From 969e9bb1550ae460a14e1a27b4d1cb3be8e08ce2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Fran=C3=A7ois=20Poulain?= <fpoulain@metrodore.fr>
Date: Mon, 3 Aug 2020 19:16:19 +0200
Subject: [PATCH] feat(convert): reconstruction des dl implicites

---
 drupal2spip_lal/base/convert.py | 34 +++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)

diff --git a/drupal2spip_lal/base/convert.py b/drupal2spip_lal/base/convert.py
index c28a866..7e02a9b 100644
--- a/drupal2spip_lal/base/convert.py
+++ b/drupal2spip_lal/base/convert.py
@@ -1,4 +1,6 @@
+import re
 from datetime import datetime
+from itertools import groupby
 
 from django.utils.timezone import make_aware, now
 
@@ -17,7 +19,39 @@ def convert_timestamp(timestamp):
     return make_aware(datetime.fromtimestamp(timestamp))
 
 
+def strong_to_dl(html):
+    """
+    Marie-Odile nous fait des dl-like à base de strong.
+    Parser html avec des regex est mal mais on essaie de reconstruire la dl
+    avant de casser la forme avec bs.
+    """
+    is_strong = r'<strong>(?P<dt>.+)</strong>(?P<dd>.*)$'
+
+    def is_strong_item(s):
+        return bool(re.match(is_strong, s))
+
+    items = re.split(r'[\r\n]+', html)
+    grouped_items = groupby(items, key=is_strong_item)
+    r = []
+    for key, group in grouped_items:
+        group = list(group)
+        if key and len(group) > 2:
+            dl = ['<dl>']
+            for elem in group:
+                match = re.match(is_strong, elem).groupdict()
+                dl += [
+                    '<dt>{}</dt>'.format(match['dt'].strip()),
+                    '<dd>{}</dd>'.format(match['dd'].strip()),
+                ]
+            dl.append('</dl>')
+            r += dl
+        else:
+            r += group
+    return '\n'.join(r)
+
+
 def sanitarize_html(html):
+    html = strong_to_dl(html)
     soup = bs(html, 'html.parser')
     return soup.prettify(formatter="html5")