From d2b4c393ee83540109050ce346981fa465b158a3 Mon Sep 17 00:00:00 2001
From: kthoden <kthoden@mpiwg-berlin.mpg.de>
Date: Fri, 4 Sep 2020 13:29:47 +0200
Subject: [PATCH] Improve regex for finding page references in aux file

---
 src/eoatex2imxml.py | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/src/eoatex2imxml.py b/src/eoatex2imxml.py
index db20681..90e6379 100755
--- a/src/eoatex2imxml.py
+++ b/src/eoatex2imxml.py
@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8; mode: python -*-
-# Time-stamp: <2020-09-01 14:12:52 (kthoden)>
+# Time-stamp: <2020-09-03 15:57:40 (kthoden)>
 
 """
 Converts Latex files into a customized DocBook XML file.
@@ -1010,12 +1010,14 @@ def process_page_references( latex_dir, set_citations ):
                 # \newlabel{BL}{{\caption@xref {facsim033}{ on input line 37}}{231}{Secondary Literature}{figure.caption.87}{}}
                 # \newlabel{BL}{{1.1}{4}{Forschungsüberblick zur Literatur über Alvarus Thomas}{section.1.1}{}}
                 # \newlabel{BL}{{1.1}{4}}
-                matched_label = re.match(r'\\newlabel\{(.*?)\}\{\{(.*?)\}\{(.*?)\}\}\{(.*?)\}', line)
+                auxpattern = re.compile("\\\\newlabel\{(?P<label>.*?)\}\{\{(?P<num1>.*?)\}\{(?P<pagenumber>.*?)\}\{(?P<sectionname>.*?)\}\{(?P<another>.*?)\}\{(?P<whatever>.*?)\}\}")
+                matched_label = re.match(auxpattern, line)
+                # matched_label = re.match(r'\\newlabel\{(.*?)\}\{\{(.*?)\}\{(.*?)\}\}\{(.*?)\}', line)
                 # matchObjectLabel = re.match(r'\newlabel\{(.*?)\}', line)
                 if matched_label:
                     # matchObjectPage = re.match(r'(.*?)\}\{(\d{1,})\}\}$', line)
                     # if matchObjectPage:
-                    dictPagelabels[matched_label.group(1)] = matched_label.group(4)
+                    dictPagelabels[matched_label.group('label')] = matched_label.group('pagenumber')
                 # parsing out information on cite works
                 matched_citation = re.match(r'\\abx@aux@cite{(.*?)}', line)
                 if matched_citation is not None: