ocrd 3.0.1__tar.gz → 3.0.3__tar.gz

{ocrd-3.0.1/src/ocrd.egg-info → ocrd-3.0.3}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: ocrd
-Version: 3.0.1
+Version: 3.0.3
 Summary: OCR-D framework
 Author-email: Konstantin Baierer <unixprog@gmail.com>
 License: Apache License 2.0

ocrd-3.0.3/VERSION

@@ -0,0 +1 @@
+3.0.3

{ocrd-3.0.1 → ocrd-3.0.3}/src/ocrd/decorators/ocrd_cli_options.py

@@ -13,16 +13,16 @@ from ocrd_network import (
 
 def ocrd_cli_options(f):
     """
-    Implement MP CLI.
+    Implement Processor CLI.
 
     Usage::
 
-        import ocrd_click_cli from ocrd.utils
+        from ocrd.decorators import ocrd_cli_options
 
         @click.command()
-        @ocrd_click_cli
-        def cli(mets_url):
-            print(mets_url)
+        @ocrd_cli_options
+        def cli(**kwargs):
+            print(kwargs['mets_url'])
     """
     # XXX Note that the `--help` output is statically generate_processor_help
     params = [

{ocrd-3.0.1 → ocrd-3.0.3}/src/ocrd/processor/base.py

@@ -518,7 +518,7 @@ class Processor():
 
                 if max_workers > 1:
                     executor_cls = ProcessPoolExecutor
-                    log_queue = mp.Queue()
+                    log_queue = mp.get_context('fork').Queue()
                 else:
                     executor_cls = DummyExecutor
                     log_queue = None
@@ -534,15 +534,20 @@ class Processor():
                     # forward messages from log queue (in subprocesses) to all root handlers
                     log_listener = logging.handlers.QueueListener(log_queue, *logging.root.handlers, respect_handler_level=True)
                     log_listener.start()
+                tasks = None
                 try:
                     self._base_logger.debug("started executor %s with %d workers", str(executor), max_workers or 1)
                     tasks = self.process_workspace_submit_tasks(executor, max_seconds)
                     stats = self.process_workspace_handle_tasks(tasks)
                 finally:
                     executor.shutdown(kill_workers=True, wait=False)
+                    self._base_logger.debug("stopped executor %s after %d tasks", str(executor), len(tasks) if tasks else -1)
                     if max_workers > 1:
-                        log_listener.stop()
-                        del log_listener
+                        # can cause deadlock:
+                        #log_listener.stop()
+                        # not much better:
+                        #log_listener.enqueue_sentinel()
+                        pass
 
             except NotImplementedError:
                 # fall back to deprecated method
@@ -670,11 +675,12 @@ class Processor():
                 nr_succeeded += 1
             # else skipped - already exists
         nr_errors = dict(nr_errors)
+        nr_all = nr_succeeded + nr_failed
         if nr_failed > 0:
-            nr_all = nr_succeeded + nr_failed
             if config.OCRD_MAX_MISSING_OUTPUTS > 0 and nr_failed / nr_all > config.OCRD_MAX_MISSING_OUTPUTS:
                 raise Exception(f"too many failures with {reason} output ({nr_failed} of {nr_all}, {str(nr_errors)})")
             self._base_logger.warning("%s %d of %d pages due to %s", reason, nr_failed, nr_all, str(nr_errors))
+        self._base_logger.debug("succeeded %d, missed %d of %d pages due to %s", nr_succeeded, nr_failed, nr_all, str(nr_errors))
         return nr_succeeded, nr_failed, nr_errors, len(tasks)
 
     def process_workspace_handle_page_task(self, page_id : str, input_files : List[Optional[OcrdFileType]], task : TFuture) -> Union[bool, Exception]:

{ocrd-3.0.1 → ocrd-3.0.3}/src/ocrd/processor/helpers.py

@@ -56,13 +56,19 @@ def run_processor(
     - :py:attr:`output_file_grp`
     - :py:attr:`parameter` (after applying any :py:attr:`parameter_override` settings)
 
-    Warning: Avoid setting the `instance_caching` flag to True. It may have unexpected side effects.
-    This flag is used for an experimental feature we would like to adopt in future.
-
     Run the processor on the workspace (creating output files in the filesystem).
 
     Finally, write back the workspace (updating the METS in the filesystem).
 
+    If :py:attr:`instance_caching` is True, then processor instances (for the same set
+    of :py:attr:`parameter` values) will be cached internally. Thus, these objects (and
+    all their memory resources, like loaded models) get re-used instead of re-instantiated
+    when a match occurs - as long as the program is being run. They only get deleted (and
+    their resources freed) when as many as :py:data:`~ocrd_utils.config.OCRD_MAX_PROCESSOR_CACHE`
+    instances have already been cached while this particular parameter set was re-used
+    least frequently. (See :py:class:`~ocrd_network.ProcessingWorker` and
+    :py:class:`~ocrd_network.ProcessorServer` for use-cases.)
+
     Args:
         processorClass (object): Python class of the module processor.
     """

{ocrd-3.0.1 → ocrd-3.0.3}/src/ocrd/workspace.py

@@ -798,7 +798,6 @@ class Workspace():
             raise Exception('Found no AlternativeImage that satisfies all requirements ' +
                             'filter="%s" in page "%s"' % (
                                 feature_filter, page_id))
-        page_image.format = 'PNG' # workaround for tesserocr#194
         # ensure DPI will be set in image meta-data again
         if 'DPI' in page_coords:
             dpi = page_coords['DPI']
@@ -1060,7 +1059,6 @@ class Workspace():
             raise Exception('Found no AlternativeImage that satisfies all requirements ' +
                             'filter="%s" in segment "%s"' % (
                                 feature_filter, segment.id))
-        segment_image.format = 'PNG' # workaround for tesserocr#194
         # ensure DPI will be set in image meta-data again
         if 'DPI' in segment_coords:
             dpi = segment_coords['DPI']

{ocrd-3.0.1 → ocrd-3.0.3/src/ocrd.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: ocrd
-Version: 3.0.1
+Version: 3.0.3
 Summary: OCR-D framework
 Author-email: Konstantin Baierer <unixprog@gmail.com>
 License: Apache License 2.0

{ocrd-3.0.1 → ocrd-3.0.3}/src/ocrd_utils/config.py

@@ -142,7 +142,7 @@ config.add('OCRD_MAX_PROCESSOR_CACHE',
     default=(True, 128))
 
 config.add('OCRD_MAX_PARALLEL_PAGES',
-    description="Maximum number of processor threads for page-parallel processing (within each Processor's selected page range, independent of the number of Processing Workers or Processor Servers). If set >1, then a METS Server must be used for METS synchronisation.",
+    description="Maximum number of processor workers for page-parallel processing (within each Processor's selected page range, independent of the number of Processing Workers or Processor Servers). If set >1, then a METS Server must be used for METS synchronisation.",
     parser=int,
     default=(True, 1))
 

ocrd-3.0.1/VERSION

@@ -1 +0,0 @@
-3.0.1

{ocrd-3.0.1 → ocrd-3.0.3}/src/ocrd_network/cli/client.py

@@ -110,8 +110,8 @@ def check_processing_job_status(address: Optional[str], processing_job_id: str):
 @click.option('-p', '--print-state', default=False, is_flag=True,
               help='If set, the client will print job states by each iteration.')
 def send_processing_job_request(
-    address: Optional[str],
     processor_name: str,
+    address: Optional[str],
     mets: str,
     input_file_grp: str,
     output_file_grp: Optional[str],