rag_korean_manufacturing_docs / src /document_processor.py
A7m0d's picture
Upload folder using huggingface_hub
e547616 verified
import logging
from abc import ABC, abstractmethod
from dataclasses import dataclass, field
from datetime import datetime
from pathlib import Path
from typing import Dict, List, Optional, Any, Union
from enum import Enum
import hashlib
import sys
import os
sys.path.append(os.path.dirname(os.path.abspath(__file__))) # Ensure current directory is in
try:
from logger.custom_logger import CustomLoggerTracker
custom_log = CustomLoggerTracker()
logger = custom_log.get_logger("documents_processor")
except ImportError:
# Fallback to standard logging if custom logger not available
logger = logging.getLogger("documents_processor")
class DocumentType(Enum):
"""Supported document types."""
PDF = "pdf"
EXCEL = "excel"
IMAGE = "image"
UNKNOWN = "unknown"
class ProcessingStatus(Enum):
"""Document processing status."""
PENDING = "pending"
PROCESSING = "processing"
COMPLETED = "completed"
FAILED = "failed"
PARTIAL = "partial"
@dataclass
class ExtractedImage:
"""Represents an image extracted from a document."""
image_id: str
filename: str
content: bytes
format: str # PNG, JPEG, etc.
width: Optional[int] = None
height: Optional[int] = None
ocr_text: Optional[str] = None
ocr_confidence: Optional[float] = None
extraction_method: str = "unknown"
metadata: Dict[str, Any] = field(default_factory=dict)
@dataclass
class ExtractedTable:
"""Represents a table extracted from a document."""
table_id: str
headers: List[str]
rows: List[List[str]]
page_number: Optional[int] = None
worksheet_name: Optional[str] = None
cell_range: Optional[str] = None
extraction_confidence: Optional[float] = None
metadata: Dict[str, Any] = field(default_factory=dict)
@dataclass
class ChunkMetadata:
"""Metadata for a document chunk."""
chunk_id: str
document_id: str
chunk_index: int
page_number: Optional[int] = None
worksheet_name: Optional[str] = None
cell_range: Optional[str] = None
section_title: Optional[str] = None
image_references: List[str] = field(default_factory=list)
table_references: List[str] = field(default_factory=list)
extraction_timestamp: datetime = field(default_factory=datetime.now)
confidence_score: Optional[float] = None
@dataclass
class DocumentChunk:
content: str
metadata: ChunkMetadata
embedding: Optional[List[float]] = None
def __post_init__(self):
"""Validate chunk content after initialization."""
if not self.content.strip():
logger.warning(f"Empty content in chunk {self.metadata.chunk_id}")
if len(self.content) > 10000: # Warn for very large chunks
logger.warning(f"Large chunk detected ({len(self.content)} chars): {self.metadata.chunk_id}")
@dataclass
class ProcessedDocument:
"""Represents a fully processed document with all extracted content."""
document_id: str
filename: str
file_path: str
document_type: DocumentType
content: str
metadata: Dict[str, Any]
images: List[ExtractedImage] = field(default_factory=list)
tables: List[ExtractedTable] = field(default_factory=list)
processing_status: ProcessingStatus = ProcessingStatus.PENDING
processing_timestamp: datetime = field(default_factory=datetime.now)
file_size: int = 0
checksum: str = ""
error_message: Optional[str] = None
def __post_init__(self):
"""Generate checksum and validate document after initialization."""
if not self.checksum and Path(self.file_path).exists():
self.checksum = self._generate_checksum()
self.file_size = Path(self.file_path).stat().st_size
def _generate_checksum(self) -> str:
try:
hash_md5 = hashlib.md5()
with open(self.file_path, "rb") as f:
for chunk in iter(lambda: f.read(1024), b""):
hash_md5.update(chunk)
return hash_md5.hexdigest()
except Exception as e:
logger.error(f"Failed to generate checksum for {self.file_path}: {e}")
return ""
class DocumentProcessingError(Exception):
"""Base exception for document processing errors."""
def __init__(self, file_path: str, error_type: str, details: str):
self.file_path = file_path
self.error_type = error_type
self.details = details
super().__init__(f"Document processing error in {file_path}: {error_type} - {details}")
class UnsupportedDocumentTypeError(DocumentProcessingError):
def __init__(self, file_path: str, detected_type: str):
super().__init__(
file_path,
"UnsupportedDocumentType",
f"Document type '{detected_type}' is not supported"
)
class DocumentProcessor(ABC):
def __init__(self, config: Dict[str, Any]):
self.config = config
self.supported_extensions = self._get_supported_extensions()
logger.info(f"Initialized {self.__class__.__name__} with config: {config}")
@abstractmethod
def _get_supported_extensions(self) -> List[str]:
pass
@abstractmethod
def process_document(self, file_path: str) -> ProcessedDocument:
pass
def can_process(self, file_path: str) -> bool:
file_extension = Path(file_path).suffix.lower()
return file_extension in self.supported_extensions
def extract_chunks(self, document: ProcessedDocument, chunk_size: int = 512,
chunk_overlap: int = 50) -> List[DocumentChunk]:
if not document.content.strip():
logger.warning(f"No content to chunk in document {document.document_id}")
return []
chunks = []
content = document.content
start = 0
chunk_index = 0
while start < len(content):
# Calculate end position
end = min(start + chunk_size, len(content))
# Try to break at word boundary if not at end of content
if end < len(content):
# Look for the last space within the chunk
last_space = content.rfind(' ', start, end)
if last_space > start:
end = last_space
# Extract chunk content
chunk_content = content[start:end].strip()
if chunk_content: # Only create chunk if it has content
# Create chunk metadata
metadata = ChunkMetadata(
chunk_id=f"{document.document_id}_chunk_{chunk_index}",
document_id=document.document_id,
chunk_index=chunk_index
)
# Create chunk
chunk = DocumentChunk(
content=chunk_content,
metadata=metadata
)
chunks.append(chunk)
chunk_index += 1
# Move start position with overlap
start = max(end - chunk_overlap, start + 1)
# Prevent infinite loop
if start >= end:
break
logger.info(f"Created {len(chunks)} chunks from document {document.document_id}")
return chunks
def _detect_document_type(self, file_path: str) -> DocumentType:
extension = Path(file_path).suffix.lower()
if extension == '.pdf':
return DocumentType.PDF
elif extension in ['.xlsx', '.xls', '.xlsm']:
return DocumentType.EXCEL
elif extension in ['.png', '.jpg', '.jpeg', '.gif', '.bmp', '.tiff']:
return DocumentType.IMAGE
else:
return DocumentType.UNKNOWN
def _generate_document_id(self, file_path: str) -> str:
"""
Generate a unique document ID based on file path and timestamp.
Args:
file_path: Path to the document file
Returns:
Unique document ID string
"""
file_name = Path(file_path).name
timestamp = datetime.now().isoformat()
content = f"{file_name}_{timestamp}"
return hashlib.md5(content.encode()).hexdigest()
def validate_file(self, file_path: str) -> None:
"""
Validate that a file exists and can be processed.
Args:
file_path: Path to the file to validate
Raises:
DocumentProcessingError: If file validation fails
"""
file_path_obj = Path(file_path)
if not file_path_obj.exists():
raise DocumentProcessingError(
file_path,
"FileNotFound",
f"File does not exist: {file_path}"
)
if not file_path_obj.is_file():
raise DocumentProcessingError(
file_path,
"NotAFile",
f"Path is not a file: {file_path}"
)
# Check file size
max_size = self.config.get('max_file_size_mb', 100) * 1024 * 1024 # Convert to bytes
file_size = file_path_obj.stat().st_size
if file_size > max_size:
raise DocumentProcessingError(
file_path,
"FileTooLarge",
f"File size ({file_size} bytes) exceeds maximum allowed size ({max_size} bytes)"
)
if not self.can_process(file_path):
detected_type = self._detect_document_type(file_path)
raise UnsupportedDocumentTypeError(file_path, detected_type.value)
logger.debug(f"File validation passed for: {file_path}")
class DocumentProcessorFactory:
"""Factory class for creating appropriate document processors."""
_processors = {}
@classmethod
def register_processor(cls, document_type: DocumentType, processor_class):
"""Register a processor class for a document type."""
cls._processors[document_type] = processor_class
logger.info(f"Registered processor {processor_class.__name__} for type {document_type.value}")
@classmethod
def create_processor(cls, file_path: str, config: Dict[str, Any]) -> DocumentProcessor:
"""
Create appropriate processor for the given file.
Args:
file_path: Path to the file to process
config: Configuration dictionary
Returns:
DocumentProcessor instance
Raises:
UnsupportedDocumentTypeError: If no processor is available for the file type
"""
# Detect document type
extension = Path(file_path).suffix.lower()
if extension == '.pdf':
document_type = DocumentType.PDF
elif extension in ['.xlsx', '.xls', '.xlsm']:
document_type = DocumentType.EXCEL
elif extension in ['.png', '.jpg', '.jpeg', '.gif', '.bmp', '.tiff']:
document_type = DocumentType.IMAGE
else:
document_type = DocumentType.UNKNOWN
# Get processor class
processor_class = cls._processors.get(document_type)
if not processor_class:
raise UnsupportedDocumentTypeError(file_path, document_type.value)
# Create and return processor instance
return processor_class(config)
@classmethod
def get_supported_types(cls) -> List[DocumentType]:
"""Get list of supported document types."""
return list(cls._processors.keys())
if __name__=="__main__":
logger.info(f"Docs processor init ..")
# Example usage (for testing purposes)
config = {'max_file_size_mb': 50}
processor = DocumentProcessorFactory.create_processor("example.pdf", config)
processed_doc = processor.process_document("example.pdf")
chunks = processor.extract_chunks(processed_doc)
for chunk in chunks:
print(chunk)