🚧 dinglehopper: Display segment id when hovering over a character difference

qurator/dinglehopper/character_error_rate.py

@@ -15,6 +15,10 @@ def character_error_rate_n(reference, compared) -> Tuple[float, int]:
     :return: character error rate and length of the reference
     """
     d = distance(reference, compared)
+    # XXX
+    from .cli import ExtractedText
+    if isinstance(reference, ExtractedText):
+        reference = reference.text
     n = len(list(grapheme_clusters(unicodedata.normalize('NFC', reference))))
 
     if d == 0:

qurator/dinglehopper/cli.py

@@ -8,11 +8,11 @@ from markupsafe import escape
 from qurator.dinglehopper import *
 
 
-def gen_diff_report(gt_things, ocr_things, css_prefix, joiner, none, align):
+def gen_diff_report(gt_in, ocr_in, css_prefix, joiner, none):
     gtx = ''
     ocrx = ''
 
-    def format_thing(t, css_classes=None):
+    def format_thing(t, css_classes=None, id_=None):
         if t is None:
             html_t = none
             css_classes += ' ellipsis'
@@ -21,19 +21,52 @@ def gen_diff_report(gt_things, ocr_things, css_prefix, joiner, none, align):
         else:
             html_t = escape(t)
 
+        html_custom_attrs = ""
+        # XXX must sanitize id_ or do we trust the XML?
+        if id_:
+            html_custom_attrs = 'data-segment-id="{}"'.format(id_)
+
         if css_classes:
-            return '<span class="{css_classes}">{html_t}</span>'.format(css_classes=css_classes, html_t=html_t)
+            return '<span class="{css_classes}" {html_custom_attrs}>{html_t}</span>'.format(css_classes=css_classes, html_t=html_t, html_custom_attrs=html_custom_attrs)
         else:
             return '{html_t}'.format(html_t=html_t)
 
-    for k, (g, o) in enumerate(align(gt_things, ocr_things)):
-        if g == o:
-            css_classes = None
-        else:
+    if isinstance(gt_in, ExtractedText):
+        print(gt_in.text)
+        if not isinstance(ocr_in, ExtractedText):
+            raise TypeError()
+        # XXX splitting should be done in ExtractedText
+        gt_things = list(grapheme_clusters(gt_in.text))
+        ocr_things = list(grapheme_clusters(ocr_in.text))
+    else:
+        gt_things = gt_in
+        ocr_things = ocr_in
+
+
+
+    g_pos = 0
+    o_pos = 0
+    for k, (g, o) in enumerate(seq_align(gt_things, ocr_things)):
+        css_classes = None
+        gt_id = None
+        ocr_id = None
+        if g != o:
             css_classes = '{css_prefix}diff{k} diff'.format(css_prefix=css_prefix, k=k)
+            if isinstance(gt_in, ExtractedText):
+                gt_id = gt_in.segment_id_for_pos(g_pos) if g is not None else None
+                ocr_id = ocr_in.segment_id_for_pos(o_pos) if o is not None else None
+                # XXX note that deletions and inserts only produce one id + None, UI must
+                #     support this, i.e. display for the one id produced
+                # XXX otherwise, it should always display for BOTH ids
+
+        gtx += joiner + format_thing(g, css_classes, gt_id)
+        ocrx += joiner + format_thing(o, css_classes, ocr_id)
+
+        if g is not None:
+            g_pos += len(g)
+        if o is not None:
+            o_pos += len(o)
 
-        gtx += joiner + format_thing(g, css_classes)
-        ocrx += joiner + format_thing(o, css_classes)
 
     return \
         '''
@@ -51,20 +84,21 @@ def process(gt, ocr, report_prefix, *, metrics=True):
     Click on a wrapper.
     """
 
-    gt_text = text(gt)
-    ocr_text = text(ocr)
+    gt_text = extract(gt)
+    ocr_text = extract(ocr)
 
-    gt_text = substitute_equivalences(gt_text)
-    ocr_text = substitute_equivalences(ocr_text)
+    # FIXME
+    #gt_text = substitute_equivalences(gt_text)
+    #ocr_text = substitute_equivalences(ocr_text)
 
     cer, n_characters = character_error_rate_n(gt_text, ocr_text)
     wer, n_words = word_error_rate_n(gt_text, ocr_text)
 
-    char_diff_report = gen_diff_report(gt_text, ocr_text, css_prefix='c', joiner='', none='·', align=align)
+    char_diff_report = gen_diff_report(gt_text, ocr_text, css_prefix='c', joiner='', none='·')
 
     gt_words = words_normalized(gt_text)
     ocr_words = words_normalized(ocr_text)
-    word_diff_report = gen_diff_report(gt_words, ocr_words, css_prefix='w', joiner=' ', none='⋯', align=seq_align)
+    word_diff_report = gen_diff_report(gt_words, ocr_words, css_prefix='w', joiner=' ', none='⋯')
 
     def json_float(value):
         """Convert a float value to an JSON float.

qurator/dinglehopper/edit_distance.py

@@ -8,6 +8,7 @@ import numpy as np
 from uniseg.graphemecluster import grapheme_clusters
 
 
+
 def levenshtein_matrix(seq1: Sequence, seq2: Sequence):
     """Compute the matrix commonly computed to produce the Levenshtein distance.
     This is also known as the Wagner-Fischer algorithm. The matrix element at the bottom right contains the desired
@@ -75,6 +76,12 @@ def distance(s1, s2):
     Note that this is different from levenshtein() as this function knows about Unicode normalization and grapheme
     clusters. This should be the correct way to compare two Unicode strings.
     """
+    # XXX
+    from .cli import ExtractedText
+    if isinstance(s1, ExtractedText):
+        s1 = s1.text
+    if isinstance(s2, ExtractedText):
+        s2 = s2.text
     s1 = list(grapheme_clusters(unicodedata.normalize('NFC', s1)))
     s2 = list(grapheme_clusters(unicodedata.normalize('NFC', s2)))
     return levenshtein(s1, s2)

qurator/dinglehopper/ocr_files.py

@@ -138,7 +138,7 @@ def page_extract(tree):
 
     # XXX Does a file have to have regions etc.? region vs lines etc.
     # Filter empty region texts
-    regions = (r for r in regions if r.text is not None)
+    regions = [r for r in regions if r.text is not None]
 
     return ExtractedText(regions, '\n')
     # FIXME needs to handle normalization

qurator/dinglehopper/templates/report.html.j2

@@ -26,12 +26,23 @@
       border: 2px solid;
       border-radius: 5px;
     }
+   #status-box {
+      position: fixed;
+      background: grey;
+      color: white;
+      width: 100%;
+      height: 2em;
+    }
     </style>
 </head>
 <body>
 
 
 
+<div id="status-box"> foo</div>
+
+
+
 <div class="container">
 
 {{ gt }}<br>

qurator/dinglehopper/templates/report.html.js

@@ -4,11 +4,16 @@ function find_diff_class(classes) {
 
 $(document).ready(function() {
     $('.diff').mouseover(function() {
-        let c = find_diff_class($(this).attr('class'))
-        $('.' + c).addClass('diff-highlight')
+        let c = find_diff_class($(this).attr('class'));
+        $('.' + c).addClass('diff-highlight');
+
+        segment_id = $(this).attr('data-segment-id');
+        $('#status-box').text(segment_id);
     });
     $('.diff').mouseout(function() {
-        let c = find_diff_class($(this).attr('class'))
-        $('.' + c).removeClass('diff-highlight')
+        let c = find_diff_class($(this).attr('class'));
+        $('.' + c).removeClass('diff-highlight');
+
+        $('#status-box').text('');
     });
 });

qurator/dinglehopper/word_error_rate.py

@@ -32,6 +32,11 @@ def words(s):
         cat = subcat[0]
         return cat in unwanted_categories or subcat in unwanted_subcategories
 
+    # XXX
+    from .cli import ExtractedText
+    if isinstance(s, ExtractedText):
+        s = s.text
+
     # We follow Unicode Standard Annex #29 on Unicode Text Segmentation here: Split on word boundaries using
     # uniseg.wordbreak.words() and ignore all "words" that contain only whitespace, punctation "or similar characters."
     for word in uniseg.wordbreak.words(s):
@@ -42,10 +47,20 @@ def words(s):
 
 
 def words_normalized(s):
+    # XXX
+    from .cli import ExtractedText
+    if isinstance(s, ExtractedText):
+        s = s.text
     return words(unicodedata.normalize('NFC', s))
 
 
 def word_error_rate_n(reference, compared) -> Tuple[float, int]:
+    # XXX
+    from .cli import ExtractedText
+    if isinstance(reference, ExtractedText):
+        reference = reference.text
+    if isinstance(compared, ExtractedText):
+        compared = compared.text
     if isinstance(reference, str):
         reference_seq = list(words_normalized(reference))
         compared_seq = list(words_normalized(compared))