ocrd 3.0.0b2__tar.gz → 3.0.0b4__tar.gz

{ocrd-3.0.0b2/src/ocrd.egg-info → ocrd-3.0.0b4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ocrd
-Version: 3.0.0b2
+Version: 3.0.0b4
 Summary: OCR-D framework
 Author-email: Konstantin Baierer <unixprog@gmail.com>
 License: Apache License 2.0

ocrd-3.0.0b4/VERSION

@@ -0,0 +1 @@
+3.0.0b4

{ocrd-3.0.0b2 → ocrd-3.0.0b4}/src/ocrd/cli/__init__.py

@@ -61,11 +61,11 @@ Variables:
 \b
 {config.describe('OCRD_DOWNLOAD_INPUT')}
 \b
-{config.describe('OCRD_MISSING_INPUT')}
+{config.describe('OCRD_MISSING_INPUT', wrap_text=False)}
 \b
-{config.describe('OCRD_MISSING_OUTPUT')}
+{config.describe('OCRD_MISSING_OUTPUT', wrap_text=False)}
 \b
-{config.describe('OCRD_EXISTING_OUTPUT')}
+{config.describe('OCRD_EXISTING_OUTPUT', wrap_text=False)}
 \b
 {config.describe('OCRD_METS_CACHING')}
 \b

{ocrd-3.0.0b2 → ocrd-3.0.0b4}/src/ocrd/processor/base.py

@@ -166,11 +166,14 @@ class Processor():
 
         (Override if ``ocrd-tool.json`` is not distributed with the Python package.)
         """
-        # XXX HACK
-        module_tokens = self.__module__.split('.')
-        if module_tokens[0] == 'src':
-            module_tokens.pop(0)
-        return resource_filename(module_tokens[0], self.metadata_filename)
+        module = inspect.getmodule(self)
+        module_tokens = module.__package__.split('.')
+        # for namespace packages, we cannot just use the first token
+        for i in range(len(module_tokens)):
+            prefix = '.'.join(module_tokens[:i + 1])
+            if sys.modules[prefix].__spec__.has_location:
+                return resource_filename(prefix, self.metadata_filename)
+        raise Exception("cannot find top-level module prefix for %s", module.__package__)
 
     @cached_property
     def metadata_rawdict(self) -> dict:
@@ -455,22 +458,22 @@ class Processor():
                 nr_copied = 0
 
                 # set up multithreading
-                if self.max_workers <= 0:
-                    max_workers = max(0, config.OCRD_MAX_PARALLEL_PAGES)
-                else:
-                    max_workers = max(0, min(config.OCRD_MAX_PARALLEL_PAGES, self.max_workers))
+                max_workers = max(0, config.OCRD_MAX_PARALLEL_PAGES)
+                if self.max_workers > 0 and self.max_workers < config.OCRD_MAX_PARALLEL_PAGES:
+                    self._base_logger.info("limiting number of threads from %d to %d", max_workers, self.max_workers)
+                    max_workers = self.max_workers
                 if max_workers > 1:
                     assert isinstance(workspace.mets, ClientSideOcrdMets), \
                         "OCRD_MAX_PARALLEL_PAGES>1 requires also using --mets-server-url"
-                if self.max_page_seconds <= 0:
-                    max_seconds = max(0, config.OCRD_PROCESSING_PAGE_TIMEOUT)
-                else:
-                    max_seconds = max(0, min(config.OCRD_PROCESSING_PAGE_TIMEOUT, self.max_page_seconds))
+                max_seconds = max(0, config.OCRD_PROCESSING_PAGE_TIMEOUT)
+                if self.max_page_seconds > 0 and self.max_page_seconds < config.OCRD_PROCESSING_PAGE_TIMEOUT:
+                    self._base_logger.info("limiting page timeout from %d to %d sec", max_seconds, self.max_page_seconds)
+                    max_seconds = self.max_page_seconds
                 executor = ThreadPoolExecutor(
                     max_workers=max_workers or 1,
                     thread_name_prefix=f"pagetask.{workspace.mets.unique_identifier}"
                 )
-                self._base_logger.debug("started executor %s", str(executor))
+                self._base_logger.debug("started executor %s with %d workers", str(executor), max_workers or 1)
                 tasks = {}
 
                 for input_file_tuple in self.zip_input_files(on_error='abort', require_first=False):
@@ -478,7 +481,7 @@ class Processor():
                     page_id = next(input_file.pageId
                                    for input_file in input_file_tuple
                                    if input_file)
-                    self._base_logger.info(f"processing page {page_id}")
+                    self._base_logger.info(f"preparing page {page_id}")
                     for i, input_file in enumerate(input_file_tuple):
                         if input_file is None:
                             # file/page not found in this file grp
@@ -521,9 +524,10 @@ class Processor():
                     # broad coverage of output failures (including TimeoutError)
                     except (Exception, TimeoutError) as err:
                         # FIXME: add re-usable/actionable logging
-                        self._base_logger.error(f"Failure on page {page_id}: {str(err) or err.__class__.__name__}")
                         if config.OCRD_MISSING_OUTPUT == 'ABORT':
+                            self._base_logger.error(f"Failure on page {page_id}: {str(err) or err.__class__.__name__}")
                             raise err
+                        self._base_logger.exception(f"Failure on page {page_id}: {str(err) or err.__class__.__name__}")
                         if config.OCRD_MISSING_OUTPUT == 'SKIP':
                             nr_skipped += 1
                             continue
@@ -587,6 +591,7 @@ class Processor():
         input_pcgts : List[Optional[OcrdPage]] = [None] * len(input_files)
         assert isinstance(input_files[0], get_args(OcrdFileType))
         page_id = input_files[0].pageId
+        self._base_logger.info("processing page %s", page_id)
         for i, input_file in enumerate(input_files):
             assert isinstance(input_file, get_args(OcrdFileType))
             self._base_logger.debug(f"parsing file {input_file.ID} for page {page_id}")

{ocrd-3.0.0b2 → ocrd-3.0.0b4}/src/ocrd/workspace.py

@@ -121,7 +121,10 @@ class Workspace():
         """
         Reload METS from the filesystem.
         """
-        self.mets = OcrdMets(filename=self.mets_target)
+        if self.is_remote:
+            self.mets.reload()
+        else:
+            self.mets = OcrdMets(filename=self.mets_target)
 
     @deprecated_alias(pageId="page_id")
     @deprecated_alias(ID="file_id")

{ocrd-3.0.0b2 → ocrd-3.0.0b4/src/ocrd.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ocrd
-Version: 3.0.0b2
+Version: 3.0.0b4
 Summary: OCR-D framework
 Author-email: Konstantin Baierer <unixprog@gmail.com>
 License: Apache License 2.0

{ocrd-3.0.0b2 → ocrd-3.0.0b4}/src/ocrd_utils/config.py

@@ -78,14 +78,26 @@ class OcrdEnvConfig():
             raise ValueError(f"Unregistered env variable {name}")
         return self._variables[name].has_default
 
+    def reset_defaults(self):
+        for name in self._variables:
+            try:
+                # we cannot use hasattr, because that delegates to getattr,
+                # which we override and provide defaults for (which of course
+                # cannot be removed)
+                if self.__getattribute__(name):
+                    delattr(self, name)
+            except AttributeError:
+                pass
+
     def describe(self, name, *args, **kwargs):
         if not name in self._variables:
             raise ValueError(f"Unregistered env variable {name}")
         return self._variables[name].describe(*args, **kwargs)
 
     def __getattr__(self, name):
+        # will be called if name is not accessible (has not been added directly yet)
         if not name in self._variables:
-            raise ValueError(f"Unregistered env variable {name}")
+            raise AttributeError(f"Unregistered env variable {name}")
         var_obj = self._variables[name]
         try:
             raw_value = self.raw_value(name)

ocrd-3.0.0b2/VERSION

@@ -1 +0,0 @@
-3.0.0b2