Merge pull request #61 from UB-Mannheim/id-system

Fix metadata check: ocr-id -> ocr-system, #59
ocropus · Sep 16, 2016 · fff7c99 · fff7c99
2 parents dc629e9 + 0132d54
commit fff7c99
Show file tree

Hide file tree

Showing 10 changed files with 25 additions and 25 deletions.
diff --git a/hocr-check b/hocr-check
@@ -90,8 +90,8 @@ else: doc = html.parse(sys.stdin)
 ################################################################
 
 # check for presence of meta information
-test_ok(doc.xpath("//meta[@name='ocr-id']")!=[], "//meta[@name='ocr-id']")
-test_ok(doc.xpath("//meta[@name='ocr-recognized']")!=[], "//meta[@name='ocr-recognized']")
+test_ok(doc.xpath("//meta[@name='ocr-system']")!=[], "//meta[@name='ocr-system']")
+test_ok(doc.xpath("//meta[@name='ocr-capabilites']")!=[], "//meta[@name='ocr-capabilites']")
 
 # check for presence of page
 test_ok(doc.xpath("//*[@class='ocr_page']")!=[], "has a page")
@@ -136,7 +136,7 @@ if not nooverlap:
 
 # FIXME add many other checks:
 # - containment of paragraphs, careas, etc.
-# - ocr-recognized vs. actual tags
+# - ocr-capabilites vs. actual tags
 # - warn about text outside ocr_ elements
 # - check title= attribute format
 # - check that only the right attributes are present on the right elements

diff --git a/test/hocr-check/ancestor/notok-carea.html b/test/hocr-check/ancestor/notok-carea.html
@@ -1,7 +1,7 @@
 <html>
   <head>
-    <meta name='ocr-id' content="foo"/>
-    <meta name='ocr-recognized' content="foo"/>
+    <meta name='ocr-system' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
   </head>
   <body>
     <span class="ocr_page"></span>

diff --git a/test/hocr-check/ancestor/notok-line.html b/test/hocr-check/ancestor/notok-line.html
@@ -1,7 +1,7 @@
 <html>
   <head>
-    <meta name='ocr-id' content="foo"/>
-    <meta name='ocr-recognized' content="foo"/>
+    <meta name='ocr-system' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
   </head>
   <body>
     <span class="ocr_page"></span>

diff --git a/test/hocr-check/ancestor/notok-par.html b/test/hocr-check/ancestor/notok-par.html
@@ -1,7 +1,7 @@
 <html>
   <head>
-    <meta name='ocr-id' content="foo"/>
-    <meta name='ocr-recognized' content="foo"/>
+    <meta name='ocr-system' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
   </head>
   <body>
     <span class="ocr_page"></span>

diff --git a/test/hocr-check/ancestor/ok-column.html b/test/hocr-check/ancestor/ok-column.html
@@ -1,7 +1,7 @@
 <html>
   <head>
-    <meta name='ocr-id' content="foo"/>
-    <meta name='ocr-recognized' content="foo"/>
+    <meta name='ocr-system' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
   </head>
   <body>
     <span class="ocr_page">

diff --git a/test/hocr-check/ancestor/ok-line.html b/test/hocr-check/ancestor/ok-line.html
@@ -1,7 +1,7 @@
 <html>
   <head>
-    <meta name='ocr-id' content="foo"/>
-    <meta name='ocr-recognized' content="foo"/>
+    <meta name='ocr-system' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
   </head>
   <body>
     <span class="ocr_page">

diff --git a/test/hocr-check/ancestor/ok-par.html b/test/hocr-check/ancestor/ok-par.html
@@ -1,7 +1,7 @@
 <html>
   <head>
-    <meta name='ocr-id' content="foo"/>
-    <meta name='ocr-recognized' content="foo"/>
+    <meta name='ocr-system' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
   </head>
   <body>
     <span class="ocr_page">

diff --git a/test/hocr-check/meta/notok-id.html → test/hocr-check/meta/notok-system.html b/test/hocr-check/meta/notok-id.html → test/hocr-check/meta/notok-system.html
@@ -1,6 +1,6 @@
 <html>
   <head>
-    <meta name='ocr-recognized' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
   </head>
   <body>
     <span class="ocr_page"></span>

diff --git a/test/hocr-check/meta/ok-id.html b/test/hocr-check/meta/ok-id.html
diff --git a/test/hocr-check/meta/ok-system.html b/test/hocr-check/meta/ok-system.html
@@ -0,0 +1,9 @@
+<html>
+  <head>
+    <meta name='ocr-system' content="foo"/>
+    <meta name='ocr-capabilites' content="foo"/>
+  </head>
+  <body>
+    <span class="ocr_page"> </span>
+  </body>
+</html>