ocrd 3.0.1__py3-none-any.whl → 3.0.3__py3-none-any.whl

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/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/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/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.dist-info → ocrd-3.0.3.dist-info}/METADATA

@@ -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
@@ -24,13 +24,15 @@ Requires-Dist: Flask
 Requires-Dist: frozendict>=2.3.4
 Requires-Dist: gdown
 Requires-Dist: httpx>=0.22.0
+Requires-Dist: importlib_metadata; python_version < "3.8"
+Requires-Dist: importlib_resources; python_version < "3.10"
 Requires-Dist: jsonschema>=4
 Requires-Dist: loky
 Requires-Dist: lxml
 Requires-Dist: memory-profiler>=0.58.0
 Requires-Dist: numpy
 Requires-Dist: ocrd-fork-bagit>=1.8.1.post2
-Requires-Dist: ocrd-fork-bagit-profile>=1.3.0.post1
+Requires-Dist: ocrd-fork-bagit_profile>=1.3.0.post1
 Requires-Dist: opencv-python-headless
 Requires-Dist: paramiko
 Requires-Dist: pika>=1.2.0
@@ -40,12 +42,10 @@ Requires-Dist: python-magic
 Requires-Dist: python-multipart
 Requires-Dist: pyyaml
 Requires-Dist: requests
-Requires-Dist: requests-unixsocket2
+Requires-Dist: requests_unixsocket2
 Requires-Dist: shapely
 Requires-Dist: uvicorn
 Requires-Dist: uvicorn>=0.17.6
-Requires-Dist: importlib-resources; python_version < "3.10"
-Requires-Dist: importlib-metadata; python_version < "3.8"
 
 # OCR-D/core
 

{ocrd-3.0.1.dist-info → ocrd-3.0.3.dist-info}/RECORD

@@ -7,7 +7,7 @@ ocrd/resolver.py,sha256=Ba9ALQbTXz6_mla4VqN9tAfHoj6aKuNJAU4tIDnjcHE,14952
 ocrd/resource_list.yml,sha256=82-PiqkZnka1kTj3MQqNn4wXWKHHtoFchsQuetWuqFs,2633
 ocrd/resource_manager.py,sha256=kIWDoKxWH4IJE1gcoTcCRQjYjieCqiQclyuyF6Y9b8A,16813
 ocrd/task_sequence.py,sha256=spiaUQaMM7M8WdBDoQGmLuTPm7tOugYXD6rcJ2UXzxw,6991
-ocrd/workspace.py,sha256=cedqK7es2i2nwQCiUiVyWk3j4-nH7bsi6TF7v8siTio,65794
+ocrd/workspace.py,sha256=-j3X83K0f4vtd5jwfu6_R53RJ2R8gt1HYpyrv8YP2bg,65661
 ocrd/workspace_backup.py,sha256=iab_JjZ_mMP-G8NIUk4PZmfpNlQuGRoqc3NbTSSew1w,3621
 ocrd/workspace_bagger.py,sha256=yU8H3xR5WmQKvgQewac71ie-DUWcfLnMS01D55zsEHQ,11971
 ocrd/cli/__init__.py,sha256=-BiwIakeCkWx0Jd2yX9_ahfdV4VKz_5yqGEJ_2zKakQ,2734
@@ -23,11 +23,11 @@ ocrd/cli/zip.py,sha256=MMJLw3OXWiJVfVtrdJcBkbB8vA1IzSautluazZRuCQ0,5910
 ocrd/decorators/__init__.py,sha256=PyXX7vxdWkRHixas9dWUtyO3YLczcly8ZEpfZDSMVp8,7639
 ocrd/decorators/loglevel_option.py,sha256=tgipROEu3t4hkwWvFssd80k2SbTBwBIC4WNE6Gc-XAg,798
 ocrd/decorators/mets_find_options.py,sha256=d4oATKMP6bFQHNqOK6nLqgUiWF2FYdkPvzkTVRMYpKo,635
-ocrd/decorators/ocrd_cli_options.py,sha256=hr2EugwAY_-GJ7F7g77Od9o9eAqhfLBHSpfmCql2OCU,2665
+ocrd/decorators/ocrd_cli_options.py,sha256=lIvtE8re1VmpHm45u71ltE0QJS8nyd28HhLC7zGSvlo,2691
 ocrd/decorators/parameter_option.py,sha256=n8hYw7XVTd3i3tvpK8F1Jx_CqRp6EGF9qJVH95yj92Q,1076
 ocrd/processor/__init__.py,sha256=39ymNwYRdc-b_OJzzKmWCvo2ga3KdsGSYDHE1Hzkn_w,274
-ocrd/processor/base.py,sha256=rVTQeUpZR_rBDh44Q7Xtl5TGcEdyBhMTDfpAgx4eLPg,59295
-ocrd/processor/helpers.py,sha256=8ngrqAJ01BSoSJNsIoK_YfA8QdryM5y0MqaqA9f7ELM,10483
+ocrd/processor/base.py,sha256=wAu3d5E2Vt-rWBejzgZOOnxuI0njgShuLfRZPxyuUGk,59731
+ocrd/processor/helpers.py,sha256=rk7OYucvUgS0fTsxNpYaiPmuyD2l0hVg9uSXb4Jdgow,10975
 ocrd/processor/ocrd_page_result.py,sha256=eDkpyVHcpaBzTHXiGrcNk9PP9Xr-XZru2w_uoX_ZeNA,510
 ocrd/processor/builtin/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ocrd/processor/builtin/dummy_processor.py,sha256=a-4kKJ1JeXQuBIyyN8w2R3s7ov-wAfyEdEz3nxrf0sU,3479
@@ -64,7 +64,7 @@ ocrd_network/server_utils.py,sha256=Uge5F2VagPAEpcyU_Qf8AiecObIGXE0ilD8DaK7bTdE,
 ocrd_network/tcp_to_uds_mets_proxy.py,sha256=yRW-O6ihd31gf7xqQBIBb_ZQQgqisMyOdRI216ehq_A,3160
 ocrd_network/utils.py,sha256=XzPXeSPCVjWLQM540PCpxfJ5hqjJ85_OQBjnf9HlDtE,6759
 ocrd_network/cli/__init__.py,sha256=F7YVqxw-9glz6-ghG0Kp5XXeV1-rL1emVSXLCWxdTF0,306
-ocrd_network/cli/client.py,sha256=XYWbeSiPK4BQXuyTq_FTOXEKljXVLkukWfx07aKbthY,8424
+ocrd_network/cli/client.py,sha256=gFEXjz-d074CpvimqaM4kJRbJVNYRAOK-jsUl2EAUVs,8424
 ocrd_network/cli/processing_server.py,sha256=rAci6RsHlZ0c87GuLdfdCQCiGNcDEu4NEEQiwKJqVUo,796
 ocrd_network/cli/processing_worker.py,sha256=ZuaCkbKV_WKJV7cGOjZ6RLrjjppymnwNCiznFMlclAg,1897
 ocrd_network/cli/processor_server.py,sha256=Vto7UekFo_g83aHqwDmhge9bhPzk0b7O-L46dSfIpJc,1259
@@ -89,7 +89,7 @@ ocrd_network/runtime_data/hosts.py,sha256=ml19ptzH4TFofyJR-Qp_Mn3sZUFbWoNe__rRXZ
 ocrd_network/runtime_data/network_agents.py,sha256=5p_zKLqECBIHLw-Ya6eKcKSZcUM4ESiipEIphVxHBEA,5192
 ocrd_network/runtime_data/network_services.py,sha256=xrPpFUU_Pa-XzGe2FEt5RmO17xqykIUmTr_9g6S7XSs,7892
 ocrd_utils/__init__.py,sha256=U_zAQJwxg_aJ4CR84CKMNAUP6Cob8Er8Ikj42JmnUKo,5977
-ocrd_utils/config.py,sha256=smjUAGK5n0iKZCs4dZAtrZurelcaavlFqhIUJWNMOi0,11796
+ocrd_utils/config.py,sha256=1_8j4kpKK5gxhFKObbBEzmq4JazTtKEkGe14Ch_tpw0,11796
 ocrd_utils/constants.py,sha256=ImbG1d8t2MW3uuFi-mN6aY90Zn74liAKZBKlfuKN86w,3278
 ocrd_utils/deprecate.py,sha256=4i50sZsA3Eevqn5D-SL5yGf9KEZfGCV4A5Anzn1GRMs,1026
 ocrd_utils/image.py,sha256=zNNX1cnRy6yvrxx8mnYQiqWraAh5-i4a1AOfCCg4SmI,24781
@@ -120,9 +120,9 @@ ocrd_validators/xlink.xsd,sha256=8fW7YAMWXN2PbB_MMvj9H5ZeFoEBDzuYBtlGC8_6ijw,318
 ocrd_validators/xsd_mets_validator.py,sha256=4GWfLyqkmca0x7osDuXuExYuM0HWVrKoqn0S35sFhHU,467
 ocrd_validators/xsd_page_validator.py,sha256=BNz_9u-Ek4UCeyZu3KxSQoolfW9lvuaSR9nIu1XXxeE,467
 ocrd_validators/xsd_validator.py,sha256=6HrVAf6SzCvfUIuQdIzz9bOq4V-zhyii9yrUPoK2Uvo,2094
-ocrd-3.0.1.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-ocrd-3.0.1.dist-info/METADATA,sha256=y6mdBSjKmkTMSJ7F1LAvjCWeagTBRAeiYaka9_Z2Djc,10442
-ocrd-3.0.1.dist-info/WHEEL,sha256=P9jw-gEje8ByB7_hXoICnHtVCrEwMQh-630tKvQWehc,91
-ocrd-3.0.1.dist-info/entry_points.txt,sha256=4hcJ2LkK_OlIabHnKgFit35Ap7b5Lz1Gb4hzkxV0Kiw,152
-ocrd-3.0.1.dist-info/top_level.txt,sha256=pUgiN42t4KXC5rvpi6V8atza31XP4SCznXpXlVlvomM,75
-ocrd-3.0.1.dist-info/RECORD,,
+ocrd-3.0.3.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+ocrd-3.0.3.dist-info/METADATA,sha256=gZiKn6AO1tPL9NAfVsX0f3UTUZipDtEUOIXmMLXr76k,10442
+ocrd-3.0.3.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+ocrd-3.0.3.dist-info/entry_points.txt,sha256=4hcJ2LkK_OlIabHnKgFit35Ap7b5Lz1Gb4hzkxV0Kiw,152
+ocrd-3.0.3.dist-info/top_level.txt,sha256=pUgiN42t4KXC5rvpi6V8atza31XP4SCznXpXlVlvomM,75
+ocrd-3.0.3.dist-info/RECORD,,

{ocrd-3.0.1.dist-info → ocrd-3.0.3.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.3.0)
+Generator: setuptools (75.8.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

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],

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))