X-Git-Url: https://git.madduck.net/etc/taskwarrior.git/blobdiff_plain/ade25ab6ef26a73da25b3be09658ea39875dd6f3..516b8d6c03058552cfe7c3bfda4362640a3077e1:/tasklib/task.py diff --git a/tasklib/task.py b/tasklib/task.py index 5a223e2..c7c9e6d 100644 --- a/tasklib/task.py +++ b/tasklib/task.py @@ -1,111 +1,439 @@ +from __future__ import print_function import copy -import datetime +import importlib import json +import logging import os -import subprocess -import tempfile -import uuid +import six +import sys +from .serializing import SerializingObject DATE_FORMAT = '%Y%m%dT%H%M%SZ' - REPR_OUTPUT_SIZE = 10 - PENDING = 'pending' +COMPLETED = 'completed' +DELETED = 'deleted' +WAITING = 'waiting' +RECURRING = 'recurring' +logger = logging.getLogger(__name__) -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 -class Task(object): + def __getitem__(self, key): + return copy.deepcopy(self.viewed_dict.__getitem__(key)) - class DoesNotExist(Exception): - pass + 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 __unicode__(self): + return six.u('ReadOnlyDictView: {0}'.format(repr(self.viewed_dict))) + + __repr__ = __unicode__ + + 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()] - def __init__(self, warrior, data={}): - self.warrior = warrior - self._data = data + +class TaskResource(SerializingObject): + read_only_fields = [] + + def _load_data(self, data): + self._data = dict((key, self._deserialize(key, value)) + for key, value in data.items()) + # We need to use a copy for original data, so that changes + # are not propagated. + self._original_data = copy.deepcopy(self._data) + + 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 + original_data dict is updated as well. + """ + 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) def __getitem__(self, key): - return self._get_field(key) + # This is a workaround to make TaskResource non-iterable + # over simple index-based iteration + try: + int(key) + raise StopIteration + except ValueError: + pass + + if key not in self._data: + self._data[key] = self._deserialize(key, None) + + return self._data.get(key) + + def __setitem__(self, key, value): + if key in self.read_only_fields: + raise RuntimeError('Field \'%s\' is read-only' % key) - def __setitem__(self, key, val): - self._data[key] = val + # Normalize the user input before saving it + value = self._normalize(key, value) + self._data[key] = value + + def __str__(self): + s = six.text_type(self.__unicode__()) + if not six.PY3: + s = s.encode('utf-8') + return s + + def __repr__(self): + return str(self) + + def export_data(self): + """ + Exports current data contained in the Task as JSON + """ + + # We need to remove spaces for TW-1504, use custom separators + data_tuples = ((key, self._serialize(key, value)) + for key, value in six.iteritems(self._data)) + + # Empty string denotes empty serialized value, we do not want + # to pass that to TaskWarrior. + data_tuples = filter(lambda t: t[1] is not '', data_tuples) + data = dict(data_tuples) + return json.dumps(data, separators=(',', ':')) + + @property + def _modified_fields(self): + writable_fields = set(self._data.keys()) - set(self.read_only_fields) + for key in writable_fields: + new_value = self._data.get(key) + old_value = self._original_data.get(key) + + # Make sure not to mark data removal as modified field if the + # field originally had some empty value + if key in self._data and not new_value and not old_value: + continue + + if new_value != old_value: + yield key + + @property + def modified(self): + return bool(list(self._modified_fields)) + + +class TaskAnnotation(TaskResource): + read_only_fields = ['entry', 'description'] + + def __init__(self, task, data=None): + self.task = task + self._load_data(data or dict()) + super(TaskAnnotation, self).__init__(task.backend) + + def remove(self): + self.task.remove_annotation(self) def __unicode__(self): - return self._data.get('description') + return self['description'] + + def __eq__(self, other): + # consider 2 annotations equal if they belong to the same task, and + # their data dics are the same + return self.task == other.task and self._data == other._data - def _get_field(self, key): - hydrate_func = getattr(self, 'deserialize_{0}'.format(key), lambda x:x) - return hydrate_func(self._data.get(key)) + def __ne__(self, other): + return not self.__eq__(other) - def _set_field(self, key, value): - dehydrate_func = getattr(self, 'serialize_{0}'.format(key), lambda x:x) - self._data[key] = dehydrate_func(value) + __repr__ = __unicode__ - def serialize_due(self, date): - return date.strftime(DATE_FORMAT) - def deserialize_due(self, date_str): - if not date_str: - return None - return datetime.datetime.strptime(date_str, DATE_FORMAT) +class Task(TaskResource): + read_only_fields = ['id', 'entry', 'urgency', 'uuid', 'modified'] - def serialize_annotations(self, annotations): - ann_list = list(annotations) - for ann in ann_list: - ann['entry'] = ann['entry'].strftime(DATE_FORMAT) - return ann_list + class DoesNotExist(Exception): + pass - def deserialize_annotations(self, annotations): - ann_list = list(annotations) - for ann in ann_list: - ann['entry'] = datetime.datetime.strptime( - ann['entry'], DATE_FORMAT) - return ann_list + class CompletedTask(Exception): + """ + Raised when the operation cannot be performed on the completed task. + """ + pass - def regenerate_uuid(self): - self['uuid'] = str(uuid.uuid4()) + class DeletedTask(Exception): + """ + Raised when the operation cannot be performed on the deleted task. + """ + 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 + it has not been saved to TaskWarrior yet. + """ + pass + + @classmethod + def from_input(cls, input_file=sys.stdin, modify=None, backend=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 + original state of the Task object, and second line as its new, + modified value. This is consistent with the TaskWarrior's hook + system. + + Object created by this method should not be saved, deleted + or refreshed, as t could create a infinite loop. For this + reason, TaskWarrior instance is set to None. + + Input_file argument can be used to specify the input file, + but defaults to sys.stdin. + """ + + # 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 backend is None: + backends = importlib.import_module('tasklib.backends') + hook_parent_dir = os.path.dirname(os.path.dirname(sys.argv[0])) + backend = backends.TaskWarrior(data_location=hook_parent_dir) + + # TaskWarrior instance is set to None + task = cls(backend) + + # 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()), + remove_missing=True) + + return task + + def __init__(self, backend, **kwargs): + super(Task, self).__init__(backend) + + # Check that user is not able to set read-only value in __init__ + for key in kwargs.keys(): + if key in self.read_only_fields: + raise RuntimeError('Field \'%s\' is read-only' % key) + + # We serialize the data in kwargs so that users of the library + # do not have to pass different data formats via __setitem__ and + # __init__ methods, that would be confusing + + # Rather unfortunate syntax due to python2.6 comaptiblity + self._data = dict((key, self._normalize(key, value)) + 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 __eq__(self, other): + if self['uuid'] and other['uuid']: + # For saved Tasks, just define equality by equality of uuids + return self['uuid'] == other['uuid'] + else: + # If the tasks are not saved, compare the actual instances + return id(self) == id(other) + + def __ne__(self, other): + return not self.__eq__(other) + + def __hash__(self): + if self['uuid']: + # For saved Tasks, just define equality by equality of uuids + return self['uuid'].__hash__() + else: + # If the tasks are not saved, return hash of instance id + return id(self).__hash__() + + @property + def completed(self): + return self['status'] == six.text_type('completed') + + @property + def deleted(self): + return self['status'] == six.text_type('deleted') + + @property + def waiting(self): + return self['status'] == six.text_type('waiting') + + @property + def pending(self): + return self['status'] == six.text_type('pending') + + @property + def recurring(self): + return self['status'] == six.text_type('recurring') + + @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 + + def serialize_depends(self, cur_dependencies): + # Check that all the tasks are saved + for task in (cur_dependencies or set()): + if not task.saved: + raise Task.NotSaved('Task \'%s\' needs to be saved before ' + 'it can be set as dependency.' % task) + + return super(Task, self).serialize_depends(cur_dependencies) def delete(self): - self.warrior.delete_task(self['uuid']) + if not self.saved: + raise Task.NotSaved("Task needs to be saved before it can be deleted") + + # Refresh the status, and raise exception if the task is deleted + self.refresh(only_fields=['status']) + + if self.deleted: + raise Task.DeletedTask("Task was already deleted") + + self.backend.delete_task(self) + + # 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']) + + 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): - self.warrior.complete_task(self['uuid']) + if not self.saved: + raise Task.NotSaved("Task needs to be saved before it can be completed") - def save(self, delete_first=True): - if self['uuid'] and delete_first: - self.delete() - if not self['uuid'] or delete_first: - self.regenerate_uuid() - self.warrior.import_tasks([self._data]) + # Refresh, and raise exception if task is already completed/deleted + self.refresh(only_fields=['status']) - __repr__ = __unicode__ + if self.completed: + raise Task.CompletedTask("Cannot complete a completed task") + elif self.deleted: + raise Task.DeletedTask("Deleted task cannot be completed") + self.backend.complete_task(self) -class TaskFilter(object): - """ - A set of parameters to filter the task list with. - """ + # Refresh the status again, so that we have updated info stored + self.refresh(only_fields=['status', 'start', 'end']) - def __init__(self, filter_params=[]): - self.filter_params = filter_params + def save(self): + if self.saved and not self.modified: + return - def add_filter(self, filter_str): - self.filter_params.append(filter_str) + # All the actual work is done by the backend + self.backend.save_task(self) - def add_filter_param(self, key, value): - key = key.replace('__', '.') - self.filter_params.append('{0}:{1}'.format(key, value)) + def add_annotation(self, annotation): + if not self.saved: + raise Task.NotSaved("Task needs to be saved to add annotation") - def get_filter_params(self): - return [f for f in self.filter_params if f] + self.backend.annotate_task(self, annotation) + self.refresh(only_fields=['annotations']) - def clone(self): - c = self.__class__() - c.filter_params = list(self.filter_params) - return c + def remove_annotation(self, annotation): + if not self.saved: + raise Task.NotSaved("Task needs to be saved to remove annotation") + + if isinstance(annotation, TaskAnnotation): + annotation = annotation['description'] + + self.backend.denotate_task(self, annotation) + self.refresh(only_fields=['annotations']) + + 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") + + new_data = self.backend.refresh_task(self, after_save=after_save) + + if only_fields: + to_update = dict( + [(k, new_data.get(k)) for k in only_fields]) + self._update_data(to_update, update_original=True) + else: + self._load_data(new_data) class TaskQuerySet(object): @@ -113,18 +441,18 @@ class TaskQuerySet(object): Represents a lazy lookup for a task objects. """ - def __init__(self, warrior=None, filter_obj=None): - self.warrior = warrior + def __init__(self, backend, filter_obj=None): + self.backend = backend self._result_cache = None - self.filter_obj = filter_obj or TaskFilter() + self.filter_obj = filter_obj or self.backend.filter_class(backend) def __deepcopy__(self, memo): """ Deep copy of a QuerySet doesn't populate the cache """ - obj = self.__class__() - for k,v in self.__dict__.items(): - if k in ('_iter','_result_cache'): + obj = self.__class__(backend=self.backend) + for k, v in self.__dict__.items(): + if k in ('_iter', '_result_cache'): obj.__dict__[k] = None else: obj.__dict__[k] = copy.deepcopy(v, memo) @@ -167,7 +495,7 @@ class TaskQuerySet(object): if klass is None: klass = self.__class__ filter_obj = self.filter_obj.clone() - c = klass(warrior=self.warrior, filter_obj=filter_obj) + c = klass(backend=self.backend, filter_obj=filter_obj) c.__dict__.update(kwargs) return c @@ -175,7 +503,7 @@ class TaskQuerySet(object): """ Fetch the tasks which match the current filters. """ - return self.warrior.filter_tasks(self.filter_obj) + return self.backend.filter_tasks(self.filter_obj) def all(self): """ @@ -186,6 +514,18 @@ class TaskQuerySet(object): def pending(self): return self.filter(status=PENDING) + def completed(self): + return self.filter(status=COMPLETED) + + def deleted(self): + return self.filter(status=DELETED) + + def waiting(self): + return self.filter(status=WAITING) + + def recurring(self): + return self.filter(status=RECURRING) + def filter(self, *args, **kwargs): """ Returns a new TaskQuerySet with the given filters added. @@ -213,69 +553,3 @@ class TaskQuerySet(object): 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): - if not os.path.exists(data_location): - os.makedirs(data_location) - self.config = { - 'data.location': os.path.expanduser(data_location), - } - self.tasks = TaskQuerySet(self) - - 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(args) - return command_args - - def execute_command(self, args, config_override={}): - command_args = self._get_command_args( - args, config_override=config_override) - p = subprocess.Popen(command_args, stdout=subprocess.PIPE, - stderr=subprocess.PIPE) - stdout, stderr = p.communicate() - if p.returncode: - error_msg = stderr.strip().splitlines()[-1] - raise TaskWarriorException(error_msg) - return stdout.strip().split('\n') - - def filter_tasks(self, filter_obj): - args = ['export', '--'] + filter_obj.get_filter_params() - tasks = [] - for line in self.execute_command(args): - if line: - tasks.append(Task(self, json.loads(line.strip(',')))) - return tasks - - def add_task(self, description, project=None): - args = ['add', description] - if project is not None: - args.append('project:{0}'.format(project)) - self.execute_command(args) - - def delete_task(self, task_id): - args = [task_id, 'rc.confirmation:no', 'delete'] - self.execute_command(args) - - def complete_task(self, task_id): - args = [task_id, 'done'] - self.execute_command(args) - - def import_tasks(self, tasks): - fd, path = tempfile.mkstemp() - with open(path, 'w') as f: - f.write(json.dumps(tasks)) - args = ['import', path] - self.execute_command(args) - - def merge_with(self, path, push=False): - path = path.rstrip('/') + '/' - args = ['merge', path] - self.execute_command(args, config_override={ - 'merge.autopush': 'yes' if push else 'no', - })