]> git.madduck.net Git - etc/taskwarrior.git/blobdiff - tasklib/task.py

madduck's git repository

Every one of the projects in this repository is available at the canonical URL git://git.madduck.net/madduck/pub/<projectpath> — see each project's metadata for the exact URL.

All patches and comments are welcome. Please squash your changes to logical commits before using git-format-patch and git-send-email to patches@git.madduck.net. If you'd read over the Git project's submission guidelines and adhered to them, I'd be especially grateful.

SSH access, as well as push access can be individually arranged.

If you use my repositories frequently, consider adding the following snippet to ~/.gitconfig and using the third clone URL listed for each project:

[url "git://git.madduck.net/madduck/"]
  insteadOf = madduck:

backends: Add complete_task to the backend interface
[etc/taskwarrior.git] / tasklib / task.py
index 51c97d31736dc0a9e3d6aa631f940c3e6fae2344..b7bba4bd541df0ef12acd63c9548af8fab6add93 100644 (file)
@@ -7,25 +7,49 @@ import os
 import pytz
 import six
 import sys
 import pytz
 import six
 import sys
-import subprocess
 import tzlocal
 
 import tzlocal
 
+from backends import TaskWarrior, TaskWarriorException
+
 DATE_FORMAT = '%Y%m%dT%H%M%SZ'
 DATE_FORMAT = '%Y%m%dT%H%M%SZ'
+DATE_FORMAT_CALC = '%Y-%m-%dT%H:%M:%S'
 REPR_OUTPUT_SIZE = 10
 PENDING = 'pending'
 COMPLETED = 'completed'
 
 REPR_OUTPUT_SIZE = 10
 PENDING = 'pending'
 COMPLETED = 'completed'
 
-VERSION_2_1_0 = six.u('2.1.0')
-VERSION_2_2_0 = six.u('2.2.0')
-VERSION_2_3_0 = six.u('2.3.0')
-VERSION_2_4_0 = six.u('2.4.0')
-
 logger = logging.getLogger(__name__)
 local_zone = tzlocal.get_localzone()
 
 
 logger = logging.getLogger(__name__)
 local_zone = tzlocal.get_localzone()
 
 
-class TaskWarriorException(Exception):
-    pass
+class ReadOnlyDictView(object):
+    """
+    Provides simplified read-only view upon dict object.
+    """
+
+    def __init__(self, viewed_dict):
+        self.viewed_dict = viewed_dict
+
+    def __getitem__(self, key):
+        return copy.deepcopy(self.viewed_dict.__getitem__(key))
+
+    def __contains__(self, k):
+        return self.viewed_dict.__contains__(k)
+
+    def __iter__(self):
+        for value in self.viewed_dict:
+            yield copy.deepcopy(value)
+
+    def __len__(self):
+        return len(self.viewed_dict)
+
+    def get(self, key, default=None):
+        return copy.deepcopy(self.viewed_dict.get(key, default))
+
+    def items(self):
+        return [copy.deepcopy(v) for v in self.viewed_dict.items()]
+
+    def values(self):
+        return [copy.deepcopy(v) for v in self.viewed_dict.values()]
 
 
 class SerializingObject(object):
 
 
 class SerializingObject(object):
@@ -44,8 +68,20 @@ class SerializingObject(object):
         not export empty-valued attributes) if the attribute
         is not iterable (e.g. list or set), in which case
         a empty iterable should be used.
         not export empty-valued attributes) if the attribute
         is not iterable (e.g. list or set), in which case
         a empty iterable should be used.
+
+    Normalizing methods should hold the following contract:
+      - They are used to validate and normalize the user input.
+        Any attribute value that comes from the user (during Task
+        initialization, assignign values to Task attributes, or
+        filtering by user-provided values of attributes) is first
+        validated and normalized using the normalize_{key} method.
+      - If validation or normalization fails, normalizer is expected
+        to raise ValueError.
     """
 
     """
 
+    def __init__(self, warrior):
+        self.warrior = warrior
+
     def _deserialize(self, key, value):
         hydrate_func = getattr(self, 'deserialize_{0}'.format(key),
                                lambda x: x if x != '' else None)
     def _deserialize(self, key, value):
         hydrate_func = getattr(self, 'deserialize_{0}'.format(key),
                                lambda x: x if x != '' else None)
@@ -63,6 +99,10 @@ class SerializingObject(object):
         or entered as a value of Task attribute.
         """
 
         or entered as a value of Task attribute.
         """
 
+        # None value should not be converted by normalizer
+        if value is None:
+            return None
+
         normalize_func = getattr(self, 'normalize_{0}'.format(key),
                                  lambda x: x)
 
         normalize_func = getattr(self, 'normalize_{0}'.format(key),
                                  lambda x: x)
 
@@ -105,6 +145,24 @@ class SerializingObject(object):
     def normalize_modified(self, value):
         return self.datetime_normalizer(value)
 
     def normalize_modified(self, value):
         return self.datetime_normalizer(value)
 
+    def serialize_start(self, value):
+        return self.timestamp_serializer(value)
+
+    def deserialize_start(self, value):
+        return self.timestamp_deserializer(value)
+
+    def normalize_start(self, value):
+        return self.datetime_normalizer(value)
+
+    def serialize_end(self, value):
+        return self.timestamp_serializer(value)
+
+    def deserialize_end(self, value):
+        return self.timestamp_deserializer(value)
+
+    def normalize_end(self, value):
+        return self.datetime_normalizer(value)
+
     def serialize_due(self, value):
         return self.timestamp_serializer(value)
 
     def serialize_due(self, value):
         return self.timestamp_serializer(value)
 
@@ -167,8 +225,15 @@ class SerializingObject(object):
         return ','.join(task['uuid'] for task in value)
 
     def deserialize_depends(self, raw_uuids):
         return ','.join(task['uuid'] for task in value)
 
     def deserialize_depends(self, raw_uuids):
-        raw_uuids = raw_uuids or ''  # Convert None to empty string
-        uuids = raw_uuids.split(',')
+        raw_uuids = raw_uuids or []  # Convert None to empty list
+
+        # TW 2.4.4 encodes list of dependencies as a single string
+        if type(raw_uuids) is not list:
+            uuids = raw_uuids.split(',')
+        # TW 2.4.5 and later exports them as a list, no conversion needed
+        else:
+            uuids = raw_uuids
+
         return set(self.warrior.tasks.get(uuid=uuid) for uuid in uuids if uuid)
 
     def datetime_normalizer(self, value):
         return set(self.warrior.tasks.get(uuid=uuid) for uuid in uuids if uuid)
 
     def datetime_normalizer(self, value):
@@ -186,16 +251,36 @@ class SerializingObject(object):
             # Convert to local midnight
             value_full = datetime.datetime.combine(value, datetime.time.min)
             localized = local_zone.localize(value_full)
             # Convert to local midnight
             value_full = datetime.datetime.combine(value, datetime.time.min)
             localized = local_zone.localize(value_full)
-        elif isinstance(value, datetime.datetime) and value.tzinfo is None:
-            # Convert to localized datetime object
-            localized = local_zone.localize(value)
+        elif isinstance(value, datetime.datetime):
+            if value.tzinfo is None:
+                # Convert to localized datetime object
+                localized = local_zone.localize(value)
+            else:
+                # If the value is already localized, there is no need to change
+                # time zone at this point. Also None is a valid value too.
+                localized = value
+        elif (isinstance(value, six.string_types)
+                and self.warrior.version >= VERSION_2_4_0):
+            # For strings, use 'task calc' to evaluate the string to datetime
+            # available since TW 2.4.0
+            args = value.split()
+            result = self.warrior.execute_command(['calc'] + args)
+            naive = datetime.datetime.strptime(result[0], DATE_FORMAT_CALC)
+            localized = local_zone.localize(naive)
         else:
         else:
-            # If the value is already localized, there is no need to change
-            # time zone at this point. Also None is a valid value too.
-            localized = value
-        
+            raise ValueError("Provided value could not be converted to "
+                             "datetime, its type is not supported: {}"
+                             .format(type(value)))
+
         return localized
         return localized
-            
+
+    def normalize_uuid(self, value):
+        # Enforce sane UUID
+        if not isinstance(value, six.string_types) or value == '':
+            raise ValueError("UUID must be a valid non-empty string, "
+                             "not: {}".format(value))
+
+        return value
 
 
 class TaskResource(SerializingObject):
 
 
 class TaskResource(SerializingObject):
@@ -208,7 +293,7 @@ class TaskResource(SerializingObject):
         # are not propagated.
         self._original_data = copy.deepcopy(self._data)
 
         # are not propagated.
         self._original_data = copy.deepcopy(self._data)
 
-    def _update_data(self, data, update_original=False):
+    def _update_data(self, data, update_original=False, remove_missing=False):
         """
         Low level update of the internal _data dict. Data which are coming as
         updates should already be serialized. If update_original is True, the
         """
         Low level update of the internal _data dict. Data which are coming as
         updates should already be serialized. If update_original is True, the
@@ -217,6 +302,11 @@ class TaskResource(SerializingObject):
         self._data.update(dict((key, self._deserialize(key, value))
                                for key, value in data.items()))
 
         self._data.update(dict((key, self._deserialize(key, value))
                                for key, value in data.items()))
 
+        # In certain situations, we want to treat missing keys as removals
+        if remove_missing:
+            for key in set(self._data.keys()) - set(data.keys()):
+                self._data[key] = None
+
         if update_original:
             self._original_data = copy.deepcopy(self._data)
 
         if update_original:
             self._original_data = copy.deepcopy(self._data)
 
@@ -290,9 +380,10 @@ class TaskResource(SerializingObject):
 class TaskAnnotation(TaskResource):
     read_only_fields = ['entry', 'description']
 
 class TaskAnnotation(TaskResource):
     read_only_fields = ['entry', 'description']
 
-    def __init__(self, task, data={}):
+    def __init__(self, task, data=None):
         self.task = task
         self.task = task
-        self._load_data(data)
+        self._load_data(data or dict())
+        super(TaskAnnotation, self).__init__(task.warrior)
 
     def remove(self):
         self.task.remove_annotation(self)
 
     def remove(self):
         self.task.remove_annotation(self)
@@ -326,6 +417,18 @@ class Task(TaskResource):
         """
         pass
 
         """
         pass
 
+    class ActiveTask(Exception):
+        """
+        Raised when the operation cannot be performed on the active task.
+        """
+        pass
+
+    class InactiveTask(Exception):
+        """
+        Raised when the operation cannot be performed on an inactive task.
+        """
+        pass
+
     class NotSaved(Exception):
         """
         Raised when the operation cannot be performed on the task, because
     class NotSaved(Exception):
         """
         Raised when the operation cannot be performed on the task, because
@@ -334,7 +437,7 @@ class Task(TaskResource):
         pass
 
     @classmethod
         pass
 
     @classmethod
-    def from_input(cls, input_file=sys.stdin, modify=None):
+    def from_input(cls, input_file=sys.stdin, modify=None, warrior=None):
         """
         Creates a Task object, directly from the stdin, by reading one line.
         If modify=True, two lines are used, first line interpreted as the
         """
         Creates a Task object, directly from the stdin, by reading one line.
         If modify=True, two lines are used, first line interpreted as the
@@ -350,25 +453,31 @@ class Task(TaskResource):
         but defaults to sys.stdin.
         """
 
         but defaults to sys.stdin.
         """
 
-        # TaskWarrior instance is set to None
-        task = cls(None)
-
         # Detect the hook type if not given directly
         name = os.path.basename(sys.argv[0])
         modify = name.startswith('on-modify') if modify is None else modify
 
         # Detect the hook type if not given directly
         name = os.path.basename(sys.argv[0])
         modify = name.startswith('on-modify') if modify is None else modify
 
+        # Create the TaskWarrior instance if none passed
+        if warrior is None:
+            hook_parent_dir = os.path.dirname(os.path.dirname(sys.argv[0]))
+            warrior = TaskWarrior(data_location=hook_parent_dir)
+
+        # TaskWarrior instance is set to None
+        task = cls(warrior)
+
         # Load the data from the input
         task._load_data(json.loads(input_file.readline().strip()))
 
         # If this is a on-modify event, we are provided with additional
         # line of input, which provides updated data
         if modify:
         # Load the data from the input
         task._load_data(json.loads(input_file.readline().strip()))
 
         # If this is a on-modify event, we are provided with additional
         # line of input, which provides updated data
         if modify:
-            task._update_data(json.loads(input_file.readline().strip()))
+            task._update_data(json.loads(input_file.readline().strip()),
+                              remove_missing=True)
 
         return task
 
     def __init__(self, warrior, **kwargs):
 
         return task
 
     def __init__(self, warrior, **kwargs):
-        self.warrior = warrior
+        super(Task, self).__init__(warrior)
 
         # Check that user is not able to set read-only value in __init__
         for key in kwargs.keys():
 
         # Check that user is not able to set read-only value in __init__
         for key in kwargs.keys():
@@ -384,6 +493,9 @@ class Task(TaskResource):
                           for (key, value) in six.iteritems(kwargs))
         self._original_data = copy.deepcopy(self._data)
 
                           for (key, value) in six.iteritems(kwargs))
         self._original_data = copy.deepcopy(self._data)
 
+        # Provide read only access to the original data
+        self.original = ReadOnlyDictView(self._original_data)
+
     def __unicode__(self):
         return self['description']
 
     def __unicode__(self):
         return self['description']
 
@@ -420,6 +532,10 @@ class Task(TaskResource):
     def pending(self):
         return self['status'] == six.text_type('pending')
 
     def pending(self):
         return self['status'] == six.text_type('pending')
 
+    @property
+    def active(self):
+        return self['start'] is not None
+
     @property
     def saved(self):
         return self['uuid'] is not None or self['id'] is not None
     @property
     def saved(self):
         return self['uuid'] is not None or self['id'] is not None
@@ -458,7 +574,7 @@ class Task(TaskResource):
         if self.warrior.version < VERSION_2_4_0:
             return self._data['description']
         else:
         if self.warrior.version < VERSION_2_4_0:
             return self._data['description']
         else:
-            return "description:'{0}'".format(self._data['description'] or '')
+            return six.u("description:'{0}'").format(self._data['description'] or '')
 
     def delete(self):
         if not self.saved:
 
     def delete(self):
         if not self.saved:
@@ -470,11 +586,44 @@ class Task(TaskResource):
         if self.deleted:
             raise Task.DeletedTask("Task was already deleted")
 
         if self.deleted:
             raise Task.DeletedTask("Task was already deleted")
 
-        self.warrior.execute_command([self['uuid'], 'delete'])
+        self.backend.delete_task(self)
 
         # Refresh the status again, so that we have updated info stored
 
         # Refresh the status again, so that we have updated info stored
+        self.refresh(only_fields=['status', 'start', 'end'])
+
+    def start(self):
+        if not self.saved:
+            raise Task.NotSaved("Task needs to be saved before it can be started")
+
+        # Refresh, and raise exception if task is already completed/deleted
         self.refresh(only_fields=['status'])
 
         self.refresh(only_fields=['status'])
 
+        if self.completed:
+            raise Task.CompletedTask("Cannot start a completed task")
+        elif self.deleted:
+            raise Task.DeletedTask("Deleted task cannot be started")
+        elif self.active:
+            raise Task.ActiveTask("Task is already active")
+
+        self.backend.start_task(self)
+
+        # Refresh the status again, so that we have updated info stored
+        self.refresh(only_fields=['status', 'start'])
+
+    def stop(self):
+        if not self.saved:
+            raise Task.NotSaved("Task needs to be saved before it can be stopped")
+
+        # Refresh, and raise exception if task is already completed/deleted
+        self.refresh(only_fields=['status'])
+
+        if not self.active:
+            raise Task.InactiveTask("Cannot stop an inactive task")
+
+        self.backend.stop_task(self)
+
+        # Refresh the status again, so that we have updated info stored
+        self.refresh(only_fields=['status', 'start'])
 
     def done(self):
         if not self.saved:
 
     def done(self):
         if not self.saved:
@@ -488,36 +637,21 @@ class Task(TaskResource):
         elif self.deleted:
             raise Task.DeletedTask("Deleted task cannot be completed")
 
         elif self.deleted:
             raise Task.DeletedTask("Deleted task cannot be completed")
 
+        # Older versions of TW do not stop active task at completion
+        if self.warrior.version < VERSION_2_4_0 and self.active:
+            self.stop()
+
         self.warrior.execute_command([self['uuid'], 'done'])
 
         # Refresh the status again, so that we have updated info stored
         self.warrior.execute_command([self['uuid'], 'done'])
 
         # Refresh the status again, so that we have updated info stored
-        self.refresh(only_fields=['status'])
+        self.refresh(only_fields=['status', 'start', 'end'])
 
     def save(self):
         if self.saved and not self.modified:
             return
 
 
     def save(self):
         if self.saved and not self.modified:
             return
 
-        args = [self['uuid'], 'modify'] if self.saved else ['add']
-        args.extend(self._get_modified_fields_as_args())
-        output = self.warrior.execute_command(args)
-
-        # Parse out the new ID, if the task is being added for the first time
-        if not self.saved:
-            id_lines = [l for l in output if l.startswith('Created task ')]
-
-            # Complain loudly if it seems that more tasks were created
-            # Should not happen
-            if len(id_lines) != 1 or len(id_lines[0].split(' ')) != 3:
-                raise TaskWarriorException("Unexpected output when creating "
-                                           "task: %s" % '\n'.join(id_lines))
-
-            # Circumvent the ID storage, since ID is considered read-only
-            self._data['id'] = int(id_lines[0].split(' ')[2].rstrip('.'))
-
-        # Refreshing is very important here, as not only modification time
-        # is updated, but arbitrary attribute may have changed due hooks
-        # altering the data before saving
-        self.refresh()
+        # All the actual work is done by the backend
+        self.backend.save_task(self)
 
     def add_annotation(self, annotation):
         if not self.saved:
 
     def add_annotation(self, annotation):
         if not self.saved:
@@ -550,9 +684,9 @@ class Task(TaskResource):
             if serialized_value is '':
                 escaped_serialized_value = ''
             else:
             if serialized_value is '':
                 escaped_serialized_value = ''
             else:
-                escaped_serialized_value = "'{0}'".format(serialized_value)
+                escaped_serialized_value = six.u("'{0}'").format(serialized_value)
 
 
-            format_default = lambda: "{0}:{1}".format(field,
+            format_default = lambda: six.u("{0}:{1}").format(field,
                                                       escaped_serialized_value)
 
             format_func = getattr(self, 'format_{0}'.format(field),
                                                       escaped_serialized_value)
 
             format_func = getattr(self, 'format_{0}'.format(field),
@@ -573,7 +707,7 @@ class Task(TaskResource):
 
         return args
 
 
         return args
 
-    def refresh(self, only_fields=[]):
+    def refresh(self, only_fields=None, after_save=False):
         # Raise error when trying to refresh a task that has not been saved
         if not self.saved:
             raise Task.NotSaved("Task needs to be saved to be refreshed")
         # Raise error when trying to refresh a task that has not been saved
         if not self.saved:
             raise Task.NotSaved("Task needs to be saved to be refreshed")
@@ -582,7 +716,39 @@ class Task(TaskResource):
         # of newly saved tasks. Any other place in the code is fine
         # with using UUID only.
         args = [self['uuid'] or self['id'], 'export']
         # of newly saved tasks. Any other place in the code is fine
         # with using UUID only.
         args = [self['uuid'] or self['id'], 'export']
-        new_data = json.loads(self.warrior.execute_command(args)[0])
+        output = self.warrior.execute_command(args)
+
+        def valid(output):
+            return len(output) == 1 and output[0].startswith('{')
+
+        # For older TW versions attempt to uniquely locate the task
+        # using the data we have if it has been just saved.
+        # This can happen when adding a completed task on older TW versions.
+        if (not valid(output) and self.warrior.version < VERSION_2_4_5
+                and after_save):
+
+            # Make a copy, removing ID and UUID. It's most likely invalid
+            # (ID 0) if it failed to match a unique task.
+            data = copy.deepcopy(self._data)
+            data.pop('id', None)
+            data.pop('uuid', None)
+
+            taskfilter = TaskFilter(self.warrior)
+            for key, value in data.items():
+                taskfilter.add_filter_param(key, value)
+
+            output = self.warrior.execute_command(['export', '--'] +
+                taskfilter.get_filter_params())
+
+        # If more than 1 task has been matched still, raise an exception
+        if not valid(output):
+            raise TaskWarriorException(
+                "Unique identifiers {0} with description: {1} matches "
+                "multiple tasks: {2}".format(
+                self['uuid'] or self['id'], self['description'], output)
+            )
+
+        new_data = json.loads(output[0])
         if only_fields:
             to_update = dict(
                 [(k, new_data.get(k)) for k in only_fields])
         if only_fields:
             to_update = dict(
                 [(k, new_data.get(k)) for k in only_fields])
@@ -595,8 +761,9 @@ class TaskFilter(SerializingObject):
     A set of parameters to filter the task list with.
     """
 
     A set of parameters to filter the task list with.
     """
 
-    def __init__(self, filter_params=[]):
-        self.filter_params = filter_params
+    def __init__(self, warrior, filter_params=None):
+        self.filter_params = filter_params or []
+        super(TaskFilter, self).__init__(warrior)
 
     def add_filter(self, filter_str):
         self.filter_params.append(filter_str)
 
     def add_filter(self, filter_str):
         self.filter_params.append(filter_str)
@@ -609,7 +776,7 @@ class TaskFilter(SerializingObject):
         attribute_key = key.split('.')[0]
 
         # Since this is user input, we need to normalize before we serialize
         attribute_key = key.split('.')[0]
 
         # Since this is user input, we need to normalize before we serialize
-        value = self._normalize(key, value)
+        value = self._normalize(attribute_key, value)
         value = self._serialize(attribute_key, value)
 
         # If we are filtering by uuid:, do not use uuid keyword
         value = self._serialize(attribute_key, value)
 
         # If we are filtering by uuid:, do not use uuid keyword
@@ -622,16 +789,18 @@ class TaskFilter(SerializingObject):
 
             # We enforce equality match by using 'is' (or 'none') modifier
             # Without using this syntax, filter fails due to TW-1479
 
             # We enforce equality match by using 'is' (or 'none') modifier
             # Without using this syntax, filter fails due to TW-1479
-            modifier = '.is' if value else '.none'
-            key = key + modifier if '.' not in key else key
+            # which is, however, fixed in 2.4.5
+            if self.warrior.version < VERSION_2_4_5:
+                modifier = '.is' if value else '.none'
+                key = key + modifier if '.' not in key else key
 
 
-            self.filter_params.append("{0}:{1}".format(key, value))
+            self.filter_params.append(six.u("{0}:{1}").format(key, value))
 
     def get_filter_params(self):
         return [f for f in self.filter_params if f]
 
     def clone(self):
 
     def get_filter_params(self):
         return [f for f in self.filter_params if f]
 
     def clone(self):
-        c = self.__class__()
+        c = self.__class__(self.warrior)
         c.filter_params = list(self.filter_params)
         return c
 
         c.filter_params = list(self.filter_params)
         return c
 
@@ -644,7 +813,7 @@ class TaskQuerySet(object):
     def __init__(self, warrior=None, filter_obj=None):
         self.warrior = warrior
         self._result_cache = None
     def __init__(self, warrior=None, filter_obj=None):
         self.warrior = warrior
         self._result_cache = None
-        self.filter_obj = filter_obj or TaskFilter()
+        self.filter_obj = filter_obj or TaskFilter(warrior)
 
     def __deepcopy__(self, memo):
         """
 
     def __deepcopy__(self, memo):
         """
@@ -744,81 +913,3 @@ class TaskQuerySet(object):
         raise ValueError(
             'get() returned more than one Task -- it returned {0}! '
             'Lookup parameters were {1}'.format(num, kwargs))
         raise ValueError(
             'get() returned more than one Task -- it returned {0}! '
             'Lookup parameters were {1}'.format(num, kwargs))
-
-
-class TaskWarrior(object):
-    def __init__(self, data_location='~/.task', create=True):
-        data_location = os.path.expanduser(data_location)
-        if create and not os.path.exists(data_location):
-            os.makedirs(data_location)
-        self.config = {
-            'data.location': os.path.expanduser(data_location),
-            'confirmation': 'no',
-            'dependency.confirmation': 'no',  # See TW-1483 or taskrc man page
-            'recurrence.confirmation': 'no',  # Necessary for modifying R tasks
-        }
-        self.tasks = TaskQuerySet(self)
-        self.version = self._get_version()
-
-    def _get_command_args(self, args, config_override={}):
-        command_args = ['task', 'rc:/']
-        config = self.config.copy()
-        config.update(config_override)
-        for item in config.items():
-            command_args.append('rc.{0}={1}'.format(*item))
-        command_args.extend(map(str, args))
-        return command_args
-
-    def _get_version(self):
-        p = subprocess.Popen(
-                ['task', '--version'],
-                stdout=subprocess.PIPE,
-                stderr=subprocess.PIPE)
-        stdout, stderr = [x.decode('utf-8') for x in p.communicate()]
-        return stdout.strip('\n')
-
-    def execute_command(self, args, config_override={}, allow_failure=True):
-        command_args = self._get_command_args(
-            args, config_override=config_override)
-        logger.debug(' '.join(command_args))
-        p = subprocess.Popen(command_args, stdout=subprocess.PIPE,
-                             stderr=subprocess.PIPE)
-        stdout, stderr = [x.decode('utf-8') for x in p.communicate()]
-        if p.returncode and allow_failure:
-            if stderr.strip():
-                error_msg = stderr.strip().splitlines()[-1]
-            else:
-                error_msg = stdout.strip()
-            raise TaskWarriorException(error_msg)
-        return stdout.strip().split('\n')
-
-    def enforce_recurrence(self):
-        # Run arbitrary report command which will trigger generation
-        # of recurrent tasks.
-        # TODO: Make a version dependant enforcement once
-        #       TW-1531 is handled
-        self.execute_command(['next'], allow_failure=False)
-
-    def filter_tasks(self, filter_obj):
-        self.enforce_recurrence()
-        args = ['export', '--'] + filter_obj.get_filter_params()
-        tasks = []
-        for line in self.execute_command(args):
-            if line:
-                data = line.strip(',')
-                try:
-                    filtered_task = Task(self)
-                    filtered_task._load_data(json.loads(data))
-                    tasks.append(filtered_task)
-                except ValueError:
-                    raise TaskWarriorException('Invalid JSON: %s' % data)
-        return tasks
-
-    def merge_with(self, path, push=False):
-        path = path.rstrip('/') + '/'
-        self.execute_command(['merge', path], config_override={
-            'merge.autopush': 'yes' if push else 'no',
-        })
-
-    def undo(self):
-        self.execute_command(['undo'])