legal-doc-masker/src/models/processors/docx_processor.py

20 lines
687 B
Python

import docx
from models.document_processor import DocumentProcessor
class DocxDocumentProcessor(DocumentProcessor):
def __init__(self, input_path: str, output_path: str):
self.input_path = input_path
self.output_path = output_path
def read_content(self) -> str:
doc = docx.Document(self.input_path)
return '\n'.join([paragraph.text for paragraph in doc.paragraphs])
def process_content(self, content: str) -> str:
# Implementation for processing docx content
return content
def save_content(self, content: str) -> None:
doc = docx.Document()
doc.add_paragraph(content)
doc.save(self.output_path)