DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Mercurial (409f3966645a)

VCS Links

Line Code
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
# file, You can obtain one at http://mozilla.org/MPL/2.0/.

from __future__ import absolute_import, print_function, unicode_literals
import logging
import os
import yaml
import copy

from . import filter_tasks
from .graph import Graph
from .taskgraph import TaskGraph
from .task import Task
from .optimize import optimize_task_graph
from .morph import morph
from .util.python_path import find_object
from .transforms.base import TransformSequence, TransformConfig
from .util.verify import (
    verify_docs,
    verifications,
)
from .config import load_graph_config

logger = logging.getLogger(__name__)


class KindNotFound(Exception):
    """
    Raised when trying to load kind from a directory without a kind.yml.
    """


class Kind(object):

    def __init__(self, name, path, config, graph_config):
        self.name = name
        self.path = path
        self.config = config
        self.graph_config = graph_config

    def _get_loader(self):
        try:
            loader = self.config['loader']
        except KeyError:
            raise KeyError("{!r} does not define `loader`".format(self.path))
        return find_object(loader)

    def load_tasks(self, parameters, loaded_tasks):
        loader = self._get_loader()
        config = copy.deepcopy(self.config)

        kind_dependencies = config.get('kind-dependencies', [])
        kind_dependencies_tasks = [task for task in loaded_tasks
                                   if task.kind in kind_dependencies]

        inputs = loader(self.name, self.path, config, parameters, loaded_tasks)

        transforms = TransformSequence()
        for xform_path in config['transforms']:
            transform = find_object(xform_path)
            transforms.add(transform)

        # perform the transformations on the loaded inputs
        trans_config = TransformConfig(self.name, self.path, config, parameters,
                                       kind_dependencies_tasks, self.graph_config)
        tasks = [Task(self.name,
                      label=task_dict['label'],
                      attributes=task_dict['attributes'],
                      task=task_dict['task'],
                      optimization=task_dict.get('optimization'),
                      dependencies=task_dict.get('dependencies'),
                      release_artifacts=task_dict.get('release-artifacts'),
                      )
                 for task_dict in transforms(trans_config, inputs)]
        return tasks

    @classmethod
    def load(cls, root_dir, graph_config, kind_name):
        path = os.path.join(root_dir, kind_name)
        kind_yml = os.path.join(path, 'kind.yml')
        if not os.path.exists(kind_yml):
            raise KindNotFound(kind_yml)

        logger.debug("loading kind `{}` from `{}`".format(kind_name, path))
        with open(kind_yml) as f:
            config = yaml.safe_load(f)

        return cls(kind_name, path, config, graph_config)


class TaskGraphGenerator(object):
    """
    The central controller for taskgraph.  This handles all phases of graph
    generation.  The task is generated from all of the kinds defined in
    subdirectories of the generator's root directory.

    Access to the results of this generation, as well as intermediate values at
    various phases of generation, is available via properties.  This encourages
    the provision of all generation inputs at instance construction time.
    """

    # Task-graph generation is implemented as a Python generator that yields
    # each "phase" of generation.  This allows some mach subcommands to short-
    # circuit generation of the entire graph by never completing the generator.

    def __init__(self, root_dir, parameters):
        """
        @param root_dir: root directory, with subdirectories for each kind
        @param parameters: parameters for this task-graph generation
        @type parameters: dict
        """
        if root_dir is None:
            root_dir = 'taskcluster/ci'
        self.root_dir = root_dir
        self.parameters = parameters

        self.verify_parameters(self.parameters)

        filters = parameters.get('filters', [])

        # Always add legacy target tasks method until we deprecate that API.
        if 'target_tasks_method' not in filters:
            filters.insert(0, 'target_tasks_method')

        self.filters = [filter_tasks.filter_task_functions[f] for f in filters]

        # this can be set up until the time the target task set is generated;
        # it defaults to parameters['target_tasks']
        self._target_tasks = parameters.get('target_tasks')

        # start the generator
        self._run = self._run()
        self._run_results = {}

    @property
    def full_task_set(self):
        """
        The full task set: all tasks defined by any kind (a graph without edges)

        @type: TaskGraph
        """
        return self._run_until('full_task_set')

    @property
    def full_task_graph(self):
        """
        The full task graph: the full task set, with edges representing
        dependencies.

        @type: TaskGraph
        """
        return self._run_until('full_task_graph')

    @property
    def target_task_set(self):
        """
        The set of targetted tasks (a graph without edges)

        @type: TaskGraph
        """
        return self._run_until('target_task_set')

    @property
    def target_task_graph(self):
        """
        The set of targetted tasks and all of their dependencies

        @type: TaskGraph
        """
        return self._run_until('target_task_graph')

    @property
    def optimized_task_graph(self):
        """
        The set of targetted tasks and all of their dependencies; tasks that
        have been optimized out are either omitted or replaced with a Task
        instance containing only a task_id.

        @type: TaskGraph
        """
        return self._run_until('optimized_task_graph')

    @property
    def label_to_taskid(self):
        """
        A dictionary mapping task label to assigned taskId.  This property helps
        in interpreting `optimized_task_graph`.

        @type: dictionary
        """
        return self._run_until('label_to_taskid')

    @property
    def morphed_task_graph(self):
        """
        The optimized task graph, with any subsequent morphs applied. This graph
        will have the same meaning as the optimized task graph, but be in a form
        more palatable to TaskCluster.

        @type: TaskGraph
        """
        return self._run_until('morphed_task_graph')

    @property
    def graph_config(self):
        """
        The configuration for this graph.

        @type: TaskGraph
        """
        return self._run_until('graph_config')

    def _load_kinds(self, graph_config):
        for kind_name in os.listdir(self.root_dir):
            try:
                yield Kind.load(self.root_dir, graph_config, kind_name)
            except KindNotFound:
                continue

    def _run(self):
        logger.info("Loading graph configuration.")
        graph_config = load_graph_config(self.root_dir)

        yield verifications('graph_config', graph_config)

        logger.info("Loading kinds")
        # put the kinds into a graph and sort topologically so that kinds are loaded
        # in post-order
        kinds = {kind.name: kind for kind in self._load_kinds(graph_config)}
        self.verify_kinds(kinds)

        edges = set()
        for kind in kinds.itervalues():
            for dep in kind.config.get('kind-dependencies', []):
                edges.add((kind.name, dep, 'kind-dependency'))
        kind_graph = Graph(set(kinds), edges)

        logger.info("Generating full task set")
        all_tasks = {}
        for kind_name in kind_graph.visit_postorder():
            logger.debug("Loading tasks for kind {}".format(kind_name))
            kind = kinds[kind_name]
            new_tasks = kind.load_tasks(self.parameters, list(all_tasks.values()))
            for task in new_tasks:
                if task.label in all_tasks:
                    raise Exception("duplicate tasks with label " + task.label)
                all_tasks[task.label] = task
            logger.info("Generated {} tasks for kind {}".format(len(new_tasks), kind_name))
        full_task_set = TaskGraph(all_tasks, Graph(set(all_tasks), set()))
        self.verify_attributes(all_tasks)
        self.verify_run_using()
        yield verifications('full_task_set', full_task_set)

        logger.info("Generating full task graph")
        edges = set()
        for t in full_task_set:
            for depname, dep in t.dependencies.iteritems():
                edges.add((t.label, dep, depname))

        full_task_graph = TaskGraph(all_tasks,
                                    Graph(full_task_set.graph.nodes, edges))
        logger.info("Full task graph contains %d tasks and %d dependencies" % (
            len(full_task_set.graph.nodes), len(edges)))
        yield verifications('full_task_graph', full_task_graph)

        logger.info("Generating target task set")
        target_task_set = TaskGraph(dict(all_tasks),
                                    Graph(set(all_tasks.keys()), set()))
        for fltr in self.filters:
            old_len = len(target_task_set.graph.nodes)
            target_tasks = set(fltr(target_task_set, self.parameters, graph_config))
            target_task_set = TaskGraph(
                {l: all_tasks[l] for l in target_tasks},
                Graph(target_tasks, set()))
            logger.info('Filter %s pruned %d tasks (%d remain)' % (
                fltr.__name__,
                old_len - len(target_tasks),
                len(target_tasks)))

        yield verifications('target_task_set', target_task_set)

        logger.info("Generating target task graph")
        # include all docker-image build tasks here, in case they are needed for a graph morph
        docker_image_tasks = set(t.label for t in full_task_graph.tasks.itervalues()
                                 if t.attributes['kind'] == 'docker-image')
        # include all tasks with `always_target` set
        always_target_tasks = set(t.label for t in full_task_graph.tasks.itervalues()
                                  if t.attributes.get('always_target'))
        logger.info('Adding %d tasks with `always_target` attribute' % (
                    len(always_target_tasks) - len(always_target_tasks & target_tasks)))
        target_graph = full_task_graph.graph.transitive_closure(
            target_tasks | docker_image_tasks | always_target_tasks)
        target_task_graph = TaskGraph(
            {l: all_tasks[l] for l in target_graph.nodes},
            target_graph)
        yield verifications('target_task_graph', target_task_graph)

        logger.info("Generating optimized task graph")
        existing_tasks = self.parameters.get('existing_tasks')
        do_not_optimize = set(self.parameters.get('do_not_optimize', []))
        if not self.parameters.get('optimize_target_tasks', True):
            do_not_optimize = set(target_task_set.graph.nodes).union(do_not_optimize)
        optimized_task_graph, label_to_taskid = optimize_task_graph(target_task_graph,
                                                                    self.parameters,
                                                                    do_not_optimize,
                                                                    existing_tasks=existing_tasks)

        yield verifications('optimized_task_graph', optimized_task_graph)

        morphed_task_graph, label_to_taskid = morph(
            optimized_task_graph, label_to_taskid, self.parameters)

        yield 'label_to_taskid', label_to_taskid
        yield verifications('morphed_task_graph', morphed_task_graph)

    def _run_until(self, name):
        while name not in self._run_results:
            try:
                k, v = self._run.next()
            except StopIteration:
                raise AttributeError("No such run result {}".format(name))
            self._run_results[k] = v
        return self._run_results[name]

    def verify_parameters(self, parameters):
        if not parameters.strict:
            return

        parameters_dict = dict(**parameters)
        verify_docs(
            filename="parameters.rst",
            identifiers=parameters_dict.keys(),
            appearing_as="inline-literal"
         )

    def verify_kinds(self, kinds):
        verify_docs(
            filename="kinds.rst",
            identifiers=kinds.keys(),
            appearing_as="heading"
         )

    def verify_attributes(self, all_tasks):
        attribute_set = set()
        for label, task in all_tasks.iteritems():
            attribute_set.update(task.attributes.keys())
        verify_docs(
            filename="attributes.rst",
            identifiers=list(attribute_set),
            appearing_as="heading"
         )

    def verify_run_using(self):
        from .transforms.job import registry
        verify_docs(
            filename="transforms.rst",
            identifiers=registry.keys(),
            appearing_as="inline-literal"
         )