People will burn me at the stake for suggesting this ☺ and I must warn you that this is not for large datasets or high-traffic sites; but for that tiny CMS you wrote for grandma's recipes it may do fine. It probably does not work with multiple processes.
The upside is that you can seamlessly persist almost any object without any dependencies or daemons and with very little code.
from __future__ import with_statement import shelve from os import path from cPickle import HIGHEST_PROTOCOL from contextlib import closing from flask import Flask SHELVE_DB = 'shelve.db' app = Flask(__name__) app.config.from_object(__name__) db = shelve.open(path.join(app.root_path, app.config['SHELVE_DB']), protocol=HIGHEST_PROTOCOL, writeback=True) @app.route('/<message>') def write_and_list(message): db.setdefault('messages', ) db['messages'].append(message) return app.response_class('\n'.join(db['messages']), mimetype='text/plain') with closing(db): app.run()
The db object works like a dict and as you see in our view we can add a normal list to it and don't need to do anything special to persist changes.
If you want something similar that scales better have a look at ZODB. I haven't actually used ZODB myself and there seem to be controversy regarding its utility, but it is in use by most Zope sites out there.
This snippet by Dag Odenhall can be used freely for anything you like. Consider it public domain.