Port to OCR-D/core API v3

.dockerignore

@@ -0,0 +1,5 @@
+src/dinglehopper/tests
+dist
+build
+*.egg-info
+.git

Dockerfile

@@ -6,17 +6,30 @@ LABEL \
     maintainer="https://github.com/qurator-spk/dinglehopper/issues" \
     org.label-schema.vcs-ref=$VCS_REF \
     org.label-schema.vcs-url="https://github.com/qurator-spk/dinglehopper" \
-    org.label-schema.build-date=$BUILD_DATE
+    org.label-schema.build-date=$BUILD_DATE \
+    org.opencontainers.image.vendor="DFG-Funded Initiative for Optical Character Recognition Development" \
+    org.opencontainers.image.title="dinglehopper" \
+    org.opencontainers.image.description="The OCR evaluation tool" \
+    org.opencontainers.image.source="https://github.com/qurator-spk/dinglehopper" \
+    org.opencontainers.image.documentation="https://github.com/qurator-spk/dinglehopper/blob/${VCS_REF}/README.md" \
+    org.opencontainers.image.revision=$VCS_REF \
+    org.opencontainers.image.created=$BUILD_DATE \
+    org.opencontainers.image.base.name=ocrd/core
+
+ENV LANG=C.UTF-8
+ENV LC_ALL=C.UTF-8
+
+# avoid HOME/.local/share (hard to predict USER here)
+# so let XDG_DATA_HOME coincide with fixed system location
+# (can still be overridden by derived stages)
+ENV XDG_DATA_HOME /usr/local/share
+# avoid the need for an extra volume for persistent resource user db
+# (i.e. XDG_CONFIG_HOME/ocrd/resources.yml)
+ENV XDG_CONFIG_HOME /usr/local/share/ocrd-resources
 
 WORKDIR /build/dinglehopper
-COPY pyproject.toml .
-COPY src/dinglehopper/ocrd-tool.json .
-COPY src ./src
-COPY requirements.txt .
-COPY README.md .
-COPY Makefile .
-RUN make install
-RUN rm -rf /build/dinglehopper
+COPY . .
+RUN make install && rm -rf /build/dinglehopper
 
 WORKDIR /data
-VOLUME ["/data"]
+VOLUME /data

Makefile

@@ -1,8 +1,9 @@
 PYTHON = python3
 PIP = pip3
 PYTHONIOENCODING=utf8
+PYTEST_ARGS = -vv
 
-DOCKER_BASE_IMAGE = docker.io/ocrd/core:v2.69.0
+DOCKER_BASE_IMAGE = docker.io/ocrd/core:v3.3.0
 DOCKER_TAG = ocrd/dinglehopper
 
 help:
@@ -16,6 +17,12 @@ help:
 install:
 	$(PIP) install .
 
+install-dev:
+	$(PIP) install -e .
+
+test:
+	pytest $(PYTEST_ARGS)
+
 docker:
 	docker build \
 	--build-arg DOCKER_BASE_IMAGE=$(DOCKER_BASE_IMAGE) \

src/dinglehopper/ocrd-tool.json

@@ -1,17 +1,13 @@
 {
   "version": "0.9.7",
   "git_url": "https://github.com/qurator-spk/dinglehopper",
+  "dockerhub": "ocrd/dinglehopper",
   "tools": {
     "ocrd-dinglehopper": {
       "executable": "ocrd-dinglehopper",
+      "input_file_grp_cardinality": 2,
+      "output_file_grp_cardinality": 1,
       "description": "Evaluate OCR text against ground truth with dinglehopper",
-      "input_file_grp": [
-        "OCR-D-GT-PAGE",
-        "OCR-D-OCR"
-      ],
-      "output_file_grp": [
-        "OCR-D-OCR-EVAL"
-      ],
       "categories": [
         "Quality assurance"
       ],

src/dinglehopper/ocrd_cli.py

@@ -1,83 +1,73 @@
-import json
+from functools import cached_property
 import os
+from typing import Optional
 
 import click
-import importlib_resources
+from ocrd_models import OcrdFileType
 from ocrd import Processor
 from ocrd.decorators import ocrd_cli_options, ocrd_cli_wrap_processor
-from ocrd_utils import assert_file_grp_cardinality, getLogger, make_file_id
+from ocrd_utils import make_file_id
 
 from .cli import process as cli_process
 
-OCRD_TOOL = json.loads(
-    importlib_resources.files(__name__)
-    .joinpath("ocrd-tool.json")
-    .read_text(encoding="utf-8", errors="strict")
-)
-
-
 @click.command()
 @ocrd_cli_options
 def ocrd_dinglehopper(*args, **kwargs):
     return ocrd_cli_wrap_processor(OcrdDinglehopperEvaluate, *args, **kwargs)
 
-
 class OcrdDinglehopperEvaluate(Processor):
-    def __init__(self, *args, **kwargs):
-        kwargs["ocrd_tool"] = OCRD_TOOL["tools"]["ocrd-dinglehopper"]
-        kwargs["version"] = OCRD_TOOL["version"]
-        super(OcrdDinglehopperEvaluate, self).__init__(*args, **kwargs)
 
-    def process(self):
-        assert_file_grp_cardinality(self.input_file_grp, 2, "GT and OCR")
-        assert_file_grp_cardinality(self.output_file_grp, 1)
+    @cached_property
+    def executable(self):
+        return 'ocrd-dinglehopper'
 
-        log = getLogger("processor.OcrdDinglehopperEvaluate")
+    def process_page_file(self, *input_files: Optional[OcrdFileType]) -> None:
 
+        assert self.parameter
         metrics = self.parameter["metrics"]
         textequiv_level = self.parameter["textequiv_level"]
-        gt_grp, ocr_grp = self.input_file_grp.split(",")
-
-        input_file_tuples = self.zip_input_files(on_error="abort")
-        for n, (gt_file, ocr_file) in enumerate(input_file_tuples):
-            if not gt_file or not ocr_file:
-                # file/page was not found in this group
-                continue
-            gt_file = self.workspace.download_file(gt_file)
-            ocr_file = self.workspace.download_file(ocr_file)
-            page_id = gt_file.pageId
 
-            log.info("INPUT FILES %i / %s↔ %s", n, gt_file, ocr_file)
-
-            file_id = make_file_id(ocr_file, self.output_file_grp)
-            report_prefix = os.path.join(self.output_file_grp, file_id)
-
-            # Process the files
-            try:
-                os.mkdir(self.output_file_grp)
-            except FileExistsError:
-                pass
-            cli_process(
-                gt_file.local_filename,
-                ocr_file.local_filename,
-                report_prefix,
-                metrics=metrics,
-                textequiv_level=textequiv_level,
+        try:
+            gt_file, ocr_file = input_files
+            assert gt_file, 'missing GT file'
+            assert ocr_file, 'missing OCR file'
+            assert gt_file.local_filename
+            assert ocr_file.local_filename
+        except (ValueError, AssertionError) as err:
+            self.logger.warning(f'Missing either GT file, OCR file or both: {err}') # TODO how to log which page?
+            return
+
+        page_id = gt_file.pageId
+
+        file_id = make_file_id(ocr_file, self.output_file_grp)
+        report_prefix = os.path.join(self.output_file_grp, file_id)
+
+        # Process the files
+        try:
+            os.mkdir(self.output_file_grp)
+        except FileExistsError:
+            pass
+        cli_process(
+            gt_file.local_filename,
+            ocr_file.local_filename,
+            report_prefix,
+            metrics=metrics,
+            textequiv_level=textequiv_level,
+        )
+
+        # Add reports to the workspace
+        for report_suffix, mimetype in [
+            [".html", "text/html"],
+            [".json", "application/json"],
+        ]:
+            self.workspace.add_file(
+                file_id=file_id + report_suffix,
+                file_grp=self.output_file_grp,
+                page_id=page_id,
+                mimetype=mimetype,
+                local_filename=report_prefix + report_suffix,
             )
 
-            # Add reports to the workspace
-            for report_suffix, mimetype in [
-                [".html", "text/html"],
-                [".json", "application/json"],
-            ]:
-                self.workspace.add_file(
-                    file_id=file_id + report_suffix,
-                    file_grp=self.output_file_grp,
-                    page_id=page_id,
-                    mimetype=mimetype,
-                    local_filename=report_prefix + report_suffix,
-                )
-
 
 if __name__ == "__main__":
     ocrd_dinglehopper()