From 7293c43e048fc48afe1d475a1c33ebffad6c2ef0 Mon Sep 17 00:00:00 2001
From: kthoden <kthoden@mpiwg-berlin.mpg.de>
Date: Fri, 9 Jul 2021 11:16:26 +0200
Subject: [PATCH] Write failed ids to file

---
 src/imxml2django.py | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

diff --git a/src/imxml2django.py b/src/imxml2django.py
index 15fac07..cbc19f4 100755
--- a/src/imxml2django.py
+++ b/src/imxml2django.py
@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8; mode: python -*-
-# Time-stamp: <2021-02-17 09:24:51 (kthoden)>
+# Time-stamp: <2021-07-08 19:20:28 (kthoden)>
 
 """
 Create an XML file that can be inserted into the Django database
@@ -1587,6 +1587,13 @@ def bring_footnote_down_django(footnote, fragment, footnote_number, object_numbe
 logging.info("----------------------------------------------")
 logging.info("Processing Cross References")
 
+libeoaconvert.debug_xml_here(
+        xmlDjangoTree,
+        "beforecrossreference",
+        DEBUG_DIR
+)
+
+failed_ids = []
 # Substitute References with their targets (wit links)
 for xmlEOAchapter in xmlEOAchapters:
     # for hyperimage collages
@@ -2017,6 +2024,16 @@ def bring_footnote_down_django(footnote, fragment, footnote_number, object_numbe
 tmpFile.close()
 logging.debug(f"Wrote {OUTPUT_DIR}/Django.xml.")
 
+if len(failed_ids) > 0:
+    cleaned_failures = sorted(set(failed_ids))
+    tmpFile = open( OUTPUT_DIR / "debug/failed_ids.txt", "w")
+    tmpFile.writelines(f"Missing IDs by appearance ({len(failed_ids)} in total):\n")
+    tmpFile.writelines(failed_ids)
+    tmpFile.writelines(f"\nMissing IDs sorted and uniqued ({len(cleaned_failures)} in total):\n")
+    tmpFile.writelines(cleaned_failures)
+    tmpFile.close()
+    logging.debug(f"Some ids could not be referenced. Check {OUTPUT_DIR}/debug/failed_ids.txt.")
+
 if args.checkpublicationcfg:
     check_publication_cfg(INPUT_DIR / "publication.cfg")
 else: