ocrd 3.1.0__py3-none-any.whl → 3.1.1__py3-none-any.whl

ocrd/ocrd-all-tool.json

@@ -1,21 +1,45 @@
 {
-  "ocrd-dummy": {
-    "executable": "ocrd-dummy",
-    "description": "Bare-bones processor creates PAGE-XML and optionally copies file from input group to output group",
-    "steps": [
-      "preprocessing/optimization"
-    ],
-    "categories": [
-      "Image preprocessing"
-    ],
-    "input_file_grp": "DUMMY_INPUT",
-    "output_file_grp": "DUMMY_OUTPUT",
-    "parameters": {
-      "copy_files": {
-        "type": "boolean",
-        "default": false,
-        "description": "Whether to actually copy files (true) or just create PAGE-XML as a side effect (false)"
-      }
-    }
+ "ocrd-dummy": {
+  "executable": "ocrd-dummy",
+  "description": "Bare-bones processor creates PAGE-XML and optionally copies file from input group to output group",
+  "steps": [
+   "preprocessing/optimization"
+  ],
+  "categories": [
+   "Image preprocessing"
+  ],
+  "input_file_grp_cardinality": 1,
+  "output_file_grp_cardinality": 1,
+  "parameters": {
+   "copy_files": {
+    "type": "boolean",
+    "default": false,
+    "description": "Whether to actually copy files (true) or just create PAGE-XML as a side effect (false)"
+   }
   }
-}
+ },
+ "ocrd-filter": {
+  "executable": "ocrd-filter",
+  "description": "Bare-bones processor can be dynamically configured to remove segments based on XPath queries",
+  "steps": [
+   "recognition/post-correction"
+  ],
+  "categories": [
+   "Quality assurance"
+  ],
+  "input_file_grp_cardinality": 1,
+  "output_file_grp_cardinality": 1,
+  "parameters": {
+   "select": {
+    "type": "string",
+    "default": "//*[ends-with(local-name(),'Region')]",
+    "description": "Which segments to select for removal. An XPath 2.0 query expression (path and optional predicates), with 'pc' as namespace prefix for PAGE-XML and our extension functions (see help text). Only selection of segment hierarchy elements is allowed (so e.g. `*` would be equivalent to `pc:NoiseRegion|pc:LineDrawingRegion|pc:AdvertRegion|pc:ImageRegion|pc:ChartRegion|pc:MusicRegion|pc:GraphicRegion|pc:UnknownRegion|pc:CustomRegion|pc:SeparatorRegion|pc:MathsRegion|pc:TextRegion|pc:MapRegion|pc:ChemRegion|pc:TableRegion|pc:TextLine|pc:Word|pc:Glyph`, but `pc:MetadataItem` or `pc:Border` or `pc:Coords` would not match).\nFor example, to remove words or glyphs with low text confidence, select '(pc:Word|pc:Glyph)[pc:TextEquiv/@conf < 0.7]'. Or low layout confidence, '*[pc:Coords/@conf < 0.7]'.\nTo remove high pixel-to-character rate, select '*[pc:pixelarea(.) div string-length(pc:textequiv(.)) > 10000]'."
+   },
+   "plot": {
+    "type": "boolean",
+    "default": false,
+    "description": "Whether to extract an image for each filtered segment and write to the output fileGrp."
+   }
+  }
+ }
+}

{ocrd-3.1.0.dist-info → ocrd-3.1.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ocrd
-Version: 3.1.0
+Version: 3.1.1
 Summary: OCR-D framework
 Author-email: Konstantin Baierer <unixprog@gmail.com>
 License: Apache License 2.0

{ocrd-3.1.0.dist-info → ocrd-3.1.1.dist-info}/RECORD

@@ -2,7 +2,7 @@ ocrd/__init__.py,sha256=ZswMVmlqFhAEIzMR3my6IKPq9XLH21aDPC_m_8Jh4dA,1076
 ocrd/constants.py,sha256=6dn3mG54WqHsKInmLZp4kJjNqqPtBoFoSuLUuRbOps0,740
 ocrd/lib.bash,sha256=le6XqAOEacdjP3JNSlPkxwRH1y0oVjNQM2tX5d6QFO4,10901
 ocrd/mets_server.py,sha256=EL6CMtVjmVPs9--3Vn-JyhU2VDMNw-6QN_XWRKFp6uk,22451
-ocrd/ocrd-all-tool.json,sha256=9bX2VYnUwhTAzAvKaoT77BFzbgBGgyIt7qBqARpwWNc,586
+ocrd/ocrd-all-tool.json,sha256=EYXmMzP68p3KzL8nUZ16TCX2chQzKkAeISvuXqI_yIw,2094
 ocrd/resolver.py,sha256=Ba9ALQbTXz6_mla4VqN9tAfHoj6aKuNJAU4tIDnjcHE,14952
 ocrd/resource_list.yml,sha256=82-PiqkZnka1kTj3MQqNn4wXWKHHtoFchsQuetWuqFs,2633
 ocrd/resource_manager.py,sha256=kIWDoKxWH4IJE1gcoTcCRQjYjieCqiQclyuyF6Y9b8A,16813
@@ -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-3.1.0.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-ocrd-3.1.0.dist-info/METADATA,sha256=IAFV_u_0lEtvLnjHznVBweeEEBfBbe32FfbnM6EEdhc,10442
-ocrd-3.1.0.dist-info/WHEEL,sha256=P9jw-gEje8ByB7_hXoICnHtVCrEwMQh-630tKvQWehc,91
-ocrd-3.1.0.dist-info/entry_points.txt,sha256=4hcJ2LkK_OlIabHnKgFit35Ap7b5Lz1Gb4hzkxV0Kiw,152
-ocrd-3.1.0.dist-info/top_level.txt,sha256=pUgiN42t4KXC5rvpi6V8atza31XP4SCznXpXlVlvomM,75
-ocrd-3.1.0.dist-info/RECORD,,
+ocrd-3.1.1.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+ocrd-3.1.1.dist-info/METADATA,sha256=athV-jzUTYFDjcbykQcSH6JXva7mNDVyHJgf3LCof8s,10442
+ocrd-3.1.1.dist-info/WHEEL,sha256=P9jw-gEje8ByB7_hXoICnHtVCrEwMQh-630tKvQWehc,91
+ocrd-3.1.1.dist-info/entry_points.txt,sha256=4hcJ2LkK_OlIabHnKgFit35Ap7b5Lz1Gb4hzkxV0Kiw,152
+ocrd-3.1.1.dist-info/top_level.txt,sha256=pUgiN42t4KXC5rvpi6V8atza31XP4SCznXpXlVlvomM,75
+ocrd-3.1.1.dist-info/RECORD,,