Merge pull request #2907 from margau/feature-multiple-barcode-scanners

feature: Add support for zxing as barcode scanning lib

Pipfile

@@ -58,6 +58,7 @@ nltk = "*"
 pdf2image = "*"
 flower = "*"
 bleach = "*"
+zxing-cpp = {version = "*", platform_machine = "== 'x86_64'"}
 #
 # Packages locked due to issues (try to check if these are fixed in a release every so often)
 #

Pipfile.lock

@@ -1,7 +1,7 @@
 {
     "_meta": {
         "hash": {
-            "sha256": "01320f2ef2a561c37d17aaad61a7871b5a379dd1ac97fdaab586936b60dec92e"
+            "sha256": "8395f25f876a71a7307a55dd542e69a4cdcb3be3be38c4e89ed06ce3d52a5345"
         },
         "pipfile-spec": 6,
         "requires": {},
@@ -48,7 +48,7 @@
                 "sha256:2163e1640ddb52b7a8c80d0a67a08587e5d245cc9c553a74a847056bc2976b15",
                 "sha256:8ca1e4fcf50d07413d66d1a5e416e42cfdf5851c981d679a09851a6853383b3c"
             ],
-            "markers": "python_version < '3.11'",
+            "markers": "python_full_version <= '3.11.2'",
             "version": "==4.0.2"
         },
         "attrs": {
@@ -1877,7 +1877,7 @@
                 "sha256:5cb5f4a79139d699607b3ef622a1dedafa84e115ab0024e0d9c044a9479ca7cb",
                 "sha256:fb33085c39dd998ac16d1431ebc293a8b3eedd00fd4a32de0ff79002c19511b4"
             ],
-            "markers": "python_version < '3.10'",
+            "markers": "python_version >= '3.7'",
             "version": "==4.5.0"
         },
         "tzdata": {
@@ -2220,6 +2220,29 @@
             ],
             "markers": "python_version >= '3.6'",
             "version": "==0.20.0"
+        },
+        "zxing-cpp": {
+            "hashes": [
+                "sha256:1b67b221aae15aad9b5609d99c38d57875bc0a4fef864142d7ca37e9ee7880b0",
+                "sha256:1d665c45029346c70ae3df5dbc36f6335ffe4f275e98dc43772fa32a65844196",
+                "sha256:214a6a0e49b92fda8d2761c74f5bfd24a677b9bf1d0ef0e083412486af97faa9",
+                "sha256:54282d0e5c573754049113a0cdbf14cc1c6b986432a367d8a788112afa92a1d5",
+                "sha256:5ce391f21763f00d5be3431e16d075e263e4b9205c2cf55d708625cb234b1f15",
+                "sha256:5fd89065f620d6b78281308c6abfb760d95760a1c9b88eb7ac612b52b331bd41",
+                "sha256:631a0c783ad233c85295e0cf4cd7740f1fe2853124c61b1ef6bcf7eb5d2fa5e6",
+                "sha256:76caafb8fc1e12c2e5ec33ce4f340a0e15e9a2aabfbfeaec170e8a2b405b8a77",
+                "sha256:8da9c912cca5829eedb2800ce3eaa1b1e52742f536aa9e798be69bf09639f399",
+                "sha256:95dd06dc559f53c1ca0eb59dbaebd802ebc839937baaf2f8d2b3def3e814c07f",
+                "sha256:97919f07c62edf1c8e0722fd64893057ce636b7067cf47bd593e98cc7e404d74",
+                "sha256:9f0c2c03f5df470ef71a7590be5042161e7590da767d4260a6d0d61a3fa80b88",
+                "sha256:a788551ddf3a6ba1152ff9a0b81d57018a3cc586544087c39d881428745faf1f",
+                "sha256:ea54fd242f93eea7bf039a68287e5e57fdf77d78e3bd5b4cbb2d289bb3380d63",
+                "sha256:f0eefdfad91e15e3f5b7ed16d83806a36f96ca482f4b042baa6297784a58b0b3",
+                "sha256:f70eefa5dc1fd9238087c024ef22f3d99ba79cb932a2c5bc5b0f1e152037722e"
+            ],
+            "index": "pypi",
+            "markers": "platform_machine == 'x86_64'",
+            "version": "==2.0.0"
         }
     },
     "develop": {
@@ -3112,7 +3135,7 @@
                 "sha256:5cb5f4a79139d699607b3ef622a1dedafa84e115ab0024e0d9c044a9479ca7cb",
                 "sha256:fb33085c39dd998ac16d1431ebc293a8b3eedd00fd4a32de0ff79002c19511b4"
             ],
-            "markers": "python_version < '3.10'",
+            "markers": "python_version >= '3.7'",
             "version": "==4.5.0"
         },
         "urllib3": {
@@ -3676,7 +3699,7 @@
                 "sha256:5cb5f4a79139d699607b3ef622a1dedafa84e115ab0024e0d9c044a9479ca7cb",
                 "sha256:fb33085c39dd998ac16d1431ebc293a8b3eedd00fd4a32de0ff79002c19511b4"
             ],
-            "markers": "python_version < '3.10'",
+            "markers": "python_version >= '3.7'",
             "version": "==4.5.0"
         },
         "urllib3": {

docs/configuration.md

@@ -902,6 +902,16 @@ file, which are separated by one or multiple barcode pages.
 
     Defaults to false.
 
+`PAPERLESS_CONSUMER_BARCODE_SCANNER=<string>`
+
+: Sets the barcode scanner used for barcode functionality.
+
+    Currently, "PYZBAR" (the default) or "ZXING" might be selected.
+    If you have problems that your Barcodes/QR-Codes are not detected
+    (especially with bad scan quality and/or small codes), try the other one.
+
+    zxing is not available on all platforms.
+
 `PAPERLESS_CONSUMER_BARCODE_TIFF_SUPPORT=<bool>`
 
 : Whether TIFF image files should be scanned for barcodes. This will

src/documents/barcodes.py

@@ -5,10 +5,12 @@ import tempfile
 from dataclasses import dataclass
 from functools import lru_cache
 from pathlib import Path
+from subprocess import run
 from typing import Dict
 from typing import List
 from typing import Optional
 
+import img2pdf
 import magic
 from django.conf import settings
 from pdf2image import convert_from_path
@@ -16,8 +18,6 @@ from pdf2image.exceptions import PDFPageCountError
 from pikepdf import Page
 from pikepdf import Pdf
 from PIL import Image
-from PIL import ImageSequence
-from pyzbar import pyzbar
 
 logger = logging.getLogger("paperless.barcodes")
 
@@ -83,18 +83,35 @@ def barcode_reader(image: Image) -> List[str]:
     Returns a list containing all found barcodes
     """
     barcodes = []
-    # Decode the barcode image
-    detected_barcodes = pyzbar.decode(image)
 
-    if detected_barcodes:
+    if settings.CONSUMER_BARCODE_SCANNER == "PYZBAR":
-        # Traverse through all the detected barcodes in image
+        logger.debug("Scanning for barcodes using PYZBAR")
+        from pyzbar import pyzbar
+
+        # Decode the barcode image
+        detected_barcodes = pyzbar.decode(image)
+
+        if detected_barcodes:
+            # Traverse through all the detected barcodes in image
+            for barcode in detected_barcodes:
+                if barcode.data:
+                    decoded_barcode = barcode.data.decode("utf-8")
+                    barcodes.append(decoded_barcode)
+                    logger.debug(
+                        f"Barcode of type {str(barcode.type)} found: {decoded_barcode}",
+                    )
+    elif settings.CONSUMER_BARCODE_SCANNER == "ZXING":
+        logger.debug("Scanning for barcodes using ZXING")
+        import zxingcpp
+
+        detected_barcodes = zxingcpp.read_barcodes(image)
         for barcode in detected_barcodes:
-            if barcode.data:
+            if barcode.text:
-                decoded_barcode = barcode.data.decode("utf-8")
+                barcodes.append(barcode.text)
-                barcodes.append(decoded_barcode)
                 logger.debug(
-                    f"Barcode of type {str(barcode.type)} found: {decoded_barcode}",
+                    f"Barcode of type {str(barcode.format)} found: {barcode.text}",
                 )
+
     return barcodes
 
 
@@ -125,21 +142,21 @@ def convert_from_tiff_to_pdf(filepath: Path) -> Path:
             f"Cannot convert mime type {mime_type} from {filepath} to pdf.",
         )
         return None
-    with Image.open(filepath) as image:
+    with Image.open(filepath) as im:
-        images = []
+        has_alpha_layer = im.mode in ("RGBA", "LA")
-        for i, page in enumerate(ImageSequence.Iterator(image)):
+    if has_alpha_layer:
-            page = page.convert("RGB")
+        run(
-            images.append(page)
+            [
-        try:
+                settings.CONVERT_BINARY,
-            if len(images) == 1:
+                "-alpha",
-                images[0].save(newpath)
+                "off",
-            else:
+                filepath,
-                images[0].save(newpath, save_all=True, append_images=images[1:])
+                filepath,
-        except OSError as e:  # pragma: no cover
+            ],
-            logger.warning(
+        )
-                f"Could not save the file as pdf. Error: {str(e)}",
+    with filepath.open("rb") as img_file:
-            )
+        with newpath.open("wb") as pdf_file:
-            return None
+            pdf_file.write(img2pdf.convert(img_file))
     return newpath
 
 

src/documents/tests/test_barcodes.py

@@ -3,6 +3,7 @@ import shutil
 from pathlib import Path
 from unittest import mock
 
+import pytest
 from django.conf import settings
 from django.test import override_settings
 from django.test import TestCase
@@ -13,7 +14,15 @@ from documents.tests.utils import DirectoriesMixin
 from documents.tests.utils import FileSystemAssertsMixin
 from PIL import Image
 
+try:
+    import zxingcpp
 
+    ZXING_AVAILIBLE = True
+except ImportError:
+    ZXING_AVAILIBLE = False
+
+
+@override_settings(CONSUMER_BARCODE_SCANNER="PYZBAR")
 class TestBarcode(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
 
     SAMPLE_DIR = Path(__file__).parent / "samples"
@@ -1030,3 +1039,21 @@ class TestAsnBarcodes(DirectoriesMixin, TestCase):
                 tasks.consume_file,
                 dst,
             )
+
+
+@pytest.mark.skipif(
+    not ZXING_AVAILIBLE,
+    reason="No zxingcpp",
+)
+@override_settings(CONSUMER_BARCODE_SCANNER="ZXING")
+class TestBarcodeZxing(TestBarcode):
+    pass
+
+
+@pytest.mark.skipif(
+    not ZXING_AVAILIBLE,
+    reason="No zxingcpp",
+)
+@override_settings(CONSUMER_BARCODE_SCANNER="ZXING")
+class TestAsnBarcodesZxing(TestAsnBarcodes):
+    pass

src/paperless/checks.py

@@ -166,4 +166,17 @@ def settings_values_check(app_configs, **kwargs):
             )
         return msgs
 
-    return _ocrmypdf_settings_check() + _timezone_validate()
+    def _barcode_scanner_validate():
+        """
+        Validates the barcode scanner type
+        """
+        msgs = []
+        if settings.CONSUMER_BARCODE_SCANNER not in ["PYZBAR", "ZXING"]:
+            msgs.append(
+                Error(f'Invalid Barcode Scanner "{settings.CONSUMER_BARCODE_SCANNER}"'),
+            )
+        return msgs
+
+    return (
+        _ocrmypdf_settings_check() + _timezone_validate() + _barcode_scanner_validate()
+    )

src/paperless/settings.py

@@ -732,6 +732,12 @@ CONSUMER_BARCODE_STRING: Final[str] = os.getenv(
     "PATCHT",
 )
 
+consumer_barcode_scanner_tmp: Final[str] = os.getenv(
+    "PAPERLESS_CONSUMER_BARCODE_SCANNER",
+    "PYZBAR",
+)
+CONSUMER_BARCODE_SCANNER = consumer_barcode_scanner_tmp.upper()
+
 CONSUMER_ENABLE_ASN_BARCODE: Final[bool] = __get_boolean(
     "PAPERLESS_CONSUMER_ENABLE_ASN_BARCODE",
 )

src/paperless/tests/test_checks.py

@@ -176,3 +176,26 @@ class TestSettingsChecks(DirectoriesMixin, TestCase):
         msg = msgs[0]
 
         self.assertIn('Timezone "TheMoon\\MyCrater"', msg.msg)
+
+    @override_settings(CONSUMER_BARCODE_SCANNER="Invalid")
+    def test_barcode_scanner_invalid(self):
+        msgs = settings_values_check(None)
+        self.assertEqual(len(msgs), 1)
+
+        msg = msgs[0]
+
+        self.assertIn('Invalid Barcode Scanner "Invalid"', msg.msg)
+
+    @override_settings(CONSUMER_BARCODE_SCANNER="")
+    def test_barcode_scanner_empty(self):
+        msgs = settings_values_check(None)
+        self.assertEqual(len(msgs), 1)
+
+        msg = msgs[0]
+
+        self.assertIn('Invalid Barcode Scanner ""', msg.msg)
+
+    @override_settings(CONSUMER_BARCODE_SCANNER="PYZBAR")
+    def test_barcode_scanner_valid(self):
+        msgs = settings_values_check(None)
+        self.assertEqual(len(msgs), 0)