mekk.rtm provides both RememberTheMilk client library (which you
can use to write client programs/libraries using relatively simple
API) and a few command-line utilities (which you can use to manage
your data, for example import, export, massively tag, move between
lists etc).
Contrary to two other python RememberTheMilk clients (RtmAPI and
pyrtm), mekk.rtm tries to provide elaborate API with enumerated
explicit parameters and structured results. It is also tested with
some unittests.
For detailed documentation see inline docs of mekk.rtm.RtmClient. Here is simple example (see also longer sample/sample_client.py in package sources, to get api key and shared secret visit this page):
from mekk.rtm import RtmClient, create_and_authorize_connector_prompting_for_api_key
APP_NAME = "mekk.rtm sample"
connector = create_and_authorize_connector_prompting_for_api_key(APP_NAME)
#
# Function above interactively prompts for api key and shared secret and
# preserves acquired data in keyring. This is useful for testing, but in
# normal case you are more likely to do:
#
# connector = create_and_authorize_connector(APP_NAME, API_KEY, SHARED_SECRET)
client = RtmClient(connector)
print "Normal lists: "
for l in client.known_lists():
if not l.archived:
print u"%s (%s)" % (l.id, l.name)
print "Smart lists: "
for l in client.known_smart_lists():
if not l.archived:
print u"%s (%s, %s)" % (l.id, l.name, l.filter)
test_list = client.find_or_create_list(u"The testing list")
print "Test list: ", test_list
another_list = client.find_or_create_list(u"The testing list")
print "Another list: ", another_list
task1 = client.create_task(
u"Write some unit tests",
list_id = test_list.id,
tags = ["testy", "@dom"],
priority = 3,
due_date = "2010-11-30",
estimate = "1 day 10 hours",
repeat = "after 3 days",
url = "http://en.wikipedia.org/wiki/Unit_testing",
completed = False,
notes = [ (u"Runner", u"Use nose to run them all\nIt is simplest"),
(u"Helper", u"And mock can help to wrap backend apis\nwithout calling them") ])
print "Created task", task1
task2 = client.create_task(
u"Less serious one",
list_id = test_list.id)
print "Created task", task2
task3 = client.create_task(
u"Less serious one",
list_id = another_list.id,
tags = ["testy"])
print "Created task", task3
print "All incomplete tasks with notes on first list:"
for t in client.find_tasks(list_id = test_list.id, filter = "status:incomplete and hasNotes:true"):
print t
task2 = client.update_task(
task2.key,
completed = True)
print "Updated task", task2
print "All incomplete tasks tagged 'testy':"
for t in client.find_tasks(filter = "tag:testy and status:incomplete"):
print t
mekk.rtm offers a few command-line helpers (mostly related to the
things original web interface fails to provide).
The command makes it possible to add and/or remove specific tag(s) to many tasks at once. Example usage:
rtmtag --add=work --add=computer \
--remove=office \
--filter='list:"Prepared Reports" and status:"incomplete"'
(every -add names tag to add, every --remove names tag to remove, --filter defines which tasks are to be updated using normal RememberTheMilk query syntax).
To verify what the command is to do without actually performing it one can add --dry-run and --verbose. For example:
rtmtag --add=rodzina \
--filter='list:"Rodzina" and status:"incomplete"' \
--dry-run --verbose
Moves all tasks fitting given query to list of specified name (which will be created if is not present).
Example:
rtmmove --list='Language learning' \
--filter='list:"English Course" or tag:"French" or tag:German"
To verify what the command is to do without actually performing it one can add --dry-run and --verbose. For example:
rtmmove --list=Family \
--filter='(tag:paul or tag:jane) and status:"incomplete"' \
--dry-run --verbose
Not yet implemented (but planned soon)
Not yet implemented (but planned soon)
For obvious reasons (I already imported my data) I don't test this module on every release. Make dry run before actually running the import and report bugs if they happen.
Prepare .json export of Nozbe data. For details see mekk.nozbe but usually you just want to:
nozbetool export --user=MyNozbeName --json=mynozbedata.json
First make a test run:
rtmimport --nozbe-json=mynozbedata.json --verbose --dry-run
(it does not store anything, just prints what it is to do) and verify whether everything seems correct.
Then make actual import:
rtmimport --nozbe-json=mynozbedata.json --verbose
(or omit --verbose if you don't want to track progress, but I recommend you keep it)
Note: import can take some time. In case of my big list over not-so-good network it has been running for almost an hour.
Nozbe projects are saved as RememberTheMilk lists.
Nozbe contexts are converted to RememberTheMilk tags. @ is prepended to their names and non-alphanumeric characters are replaced with dashes (so for example My home/kitchen becomes @My-home-kitchen).
Next actions are tagged as Next.
Actions are saved as tasks. Name, due date, recurrence, estimated cost and completion status are all saved.
In case of recurrence, RTM every mode is used (so the task marked
on Nozbe as recurring every week will be spawned 52 times a year by
RTM, whether user completes it, or not). If you prefer alternative way
(spawning new incarnation whenever previous is completed), edit tasks
after import, patch the code (and replace every with after), or
ask me for a commandline flag).
As notes are bound to projects on Nozbe, and to tasks on RememberTheMilk, I save notes by creating artificial tasks named "Save this note" (one per every list for which appropriate project had notes) and binding notes to those tasks. This must be handled afterwards using RTM interface, to make sure it happens I mark those tasks as due immediately. Those task are also tagged as Note.
Only main context is copied, additional contexts are lost. I don't know how to grab them from Nozbe (in case somebody knows how to patch mekk.nozbe to grab all contexts, I can extend this importer easily to handle them all).
Uploads are not copied at all. I neither now how to export them from Nozbe, nor how could I handle them afterwards (RTM has no uploads).
Action name formatting is not available on RememberTheMilk, so if you used constructs like Visit "the website":http://google.com, they will show up as is.
Sharing information (= information about delegations to other users) is lost. I haven't used this feature so I don't know how do the underlying data look like.
Some contexts could probably be converted to locations, not tags, but I don't have an idea how to decide which way to go.
If import process is interrupted, re-running it would make duplicate tasks (there is no duplicate checking, it would be costly). The best way to resolve it is to open .json file in text editor (after making a backup copy of it) and simply cut all actions which are already saved (they are saved in order, so it is just a matter of locating the last action saved before the process was interrupted and cutting all actions up to this one).
I borrowed many ideas (especially how to handle authorization and construct requests) from RtmAPI by Michael Grünewald. Great thanks for sharing the code.
Development is tracked on BitBucket. Clone from there, report bugs there, offer patches there.