From 68e18ad1f44e9a6fab66adc38e97d027a58de8a4 Mon Sep 17 00:00:00 2001
From: Rob Austein <sra@hactrn.net>
Date: Sun, 14 Feb 2021 16:56:57 +0000
Subject: Another reorg, and pelican samples

---
 extract.py | 105 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 105 insertions(+)
 create mode 100755 extract.py

(limited to 'extract.py')

diff --git a/extract.py b/extract.py
new file mode 100755
index 0000000..7f76fe9
--- /dev/null
+++ b/extract.py
@@ -0,0 +1,105 @@
+#!/usr/bin/env python3
+
+import fnmatch
+import hashlib
+import json
+import os
+import shutil
+import sqlite3
+import sys
+import time
+import urllib.parse
+
+import trac2md
+
+wiki_query = '''
+  SELECT
+    name, 
+    version, 
+    time / 1000000 AS time, 
+    text 
+  FROM wiki
+  ORDER BY
+    name, version
+'''
+
+attachment_query = '''
+  SELECT
+    id,
+    filename,
+    size,
+    author,
+    description,
+    ipnr,
+    time / 1000000 AS createdtime
+  FROM
+    attachment
+  WHERE
+    type = 'wiki'
+  ORDER BY
+    filename, time
+'''
+
+def isotime(t):
+    return None if t == 0 else time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime(t))
+
+def attachment_link(row):
+    h   = lambda whatever: hashlib.sha1(whatever.encode()).hexdigest()
+    h1  = h(row.id)
+    h2  = h(row.filename)
+    fn2 = os.path.splitext(row["filename"])[1]
+    return \
+        os.path.join("attachments", "wiki", h1[:3], h1, h2 + fn2), \
+        os.path.join(urllib.parse.quote(row.id, ""), urllib.parse.quote(row.filename, ""))
+
+class Filter:
+
+    def __init__(self, filename = "filter.json"):
+        with open(filename) as f:
+            filter = json.load(f)
+        if not all(action in "-+" for action, pattern in filter):
+            sys.exit("Bad action \"{}\" in filter".format(action))
+        self.filter = tuple((action == "+", pattern) for action, pattern in filter)
+
+    def __call__(self, name):
+        for action, pattern in self.filter:
+            if fnmatch.fnmatch(name, pattern):
+                return action
+        return True
+
+class Row(sqlite3.Row):
+    def __getattr__(self, name):
+        return self[name]
+
+def main():
+
+    for dn in ("wiki", "pelican"):
+        shutil.rmtree(dn)
+
+    for dn in ("wiki", "pelican/content/images", "pelican/content/pages"):
+        os.makedirs(dn)
+
+    #os.link("pelican.conf", "pelican/pelican.conf")
+
+    keep = Filter()
+
+    db = sqlite3.connect("trac.db")
+    db.row_factory = Row
+
+    for row in db.execute(wiki_query):
+        if keep(row.name):
+            slug = urllib.parse.quote(row.name, "")
+            print(slug, row.version)
+            with open("wiki/{}.trac".format(slug), "w") as f:
+                f.write(row.text)
+            md = trac2md.WikiToMD(row.text)
+            with open("pelican/content/{}.md".format(slug), "w") as f:
+                f.write(md)
+
+    for row in db.execute(attachment_query):
+        print("{} => {}".format(*attachment_link(row)))
+
+    db.close()
+
+if __name__ == "__main__":
+    main()
-- 
cgit v1.2.3