show orphan comments after Disqus import (if any)

An orphan comment is exported by Disqus but its thread id is
non-existent (probably deleted, moved). Usually from the earlier
days (or WordPress migration).

It is not possible to get the thread without manual intervention (
aka SQLite insertions).

isso/migrate.py

@@ -9,6 +9,7 @@ from __future__ import division
 
 import sys
 import os
+import textwrap
 
 from time import mktime, strptime
 from collections import defaultdict
@@ -20,12 +21,14 @@ except ImportError:
 
 from xml.etree import ElementTree
 
-
 ns = '{http://disqus.com}'
 dsq = '{http://disqus.com/disqus-internals}'
 
+threads = set([])
+comments = set([])
 
-def insert(db, thread, comments):
+
+def insert(db, thread, posts):
 
     path = urlparse(thread.find('%sid' % ns).text).path
     remap = dict()
@@ -33,13 +36,15 @@ def insert(db, thread, comments):
     if path not in db.threads:
         db.threads.new(path, thread.find('%stitle' % ns).text.strip())
 
-    for item in sorted(comments, key=lambda k: k['created']):
+    for item in sorted(posts, key=lambda k: k['created']):
 
         dsq_id = item.pop('dsq:id')
         item['parent'] = remap.get(item.pop('dsq:parent', None))
         rv = db.comments.add(path, item)
         remap[dsq_id] = rv["id"]
 
+    comments.update(set(remap.keys()))
+
 
 def disqus(db, xmlfile):
 
@@ -67,7 +72,6 @@ def disqus(db, xmlfile):
     num = len(tree.findall('%sthread' % ns))
     cols = int(os.popen('stty size', 'r').read().split()[1])
 
-    threads = 0
     for i, thread in enumerate(tree.findall('%sthread' % ns)):
 
         if int(round((i+1)/num, 2) * 100) % 13 == 0:
@@ -76,13 +80,28 @@ def disqus(db, xmlfile):
             sys.stdout.write("\r[%i%%]  %s" % (((i+1)/num * 100), thread.find('%sid' % ns).text))
             sys.stdout.flush()
 
+        # skip (possibly?) duplicate, but empty thread elements
         if thread.find('%sid' % ns).text is None:
             continue
 
         id = thread.attrib.get(dsq + 'id')
         if id in res:
-            threads += 1
+            threads.add(id)
             insert(db, thread, res[id])
 
     sys.stdout.write("\r%s" % (" "*cols))
-    sys.stdout.write("\r[100%%]  %i threads, %i comments" % (threads, len(tree.findall('%spost' % ns))))
+    sys.stdout.write("\r[100%%]  %i threads, %i comments\n" % (len(threads), len(comments)))
+
+    orphans = set(map(lambda e: e.attrib.get(dsq + "id"), tree.findall("%spost" % ns))) - comments
+    if orphans:
+        print("Found %i orphans:" % len(orphans))
+        for post in tree.findall("%spost" % ns):
+            if post.attrib.get(dsq + "id") not in orphans:
+                continue
+
+            print(" * %s by %s <%s>" % (post.attrib.get(dsq + "id"),
+                                        post.find("%sauthor/%sname" % (ns, ns)).text,
+                                        post.find("%sauthor/%semail" % (ns, ns)).text))
+            print(textwrap.fill(post.find("%smessage" % ns).text,
+                                initial_indent="  ", subsequent_indent="  "))
+            print("")