🚧 dinglehopper: Compare line text directories (WIP)

qurator/dinglehopper/cli_line_dirs.py

@@ -44,12 +44,12 @@ def process(gt_dir, ocr_dir, report_prefix, *, metrics=True):
     n_characters = None
     char_diff_report = ""
 
-    for gt in os.listdir(gt_dir):
+    for k, gt in enumerate(os.listdir(gt_dir)):
         # Find a match by replacing the suffix
         ocr = removesuffix(gt, gt_suffix) + ocr_suffix
 
-        gt_text = plain_extract(os.path.join(gt_dir, gt))
-        ocr_text = plain_extract(os.path.join(ocr_dir, ocr))
+        gt_text = plain_extract(os.path.join(gt_dir, gt), include_filename_in_id=True)
+        ocr_text = plain_extract(os.path.join(ocr_dir, ocr), include_filename_in_id=True)
 
         # Compute CER
         l_cer, l_n_characters = character_error_rate_n(gt_text, ocr_text)
@@ -65,7 +65,7 @@ def process(gt_dir, ocr_dir, report_prefix, *, metrics=True):
         wer = 9999; n_words = 0
 
         char_diff_report += gen_diff_report(
-             gt_text, ocr_text, css_prefix="c", joiner="", none="·"
+             gt_text, ocr_text, css_prefix="l{0}-c".format(k), joiner="", none="·"
         )
 
         # TODO

qurator/dinglehopper/ocr_files.py

@@ -1,8 +1,9 @@
 from __future__ import division, print_function
 
+import os
+import sys
 from typing import Iterator
 from warnings import warn
-import sys
 
 from lxml import etree as ET
 from lxml.etree import XMLSyntaxError
@@ -130,12 +131,15 @@ def page_text(tree, *, textequiv_level="region"):
     return page_extract(tree, textequiv_level=textequiv_level).text
 
 
-def plain_extract(filename):
+def plain_extract(filename, include_filename_in_id=False):
+    id_template = "{filename} - line {no}" if include_filename_in_id else "line {no}"
     with open(filename, "r") as f:
         return ExtractedText(
             None,
             [
-                ExtractedText("line %d" % no, None, None, normalize_sbb(line))
+                ExtractedText(
+                    id_template.format(filename=os.path.basename(filename), no=no),
+                    None, None, normalize_sbb(line))
                 for no, line in enumerate(f.readlines())
             ],
             "\n",