add initial support to import WordPress comments

isso/__init__.py

@@ -207,6 +207,8 @@ def main():
     imprt.add_argument("dump", metavar="FILE")
     imprt.add_argument("-n", "--dry-run", dest="dryrun", action="store_true",
                        help="perform a trial run with no changes made")
+    imprt.add_argument("-t", "--type", dest="type", default=None,
+                       choices=["disqus", "wordpress"], help="export type")
 
     serve = subparser.add_parser("run", help="run server")
 
@@ -223,7 +225,7 @@ def main():
             dbpath = conf.get("general", "dbpath")
 
         mydb = db.SQLite3(dbpath, conf)
-        migrate.dispatch(mydb, args.dump)
+        migrate.dispatch(args.type, mydb, args.dump)
 
         sys.exit(0)
 

isso/migrate.py

@@ -1,14 +1,18 @@
 # -*- encoding: utf-8 -*-
 
-from __future__ import division
+from __future__ import division, print_function
 
 import sys
 import os
+import io
 import textwrap
 
-from time import mktime, strptime
+from time import mktime, strptime, time
 from collections import defaultdict
 
+from isso.utils import anonymize
+from isso.compat import string_types
+
 try:
     input = raw_input
 except NameError:
@@ -22,6 +26,39 @@ except ImportError:
 from xml.etree import ElementTree
 
 
+def strip(val):
+    if isinstance(val, string_types):
+        return val.strip()
+    return val
+
+
+class Progress(object):
+
+    def __init__(self, end):
+        self.end = end or 1
+
+        self.istty = sys.stdout.isatty()
+        self.last = 0
+
+    def update(self, i, message):
+
+        if not self.istty or message is None:
+            return
+
+        cols = int((os.popen('stty size', 'r').read()).split()[1])
+        message = message[:cols - 7]
+
+        if time() - self.last > 0.2:
+            sys.stdout.write("\r{0}".format(" " * cols))
+            sys.stdout.write("\r[{0:.0%}]  {1}".format(i/self.end, message))
+            sys.stdout.flush()
+            self.last = time()
+
+    def finish(self, message):
+        self.last = 0
+        self.update(self.end, message + "\n")
+
+
 class Disqus(object):
 
     ns = '{http://disqus.com}'
@@ -116,9 +153,94 @@ class Disqus(object):
                 print("")
 
 
-def dispatch(db, dump):
+class WordPress(object):
+
+    ns = "{http://wordpress.org/export/1.0/}"
+
+    def __init__(self, db, xmlfile):
+        self.db = db
+        self.xmlfile = xmlfile
+        self.count = 0
+
+    def insert(self, thread):
+
+        path = urlparse(thread.find("link").text).path
+        self.db.threads.new(path, thread.find("title").text.strip())
+
+        comments = list(map(WordPress.Comment, thread.findall(WordPress.ns + "comment")))
+        comments.sort(key=lambda k: k["id"])
+
+        remap = {}
+        ids = set(c["id"] for c in comments)
+
+        self.count += len(ids)
+
+        while comments:
+            for i, item in enumerate(comments):
+                if item["parent"] in ids:
+                    continue
+
+                item["parent"] = remap.get(item["parent"], None)
+                rv = self.db.comments.add(path, item)
+                remap[item["id"]] = rv["id"]
+
+                ids.remove(item["id"])
+                comments.pop(i)
+
+                break
+            else:
+                # should never happen, but... it's WordPress.
+                return
+
+    def migrate(self):
+
+        tree = ElementTree.parse(self.xmlfile)
+        items = tree.findall("channel/item")
+
+        progress = Progress(len(items))
+        for i, thread in enumerate(items):
+            progress.update(i, thread.find("title").text)
+            self.insert(thread)
+
+        progress.finish("{0} threads, {1} comments".format(len(items), self.count))
+
+    @classmethod
+    def Comment(cls, el):
+        return {
+            "text": strip(el.find(WordPress.ns + "comment_content").text),
+            "author": strip(el.find(WordPress.ns + "comment_author").text),
+            "email": strip(el.find(WordPress.ns + "comment_author_email").text),
+            "website": strip(el.find(WordPress.ns + "comment_author_url").text),
+            "remote_addr": anonymize(
+                strip(el.find(WordPress.ns + "comment_author_IP").text)),
+            "created": mktime(strptime(
+                strip(el.find(WordPress.ns + "comment_date_gmt").text),
+                "%Y-%m-%d %H:%M:%S")),
+            "mode": 1 if el.find(WordPress.ns + "comment_approved").text == "1" else 2,
+            "id": int(el.find(WordPress.ns + "comment_id").text),
+            "parent": int(el.find(WordPress.ns + "comment_parent").text) or None
+        }
+
+
+def dispatch(type, db, dump):
         if db.execute("SELECT * FROM comments").fetchone():
             if input("Isso DB is not empty! Continue? [y/N]: ") not in ("y", "Y"):
                 raise SystemExit("Abort.")
 
+        if type is None:
+
+            with io.open(dump) as fp:
+                peek = fp.read(2048)
+
+            if 'xmlns:wp="%s"' % WordPress.ns[1:-1] in peek:
+                type = "wordpress"
+
+            if '<disqus xmlns=' in peek:
+                type = "disqus"
+
+        if type == "wordpress":
+            WordPress(db, dump).migrate()
+        elif type == "disqus":
             Disqus(db, dump).migrate()
+        else:
+            raise SystemExit("Unknown format, abort.")

isso/tests/test_migration.py

@@ -11,7 +11,7 @@ from os.path import join, dirname
 from isso.core import Config
 
 from isso.db import SQLite3
-from isso.migrate import Disqus
+from isso.migrate import Disqus, WordPress
 
 
 class TestMigration(unittest.TestCase):
@@ -35,11 +35,30 @@ class TestMigration(unittest.TestCase):
         b = db.comments.get(2)
         self.assertEqual(b["parent"], a["id"])
 
+    def test_wordpress(self):
 
-    a = db.comments.get(1)
+        xml = join(dirname(__file__), "wordpress.xml")
+        xxx = tempfile.NamedTemporaryFile()
 
-    assert a["author"] == "peter"
+        db = SQLite3(xxx.name, Config.load(None))
-    assert a["email"] == "foo@bar.com"
+        WordPress(db, xml).migrate()
 
-    b = db.comments.get(2)
+        self.assertEqual(db.threads["/2014/test/"]["title"], "Hello, World!")
-    assert b["parent"] == a["id"]
+        self.assertEqual(db.threads["/2014/test/"]["id"], 1)
+
+        self.assertEqual(len(db.execute("SELECT id FROM comments").fetchall()), 6)
+
+        first = db.comments.get(1)
+        self.assertEqual(first["author"], "Ohai")
+        self.assertEqual(first["text"], "Erster!1")
+
+        second = db.comments.get(2)
+        self.assertEqual(second["author"], "Tester")
+        self.assertEqual(second["text"], "Zweiter.")
+
+        for i in (3, 4, 5):
+            self.assertEqual(db.comments.get(i)["parent"], second["id"])
+
+        last = db.comments.get(6)
+        self.assertEqual(last["author"], "Letzter :/")
+        self.assertEqual(last["parent"], None)

isso/tests/wordpress.xml

@@ -0,0 +1,119 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<rss version="2.0"
+     xmlns:content="http://purl.org/rss/1.0/modules/content/"
+     xmlns:dc="http://purl.org/dc/elements/1.1/"
+     xmlns:wp="http://wordpress.org/export/1.0/">
+    <!-- This WXR dump is incomplete! It only contains elements needed for an
+         import, a few are unused yet, but eventually useful later.
+
+         The <item> node is derived from a sort-of real-world WordPress blog,
+         but modified to test various things.
+    -->
+    <channel>
+        <item>
+            <title>Hello, World!</title>
+            <link>http://example.tld/2014/test/</link>
+            <pubDate>Tue, 14 Jan 2014 17:31:03 +0000</pubDate>
+            <dc:creator><![CDATA[Tester]]></dc:creator>
+            <wp:post_id>18</wp:post_id>
+            <wp:post_date>2014-01-14 17:31:03</wp:post_date>
+            <wp:post_date_gmt>2014-01-14 17:31:03</wp:post_date_gmt>
+            <wp:comment_status>open</wp:comment_status>
+            <wp:post_name>test</wp:post_name>
+            <wp:status>publish</wp:status>
+            <wp:post_type>post</wp:post_type>
+            <wp:comment>
+                <wp:comment_id>2</wp:comment_id>
+                <wp:comment_author><![CDATA[Ohai]]></wp:comment_author>
+                <wp:comment_author_email>test@example.org
+                </wp:comment_author_email>
+                <wp:comment_author_url>http://example.tld/</wp:comment_author_url>
+                <wp:comment_author_IP>::ffff:82.119.20.0</wp:comment_author_IP>
+                <wp:comment_date>2014-01-14 17:32:12</wp:comment_date>
+                <wp:comment_date_gmt>2014-01-14 17:32:12</wp:comment_date_gmt>
+                <wp:comment_content>
+                    <![CDATA[Erster!1]]></wp:comment_content>
+                <wp:comment_approved>1</wp:comment_approved>
+
+                <!-- what's that? -->
+                <wp:comment_type></wp:comment_type>
+
+                <wp:comment_parent>0</wp:comment_parent>
+                <wp:comment_user_id>0</wp:comment_user_id>
+            </wp:comment>
+            <wp:comment>
+                <wp:comment_id>6</wp:comment_id>
+                <wp:comment_author><![CDATA[Tester]]></wp:comment_author>
+                <wp:comment_author_email>info@posativ.org
+                </wp:comment_author_email>
+                <wp:comment_author_url></wp:comment_author_url>
+                <wp:comment_author_IP>::ffff:86.56.63.0</wp:comment_author_IP>
+                <wp:comment_date>2014-04-29 15:21:27</wp:comment_date>
+                <wp:comment_date_gmt>2014-04-29 15:21:27</wp:comment_date_gmt>
+                <wp:comment_content><![CDATA[Zweiter.]]></wp:comment_content>
+                <wp:comment_approved>1</wp:comment_approved>
+                <wp:comment_type></wp:comment_type>
+                <wp:comment_parent>0</wp:comment_parent>
+                <wp:comment_user_id>1</wp:comment_user_id>
+            </wp:comment>
+            <wp:comment>
+                <wp:comment_id>7</wp:comment_id>
+                <wp:comment_author><![CDATA[Tester]]></wp:comment_author>
+                <wp:comment_author_email>info@posativ.org
+                </wp:comment_author_email>
+                <wp:comment_author_url></wp:comment_author_url>
+                <wp:comment_author_IP>::ffff:86.56.63.0</wp:comment_author_IP>
+                <wp:comment_date_gmt>2014-04-29 15:21:35</wp:comment_date_gmt>
+                <wp:comment_content><![CDATA[Drölfter!]]></wp:comment_content>
+                <wp:comment_approved>1</wp:comment_approved>
+                <wp:comment_parent>6</wp:comment_parent>
+            </wp:comment>
+            <wp:comment>
+                <wp:comment_id>8</wp:comment_id>
+                <wp:comment_author><![CDATA[Tester]]></wp:comment_author>
+                <wp:comment_author_email>info@posativ.org
+                </wp:comment_author_email>
+                <wp:comment_author_url></wp:comment_author_url>
+                <wp:comment_author_IP>::ffff:86.56.63.0</wp:comment_author_IP>
+                <wp:comment_date>2014-04-29 15:21:45</wp:comment_date>
+                <wp:comment_date_gmt>2014-04-29 15:21:45</wp:comment_date_gmt>
+                <wp:comment_content>
+                    <![CDATA[Yet another reply.]]></wp:comment_content>
+                <wp:comment_approved>1</wp:comment_approved>
+                <wp:comment_type></wp:comment_type>
+                <wp:comment_parent>7</wp:comment_parent>
+                <wp:comment_user_id>1</wp:comment_user_id>
+            </wp:comment>
+            <wp:comment>
+                <wp:comment_id>9</wp:comment_id>
+                <wp:comment_author><![CDATA[Tester]]></wp:comment_author>
+                <wp:comment_author_email>info@posativ.org
+                </wp:comment_author_email>
+                <wp:comment_author_url></wp:comment_author_url>
+                <wp:comment_author_IP>::ffff:86.56.63.0</wp:comment_author_IP>
+                <wp:comment_date>2014-04-29 15:21:52</wp:comment_date>
+                <wp:comment_date_gmt>2014-04-29 15:21:52</wp:comment_date_gmt>
+                <wp:comment_content><![CDATA[...]]></wp:comment_content>
+                <wp:comment_approved>1</wp:comment_approved>
+                <wp:comment_type></wp:comment_type>
+                <wp:comment_parent>7</wp:comment_parent>
+                <wp:comment_user_id>1</wp:comment_user_id>
+            </wp:comment>
+            <wp:comment>
+                <wp:comment_id>10</wp:comment_id>
+                <wp:comment_author><![CDATA[Letzter :/]]></wp:comment_author>
+                <wp:comment_author_email>info@posativ.org
+                </wp:comment_author_email>
+                <wp:comment_author_url></wp:comment_author_url>
+                <wp:comment_author_IP>::ffff:86.56.63.0</wp:comment_author_IP>
+                <wp:comment_date>2014-04-29 15:21:56</wp:comment_date>
+                <wp:comment_date_gmt>2014-04-29 15:21:56</wp:comment_date_gmt>
+                <wp:comment_content><![CDATA[...]]></wp:comment_content>
+                <wp:comment_approved>1</wp:comment_approved>
+                <wp:comment_type></wp:comment_type>
+                <wp:comment_parent>0</wp:comment_parent>
+                <wp:comment_user_id>1</wp:comment_user_id>
+            </wp:comment>
+        </item>
+    </channel>
+</rss>