support path names with spaces, fixes #40

ocrd_browser/model/document.py

@@ -22,7 +22,7 @@
 from pathlib import Path
 from tempfile import mkdtemp
 from datetime import datetime
-from urllib.parse import urlparse
+from urllib.parse import urlparse, unquote
 # noinspection PyProtectedMember
 from lxml.etree import ElementBase as Element, _ElementTree as ElementTree
 
@@ -72,9 +72,9 @@ def load(cls, mets_url: Union[Path, str] = None, emitter: EventCallBack = None)
         """
         if not mets_url:
             return cls.create(emitter=emitter)
-        mets_url = cls._strip_local(mets_url)
+        mets_path = cls._to_path(mets_url)
 
-        workspace = Resolver().workspace_from_url(mets_url, download=False)
+        workspace = Resolver().workspace_from_url(str(mets_path), download=False)
         doc = cls(workspace, emitter=emitter, original_url=mets_url)
         doc._empty = False
         return doc
@@ -110,7 +110,7 @@ def save(self, backup_directory: Union[bool, Path, str] = True) -> None:
 
     def save_as(self, mets_url: Union[Path, str], backup_directory: Union[bool, Path, str] = True) -> None:
         log = getLogger('ocrd_browser.model.document.Document.save_as')
-        mets_path = Path(self._strip_local(mets_url, disallow_remote=True))
+        mets_path = self._to_path(mets_url)
 
         workspace_directory = mets_path.parent
         if workspace_directory.exists():
@@ -496,11 +496,19 @@ def _emit(self, event: str, *args: Any) -> None:
     def _strip_local(mets_url: Union[Path, str], disallow_remote: bool = True) -> str:
         result = urlparse(str(mets_url))
         if result.scheme == 'file' or result.scheme == '':
-            mets_url = result.path
+            mets_url = unquote(result.path)
         elif disallow_remote:
             raise ValueError('invalid url {}'.format(mets_url))
         return str(mets_url)
 
+    @staticmethod
+    def _to_path(mets_url: Union[Path, str]) -> Path:
+        result = urlparse(str(mets_url))
+        if not (result.scheme == 'file' or result.scheme == ''):
+            raise ValueError('invalid local path/url {}'.format(mets_url))
+        return Path(unquote(result.path))
+
+
     @staticmethod
     def _derive_backup_directory(workspace_directory: Path, now: datetime = None) -> Path:
         now = now or datetime.now()

tests/model/test_document.py

@@ -1,7 +1,7 @@
 from pathlib import Path
 from tempfile import TemporaryDirectory
 
-from tests import TestCase, ASSETS_PATH
+from tests import TestCase, ASSETS_PATH, TEST_BASE_PATH
 from ocrd_browser.model import Document, Page
 from datetime import datetime
 from ocrd_models.ocrd_page import PcGtsType
@@ -104,7 +104,7 @@ def test_clone(self):
 
     def test_save(self):
         doc = Document.clone(self.path)
-        with TemporaryDirectory(prefix='browse-ocrd-tests') as directory:
+        with TemporaryDirectory(prefix='browse-ocrd tests') as directory:
             saved_mets = directory + '/mets.xml'
             doc.save_as(saved_mets)
             saved = Document.load(saved_mets)
@@ -172,3 +172,13 @@ def test_modify_when_not_editable(self):
     def test_modify_when_editable(self):
         doc = Document.clone(self.path)
         doc.reorder(['PHYS_0020', 'PHYS_0017'])
+
+    def test_path_with_spaces(self):
+        path = TEST_BASE_PATH / 'example/workspaces/heavy quoting/mets.xml'
+        uri = path.as_uri()
+        doc = Document.load(uri)
+        page = doc.page_for_id('PHYS_0017', 'OCR-D-GT-PAGE')
+        image = doc.workspace.image_from_page(page.page, 'PHYS_0017')
+        # Assert no exceptions happened and a sensible return value
+        self.assertGreater(image[0].height, 100)
+