almet/dangerzone
1
0
Fork 0
mirror of https://github.com/freedomofpress/dangerzone.git synced 2025-04-29 10:12:38 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 11

qubes: Pass OCR parameters properly

Browse source 
Pass OCR parameters to conversion functions as arguments, instead of
setting environment variables.

Fixes #455
This commit is contained in:
Alex Pyrgiotis 2023-09-07 14:51:25 +03:00
parent 8a0c0a4673
commit e64d1da61f
No known key found for this signature in database
GPG key ID: B6C15EBA0357C9AA
2 changed files with 9 additions and 15 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
dangerzone/conversion/pixels_to_pdf.py
View file

@ -11,12 +11,13 @@ import json
import os
import shutil
import sys
from typing import Optional
from .common import DangerzoneConverter, running_on_qubes
class PixelsToPDF(DangerzoneConverter):
async def convert(self) -> None:
async def convert(self, ocr_lang: Optional[str] = None) -> None:
self.percentage = 50.0
num_pages = len(glob.glob("/tmp/dangerzone/page-*.rgb"))
@ -43,7 +44,7 @@ class PixelsToPDF(DangerzoneConverter):
total_size += page_size
timeout = self.calculate_timeout(page_size, 1)
if os.environ.get("OCR") == "1": # OCR the document
if ocr_lang: # OCR the document
self.update_progress(
f"Converting page {page}/{num_pages} from pixels to searchable PDF"
)
@ -71,7 +72,7 @@ class PixelsToPDF(DangerzoneConverter):
png_filename,
ocr_filename,
"-l",
os.environ.get("OCR_LANGUAGE"), # type: ignore
ocr_lang,
"--dpi",
"70",
"pdf",
@ -152,10 +153,11 @@ class PixelsToPDF(DangerzoneConverter):
async def main() -> int:
ocr_lang = os.environ.get("OCR_LANGUAGE") if os.environ.get("OCR") == "1" else None
converter = PixelsToPDF()
try:
await converter.convert()
await converter.convert(ocr_lang)
error_code = 0 # Success!
except (RuntimeError, TimeoutError, ValueError) as e:

14
dangerzone/isolation_provider/qubes.py
View file

@ -176,25 +176,17 @@ class Qubes(IsolationProvider):
f"Conversion output (doc to pixels)\n{self.sanitize_conversion_str(untrusted_log)}"
)
# FIXME pass OCR stuff properly (see #455)
old_environ = dict(os.environ)
if ocr_lang:
os.environ["OCR"] = "1"
os.environ["OCR_LANGUAGE"] = ocr_lang
def print_progress_wrapper(error: bool, text: str, percentage: float) -> None:
self.print_progress_trusted(document, error, text, percentage)
asyncio.run(PixelsToPDF(progress_callback=print_progress_wrapper).convert())
asyncio.run(
PixelsToPDF(progress_callback=print_progress_wrapper).convert(ocr_lang)
)
percentage = 100.0
text = "Safe PDF created"
self.print_progress_trusted(document, False, text, percentage)
# FIXME remove once the OCR args are no longer passed with env vars
os.environ.clear()
os.environ.update(old_environ)
shutil.move(CONVERTED_FILE_PATH, document.output_filename)
success = True