X-Git-Url: https://git.madduck.net/etc/taskwarrior.git/blobdiff_plain/088fc620ec1921c3e9ee2dcf8385393b53306824..b8c20c493eb7f74b312d79d3c15e65a1ff560949:/docs/index.rst

diff --git a/docs/index.rst b/docs/index.rst
index 44ef469..dfddeb0 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -36,6 +36,11 @@ The default location is the same as taskwarrior's::
 
     >>> tw = TaskWarrior(data_location='~/.task', create=True)
 
+The ``TaskWarrior`` instance will also use your .taskrc configuration (so that
+it recognizes the same UDAs as your task binary, uses the same configuration,
+etc.). To override the location of the .taskrc, use
+``taskrc_location=~/some/different/path``.
+
 Creating Tasks
 --------------
 
@@ -84,7 +89,7 @@ Attributes of task objects are accessible through indices, like so::
     >>> task['id']
     15
     >>> task['due']
-    datetime.datetime(2013, 12, 5, 0, 0)
+    datetime.datetime(2015, 2, 5, 0, 0, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
     >>> task['tags']
     ['work', 'servers']
 
@@ -239,6 +244,105 @@ same Python object::
     >>> task3 == task1
     True
 
+Accessing original values
+-------------------------
+
+To access the saved state of the Task, use dict-like access using the
+``original`` attribute:
+
+    >>> t = Task(tw, description="tidy up")
+    >>> t.save()
+    >>> t['description'] = "tidy up the kitchen and bathroom"
+    >>> t['description']
+    "tidy up the kitchen and bathroom"
+    >>> t.original['description']
+    "tidy up"
+
+When you save the task, original values are refreshed to reflect the
+saved state of the task:
+
+    >>> t.save()
+    >>> t.original['description']
+    "tidy up the kitchen and bathroom"
+
+Dealing with dates and time
+---------------------------
+
+Any timestamp-like attributes of the tasks are converted to timezone-aware
+datetime objects. To achieve this, Tasklib leverages ``pytz`` Python module,
+which brings the Olsen timezone databaze to Python.
+
+This shields you from annoying details of Daylight Saving Time shifts
+or conversion between different timezones. For example, to list all the
+tasks which are due midnight if you're currently in Berlin:
+
+    >>> myzone = pytz.timezone('Europe/Berlin')
+    >>> midnight = myzone.localize(datetime(2015,2,2,0,0,0))
+    >>> tw.tasks.filter(due__before=midnight)
+
+However, this is still a little bit tedious. That's why TaskWarrior object
+is capable of automatic timezone detection, using the ``tzlocal`` Python
+module. If your system timezone is set to 'Europe/Berlin', following example
+will work the same way as the previous one:
+
+    >>> tw.tasks.filter(due__before=datetime(2015,2,2,0,0,0))
+
+You can also use simple dates when filtering:
+
+    >>> tw.tasks.filter(due__before=date(2015,2,2))
+
+In such case, a 00:00:00 is used as the time component.
+
+Of course, you can use datetime naive objects when initializing Task object
+or assigning values to datetime atrributes:
+
+    >>> t = Task(tw, description="Buy new shoes", due=date(2015,2,5))
+    >>> t['due']
+    datetime.datetime(2015, 2, 5, 0, 0, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
+    >>> t['due'] = date(2015,2,6,15,15,15)
+    >>> t['due']
+    datetime.datetime(2015, 2, 6, 15, 15, 15, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
+
+However, since timezone-aware and timezone-naive datetimes are not comparable
+in Python, this can cause some unexpected behaviour:
+
+    >>> from datetime import datetime
+    >>> now = datetime.now()
+    >>> t = Task(tw, description="take out the trash now") 
+    >>> t['due'] = now
+    >>> now
+    datetime.datetime(2015, 2, 1, 19, 44, 4, 770001)
+    >>> t['due']
+    datetime.datetime(2015, 2, 1, 19, 44, 4, 770001, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
+    >>> t['due'] == now
+    Traceback (most recent call last):
+      File "<stdin>", line 1, in <module>
+      TypeError: can't compare offset-naive and offset-aware datetimes
+
+If you want to compare datetime aware value with datetime naive value, you need
+to localize the naive value first:
+
+    >>> from datetime import datetime
+    >>> from tasklib.task import local_zone
+    >>> now = local_zone.localize(datetime.now())
+    >>> t['due'] = now
+    >>> now
+    datetime.datetime(2015, 2, 1, 19, 44, 4, 770001, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
+    >>> t['due'] == now
+    True
+
+Also, note that it does not matter whether the timezone aware datetime objects
+are set in the same timezone:
+
+    >>> import pytz
+    >>> t['due']
+    datetime.datetime(2015, 2, 1, 19, 44, 4, 770001, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
+    >>> now.astimezone(pytz.utc)
+    datetime.datetime(2015, 2, 1, 18, 44, 4, 770001, tzinfo=<UTC>)
+    >>> t['due'] == now.astimezone(pytz.utc)
+    True
+
+
 Working with annotations
 ------------------------
 
@@ -253,7 +357,7 @@ Annotations have only defined ``entry`` and ``description`` values::
 
     >>> annotation = annotated_task['annotations'][0]
     >>> annotation['entry']
-    datetime.datetime(2015, 1, 3, 21, 13, 55)
+    datetime.datetime(2015, 1, 3, 21, 13, 55, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
     >>> annotation['description']
     u'Yeah, I am annotated!'
 
@@ -298,30 +402,76 @@ You can use ``config_override`` keyword argument to specify a dictionary of conf
 Setting custom configuration values
 -----------------------------------
 
-By default, TaskWarrior does not use any of configuration values stored in
-your .taskrc. To see what configuration values are passed to each executed
+By default, TaskWarrior uses configuration values stored in your .taskrc.
+To see what configuration value overrides are passed to each executed
 task command, have a peek into ``config`` attribute of ``TaskWarrior`` object::
 
     >>> tw.config
     {'confirmation': 'no', 'data.location': '/home/tbabej/.task'}
 
-To pass your own configuration, you just need to update this dictionary::
+To pass your own configuration overrides, you just need to update this dictionary::
 
     >>> tw.config.update({'hooks': 'off'})  # tasklib will not trigger hooks
 
+Creating hook scripts
+---------------------
+
+From version 2.4.0, TaskWarrior has support for hook scripts. Tasklib provides
+some very useful helpers to write those. With tasklib, writing these becomes
+a breeze::
+
+    #!/usr/bin/python
+
+    from tasklib.task import Task
+    task = Task.from_input()
+    # ... <custom logic>
+    print task.export_data()
+
+For example, plugin which would assign the priority "H" to any task containing
+three exclamation marks in the description, would go like this::
+
+    #!/usr/bin/python
+
+    from tasklib.task import Task
+    task = Task.from_input()
+
+    if "!!!" in task['description']:
+        task['priority'] = "H"
+
+    print task.export_data()
+
+Tasklib can automatically detect whether it's running in the ``on-modify`` event,
+which provides more input than ``on-add`` event and reads the data accordingly.
+
+This means the example above works both for ``on-add`` and ``on-modify`` events!
+
+Consenquently, you can create just one hook file for both ``on-add`` and
+``on-modify`` events, and you just need to create a symlink for the other one.
+This removes the need for maintaining two copies of the same code base and/or
+boilerplate code.
+
+In ``on-modify`` events, tasklib loads both the original version and the modified
+version of the task to the returned ``Task`` object. To access the original data
+(in read-only manner), use ``original`` dict-like attribute:
+
+    >>> t = Task.from_input()
+    >>> t['description']
+    "Modified description"
+    >>> t.original['description']
+    "Original description"
+
 Working with UDAs
 -----------------
 
-Since TaskWarrior does not read your .taskrc, you need to define any UDAs
-in the TaskWarrior's config dictionary, as described above.
-
-Let us demonstrate this on the same example as in the TaskWarrior's docs::
+Since TaskWarrior does read your .taskrc, you need not to define any UDAs
+in the TaskWarrior's config dictionary, as described above. Suppose we have
+a estimate UDA in the .taskrc::
 
-    >>> tw = TaskWarrior()
-    >>> tw.config.update({'uda.estimate.type': 'numeric'})
+    uda.estimate.type = numeric
 
-Now we can filter and create tasks using the estimate UDA::
+We can simply filter and create tasks using the estimate UDA out of the box::
 
+    >>> tw = TaskWarrior()
     >>> task = Task(tw, description="Long task", estimate=1000)
     >>> task.save()
     >>> task['id']
@@ -332,7 +482,7 @@ This is saved as UDA in the TaskWarrior::
     $ task 1 export
     {"id":1,"description":"Long task","estimate":1000, ...}
 
-As long as ``TaskWarrior``'s config is updated, we can approach UDAs as built in attributes::
+We can also speficy UDAs as arguments in the TaskFilter::
 
     >>> tw.tasks.filter(estimate=1000)
     Long task
@@ -340,9 +490,16 @@ As long as ``TaskWarrior``'s config is updated, we can approach UDAs as built in
 Syncing
 -------
 
-Syncing is not directly supported by tasklib, but it can be made to work in a similiar way
-as the UDAs. First we need to update the ``config`` dictionary by the values required for
-sync to work, and then we can run the sync command using the ``execute_command()`` method::
+If you have configurated the needed config variables in your .taskrc, syncing
+is as easy as::
+
+    >>> tw = TaskWarrior()
+    >>> tw.execute_command(['sync'])
+
+If you want to use non-standard server/credentials, you'll need to provide configuration
+overrides to the ``TaskWarrior`` instance. Update the ``config`` dictionary with the
+values you desire to override, and then we can run the sync command using
+the ``execute_command()`` method::
 
     >>> tw = TaskWarrior()
     >>> sync_config = {