user
Returns:
Preprocessed query string optimized for search
"""
# Convert to lowercase for consistent matching
query = query.lower()
# Remove punctuation that might interfere with search
query = re.sub(r"[^\w\s]", "", query)
# Remove stop words to focus on meaningful terms
words = query.split()
filtered_words = [word for word in words if word not in self.stop_words]
return " ".join(filtered_words)
|