Merge branch 'main' into typos

OCR-D · Apr 22, 2024 · 35832ac · 35832ac
2 parents 0e677fa + 33b4cd6
commit 35832ac
Show file tree

Hide file tree

Showing 6 changed files with 65 additions and 30 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -17,19 +17,18 @@ jobs:
       fail-fast: false
       matrix:
         python-version:
-          # - '3.7'
           - '3.8'
           - '3.9'
-          # - '3.10'
-          # - '3.11'
+          - '3.10'
+          - '3.11'
         os:
           - ubuntu-22.04
           # - macos-latest
 
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python-version }}
       - name: Install and test

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -5,6 +5,22 @@ Versioned according to [Semantic Versioning](http://semver.org/).
 
 ## Unreleased
 
+## [0.6.0] - 2024-04-22
+
+Changed:
+
+* rename parameter `replace_textstyle` -> `overwrite_style`, #8, #9
+* rename parameter `network` -> `model`, #8, #9
+* parameter `overwrite_text`: If true, replace all existing textequivs, else (default) just add a textequiv, #8, #9
+* parameter `overwrite_style`: if true (default), replace the `fontFamily` attribute of existing textstyle or create new style if non exists., #8, #9
+* parameter `min_score_style`: Score between 0 and 100, font classification results below this score will not be serialized or used for OCR, default: 0, #8, #9
+
+Fixed:
+
+  * Require OCR-D/core v2.64.1+ with proper support for `importlib{.,_}metadata`, #10
+  * CI: Use most recent actions, #15
+  * missing top-level `__init__.py`, #12
+
 ## [0.5.2] - 2024-02-01
 
 Fixed:
@@ -22,6 +38,7 @@ Fixed:
 - First release in ocrd_all
 
 <!-- link-labels -->
-[0.5.1]: ../../compare/v0.5.2...v0.5.1
+[0.6.0]: ../../compare/v0.6.0...v0.5.2
+[0.5.2]: ../../compare/v0.5.2...v0.5.1
 [0.5.1]: ../../compare/v0.5.1...v0.5.0
 [0.5.0]: ../../compare/v0.5.0...HEAD
diff --git a/ocrd_froc/__init__.py b/ocrd_froc/__init__.py
diff --git a/ocrd_froc/ocrd-tool.json b/ocrd_froc/ocrd-tool.json
@@ -1,5 +1,5 @@
 {
-  "version": "0.5.2",
+  "version": "0.6.0",
   "git_url": "https://github.com/OCR-D/ocrd_froc",
   "tools": {
     "ocrd-froc-recognize": {
@@ -21,13 +21,26 @@
           "enum": ["none", "SelOCR", "COCR", "adaptive"],
           "default": "none"
         },
-        "replace_textstyle": {
-          "description": "Whether to replace existing textStyle",
+        "overwrite_style": {
+          "description": "Whether to overwrite existing TextStyle/@fontFamily attributes",
           "type": "boolean",
           "required": false,
           "default": true
         },
-        "network": {
+        "min_score_style": {
+          "description": "The minimum score of a font classification to be serialized/used as input for OCR",
+          "type": "number",
+          "format": "float",
+          "required": false,
+          "default": 0
+        },
+        "overwrite_text": {
+          "description": "Whether to remove any existing TextEquiv before adding text",
+          "type": "boolean",
+          "required": false,
+          "default": false
+        },
+        "model": {
           "description": "The file name of the neural network to use, including sufficient path information. Defaults to the model bundled with ocrd_froc.",
           "type": "string",
           "required": false

diff --git a/ocrd_froc/processor.py b/ocrd_froc/processor.py
@@ -2,6 +2,7 @@
 Wrap FROC as an ocrd.Processor
 """
 import os
+from typing import List, Tuple
 
 from ocrd import Processor
 from ocrd_utils import (
@@ -21,7 +22,7 @@
 from ocrd_modelfactory import page_from_file
 from .froc import Froc
 
-OCRD_TOOL = loads(resource_string(__name__, 'ocrd-tool.json'))
+OCRD_TOOL = loads(resource_string(__package__, 'ocrd-tool.json'))
 
 class FROCProcessor(Processor):
 
@@ -35,29 +36,21 @@ def __init__(self, *args, **kwargs):
 
     def setup(self):
 
-        if 'network' not in self.parameter:
-            self.parameter['network'] = str(resource_filename(f'ocrd_froc.models', 'default.froc'))
+        if 'model' not in self.parameter:
+            self.parameter['model'] = str(resource_filename(f'ocrd_froc.models', 'default.froc'))
 
-        network_file = self.resolve_resource(self.parameter['network'])
-        self.froc = Froc.load(network_file)
+        model = self.resolve_resource(self.parameter['model'])
+        self.froc = Froc.load(model)
 
     def _process_segment(self, segment, image):
-        textStyle = segment.get_TextStyle()
-        if textStyle and self.parameter['replace_textstyle']:
-            textStyle = None
-            segment.set_TextStyle(textStyle)
-        if not textStyle:
-            textStyle = TextStyleType()
-            segment.set_TextStyle(textStyle)
-
         ocr_method = self.parameter['ocr_method']
 
         result = {}
 
         if ocr_method != 'COCR':
 
             result = self.froc.classify(image)
-            classification_result = ''
+            fonts_detected : List[Tuple[str, float]] = []
 
             font_class_priors = self.parameter['font_class_priors']
             output_font = True
@@ -87,12 +80,21 @@ def _process_segment(self, segment, image):
                 score = round(100 * score)
                 if score <= 0:
                     continue
-                if classification_result != '':
-                    classification_result += ', '
-                classification_result += '%s:%d' % (typegroup, score)
+                fonts_detected.append((typegroup, score))
+
+            classification_result = ', '.join([
+                f'{family}:{score}' \
+                for family, score in fonts_detected \
+                if score > self.parameter['min_score_style']
+            ])
 
             if output_font:
-                textStyle.set_fontFamily(classification_result)
+                textStyle = segment.get_TextStyle()
+                if not textStyle or self.parameter['overwrite_style']:
+                    if not textStyle:
+                        textStyle = TextStyleType()
+                        segment.set_TextStyle(textStyle)
+                    textStyle.set_fontFamily(classification_result)
 
 
         if ocr_method == 'COCR':
@@ -111,8 +113,12 @@ def _process_segment(self, segment, image):
                                           method=ocr_method,
                                           classification_result=result,
                                           fast_cocr=fast_cocr,
-                                          adaptive_threshold=adaptive_threshold)
-        segment.set_TextEquiv([TextEquivType(Unicode=transcription, conf=score)])
+                                          adaptive_treshold=adaptive_threshold)
+
+        if self.parameter['overwrite_text']:
+            segment.set_TextEquiv([TextEquivType(Unicode=transcription, conf=score)])
+        else:
+            segment.add_TextEquiv(TextEquivType(Unicode=transcription, conf=score))
 
 
     def process(self):  # type: ignore

diff --git a/requirements.txt b/requirements.txt
@@ -1,4 +1,4 @@
-ocrd >= 2.22.3
+ocrd >= 2.64.1
 pandas
 scikit-image
 torch >= 1.4.0