smoke test, circle ci

Conflicts:
	Makefile
	ocrd_calamari/__init__.py

.circleci/config.yml

@@ -0,0 +1,21 @@
+version: 2.1
+orbs:
+  codecov: codecov/codecov@1.0.5
+
+jobs:
+
+  build-python36:
+    docker:
+      - image: ubuntu:18.04
+    steps:
+      - run: apt-get update ; apt-get install -y make git curl
+      - checkout
+      - run: make install
+      - run: pip install -r requirements-test.txt
+      - run: make coverage
+      - codecov/upload
+
+workflows:
+  build:
+    jobs:
+      - build-python36

.gitignore

@@ -102,7 +102,9 @@ venv.bak/
 
 # mypy
 .mypy_cache/
+
+
 /calamari
 /calamari_models
 /repo
-/test
+/test/assets

Makefile

@@ -1,7 +1,11 @@
-GIT_CLONE = git clone --depth 1
+# '$(PYTHON)'
+PYTHON = python
+
+# '$(PIP_INSTALL)'
+PIP_INSTALL = pip install
 
-# Docker tag
-DOCKER_TAG = ocrd/calamari
+# '$(GIT_CLONE)'
+GIT_CLONE = git clone --depth 1
 
 # BEGIN-EVAL makefile-parser --make-help Makefile
 
@@ -9,31 +13,75 @@ help:
 	@echo ""
 	@echo "  Targets"
 	@echo ""
-	@echo "    calamari         git clone calamari"
-	@echo "    calamari_models  git clone calamari_models"
-	@echo "    calamari/build   Install calamari"
-	@echo "    docker           Build docker image"
+	@echo "    install          Install ocrd_calamari"
+	@echo "    calamari         Clone calamari repo"
+	@echo "    calamari_models  Clone calamari_models repo"
+	@echo "    calamari/build   pip install calamari"
+	@echo "    deps-test        Install testing python deps via pip"
+	@echo "    repo/assets      Clone OCR-D/assets to ./repo/assets"
+	@echo "    test/assets      Setup test assets"
+	@echo "    assets-clean     Remove symlinks in test/assets"
+	@echo "    test             Run unit tests"
+	@echo "    coverage         Run unit tests and determine test coverage"
 	@echo ""
 	@echo "  Variables"
 	@echo ""
-	@echo "    DOCKER_TAG  Docker tag"
+	@echo "    PYTHON       '$(PYTHON)'"
+	@echo "    PIP_INSTALL  '$(PIP_INSTALL)'"
+	@echo "    GIT_CLONE    '$(GIT_CLONE)'"
 
 # END-EVAL
 
-# git clone calamari
+# Install ocrd_calamari
+install:
+	$(PIP_INSTALL) .
+
+# Clone calamari repo
 calamari:
 	$(GIT_CLONE) https://github.com/chwick/calamari
 
-# git clone calamari_models
+# Clone calamari_models repo
 calamari_models:
 	$(GIT_CLONE) https://github.com/chwick/calamari_models
 
-# Install calamari
+# pip install calamari
 calamari/build: calamari calamari_models
-	cd calamari &&\
-		pip install -r requirements.txt ;\
-		python setup.py install
+	cd calamari && $(PIP_INSTALL) .
+
+#
+# Assets and Tests
+#
+
+# Install testing python deps via pip
+deps-test:
+	$(PIP) install -r requirements_test.txt
+
+
+# Clone OCR-D/assets to ./repo/assets
+repo/assets:
+	mkdir -p $(dir $@)
+	git clone https://github.com/OCR-D/assets "$@"
+
+
+# Setup test assets
+test/assets: repo/assets
+	mkdir -p $@
+	cp -r -t $@ repo/assets/data/*
+
+# Remove symlinks in test/assets
+assets-clean:
+	rm -rf test/assets
+
+# Run unit tests
+test: test/assets calamari_models
+	# declare -p HTTP_PROXY
+	$(PYTHON) -m pytest --continue-on-collection-errors test $(PYTEST_ARGS)
+
+# Run unit tests and determine test coverage
+coverage:
+	coverage erase
+	make test PYTHON="coverage run"
+	coverage report
+	coverage html
 
-# Build docker image
-docker:
-	docker build -t '$(DOCKER_TAG)' .
+.PHONY: assets-clean test

ocrd_calamari/__init__.py

@@ -1 +1,5 @@
+__all__ = [
+    'CalamariRecognize'
+]
+
 from .recognize import CalamariRecognize

requirements-test.txt

@@ -0,0 +1,2 @@
+pytest
+ocrd_tesserocr >= 0.4.0

test/base.py

@@ -0,0 +1,10 @@
+# pylint: disable=unused-import
+
+import os
+import sys
+from unittest import TestCase, skip, main # pylint: disable=unused-import
+
+from test.assets import assets
+
+PWD = os.path.dirname(os.path.realpath(__file__))
+sys.path.append(PWD + '/../ocrd')

test/test_recognize.py

@@ -0,0 +1,62 @@
+import os
+from os.path import join, exists
+import shutil
+
+from test.base import TestCase, main, assets, skip
+
+from ocrd.resolver import Resolver
+
+from ocrd_tesserocr import TesserocrSegmentRegion
+from ocrd_tesserocr import TesserocrSegmentLine
+
+from ocrd_calamari import CalamariRecognize
+
+#METS_HEROLD_SMALL = assets.url_of('SBB0000F29300010000/data/mets_one_file.xml')
+# as long as #96 remains, we cannot use workspaces which have local relative files:
+METS_HEROLD_SMALL = assets.url_of('kant_aufklaerung_1784-binarized/data/mets.xml')
+
+WORKSPACE_DIR = '/tmp/test-ocrd-calamari'
+
+class TestCalamariRecognize(TestCase):
+
+    def setUp(self):
+        if exists(WORKSPACE_DIR):
+            shutil.rmtree(WORKSPACE_DIR)
+        os.makedirs(WORKSPACE_DIR)
+
+    #skip("Takes too long")
+    def runTest(self):
+        resolver = Resolver()
+        workspace = resolver.workspace_from_url(METS_HEROLD_SMALL, dst_dir=WORKSPACE_DIR)
+
+        TesserocrSegmentRegion(
+            workspace,
+            input_file_grp="OCR-D-IMG",
+            output_file_grp="OCR-D-SEG-BLOCK"
+        ).process()
+        workspace.save_mets()
+
+        TesserocrSegmentLine(
+            workspace,
+            input_file_grp="OCR-D-SEG-BLOCK",
+            output_file_grp="OCR-D-SEG-LINE"
+        ).process()
+        workspace.save_mets()
+
+        CalamariRecognize(
+            workspace,
+            input_file_grp="OCR-D-SEG-LINE",
+            output_file_grp="OCR-D-OCR-CALAMARI",
+            parameter={
+                'checkpoint': 'calamari_models/fraktur_historical/*.ckpt.json'
+            }
+        ).process()
+        workspace.save_mets()
+
+        page1 = join(workspace.directory, 'OCR-D-OCR-CALAMARI/OCR-D-OCR-CALAMARI_0001.xml')
+        self.assertTrue(exists(page1))
+        with open(page1, 'r') as f:
+            self.assertIn('verſchuldeten', f.read())
+
+if __name__ == '__main__':
+    main()