Correct order for output files.

2025-07-01 19:00:34 +00:00 · 2019-05-13 15:03:43 +02:00
parent 937abb8c8d
commit 843151e547
3 changed files with 174 additions and 80 deletions
--- a/45
+++ b/45
@ -7,22 +7,31 @@ import os
 import re
 import sys

-input_files = filter(lambda x: x.endswith(".hocr"), sorted(os.listdir(sys.argv[1])))
+input_files = sorted(
+    filter(
+        lambda x: x.endswith(".hocr"),
+        os.listdir(sys.argv[1])
+    ),
+    key=lambda x: int(re.search(r'\d+', x).group(0))
+)
+# "page-1.hocr" -> "1"
 output_file = open(sys.argv[2], "w")

-output_file.write('<?xml version="1.0" encoding="UTF-8"?>\n' +
-           '<TEI xmlns="http://www.tei-c.org/ns/1.0" xml:id="dtabf">\n' +
-           '    <teiHeader>\n' +
-           '        <fileDesc>\n' +
-           '            <titleStmt/>\n' +
-           '            <publicationStmt/>\n' +
-           '            <sourceDesc/>\n' +
-           '        </fileDesc>\n' +
-           '        <encodingDesc/>\n' +
-           '        <profileDesc/>\n' +
-           '    </teiHeader>\n' +
-           '    <text>\n' +
-           '        <body>\n')
+output_file.write(
+      '<?xml version="1.0" encoding="UTF-8"?>\n'
+    + '<TEI xmlns="http://www.tei-c.org/ns/1.0" xml:id="dtabf">\n'
+    + '    <teiHeader>\n'
+    + '        <fileDesc>\n'
+    + '            <titleStmt/>\n'
+    + '            <publicationStmt/>\n'
+    + '            <sourceDesc/>\n'
+    + '        </fileDesc>\n'
+    + '        <encodingDesc/>\n'
+    + '        <profileDesc/>\n'
+    + '    </teiHeader>\n'
+    + '    <text>\n'
+    + '        <body>\n'
+)

 for input_file in input_files:
    tree = ET.parse(os.path.join(sys.argv[1], input_file))
@ -40,7 +49,9 @@ for input_file in input_files:
                output_file.write('<lb/>\n')
        output_file.write('            </p>\n')

-output_file.write('        </body>\n' +
-           '    </text>\n' +
-           '</TEI>')
+output_file.write(
+      '        </body>\n'
+    + '    </text>\n'
+    + '</TEI>')
+
 output_file.close()