mediagoblin/db/models.py

   1 # GNU MediaGoblin -- federated, autonomous media hosting
   2 # Copyright (C) 2011, 2012 MediaGoblin contributors.  See AUTHORS.
   3 #
   4 # This program is free software: you can redistribute it and/or modify
   5 # it under the terms of the GNU Affero General Public License as published by
   6 # the Free Software Foundation, either version 3 of the License, or
   7 # (at your option) any later version.
   8 #
   9 # This program is distributed in the hope that it will be useful,
  10 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 # GNU Affero General Public License for more details.
  13 #
  14 # You should have received a copy of the GNU Affero General Public License
  15 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
  16
  17 """
  18 TODO: indexes on foreignkeys, where useful.
  19 """
  20
  21 import logging
  22 import datetime
  23 import sys
  24
  25 from sqlalchemy import Column, Integer, Unicode, UnicodeText, DateTime, \
  26         Boolean, ForeignKey, UniqueConstraint, PrimaryKeyConstraint, \
  27         SmallInteger
  28 from sqlalchemy.orm import relationship, backref
  29 from sqlalchemy.orm.collections import attribute_mapped_collection
  30 from sqlalchemy.sql.expression import desc
  31 from sqlalchemy.ext.associationproxy import association_proxy
  32 from sqlalchemy.util import memoized_property
  33
  34 from mediagoblin.db.extratypes import PathTupleWithSlashes, JSONEncoded
  35 from mediagoblin.db.base import Base, DictReadAttrProxy, Session
  36 from mediagoblin.db.mixin import UserMixin, MediaEntryMixin, MediaCommentMixin, CollectionMixin, CollectionItemMixin
  37 from mediagoblin.tools.files import delete_media_files
  38
  39 # It's actually kind of annoying how sqlalchemy-migrate does this, if
  40 # I understand it right, but whatever.  Anyway, don't remove this :P
  41 #
  42 # We could do migration calls more manually instead of relying on
  43 # this import-based meddling...
  44 from migrate import changeset
  45
  46 _log = logging.getLogger(__name__)
  47
  48
  49 class User(Base, UserMixin):
  50     """
  51     TODO: We should consider moving some rarely used fields
  52     into some sort of "shadow" table.
  53     """
  54     __tablename__ = "core__users"
  55
  56     id = Column(Integer, primary_key=True)
  57     username = Column(Unicode, nullable=False, unique=True)
  58     email = Column(Unicode, nullable=False)
  59     created = Column(DateTime, nullable=False, default=datetime.datetime.now)
  60     pw_hash = Column(Unicode, nullable=False)
  61     email_verified = Column(Boolean, default=False)
  62     status = Column(Unicode, default=u"needs_email_verification", nullable=False)
  63     # Intented to be nullable=False, but migrations would not work for it
  64     # set to nullable=True implicitly.
  65     wants_comment_notification = Column(Boolean, default=True)
  66     license_preference = Column(Unicode)
  67     verification_key = Column(Unicode)
  68     is_admin = Column(Boolean, default=False, nullable=False)
  69     url = Column(Unicode)
  70     bio = Column(UnicodeText)  # ??
  71     fp_verification_key = Column(Unicode)
  72     fp_token_expire = Column(DateTime)
  73
  74     ## TODO
  75     # plugin data would be in a separate model
  76
  77     def __repr__(self):
  78         return '<{0} #{1} {2} {3} "{4}">'.format(
  79                 self.__class__.__name__,
  80                 self.id,
  81                 'verified' if self.email_verified else 'non-verified',
  82                 'admin' if self.is_admin else 'user',
  83                 self.username)
  84
  85     def delete(self, **kwargs):
  86         """Deletes a User and all related entries/comments/files/..."""
  87         # Delete this user's Collections and all contained CollectionItems
  88         for collection in self.collections:
  89             collection.delete(commit=False)
  90
  91         media_entries = MediaEntry.query.filter(MediaEntry.uploader == self.id)
  92         for media in media_entries:
  93             # TODO: Make sure that "MediaEntry.delete()" also deletes
  94             # all related files/Comments
  95             media.delete(del_orphan_tags=False, commit=False)
  96
  97         # Delete now unused tags
  98         # TODO: import here due to cyclic imports!!! This cries for refactoring
  99         from mediagoblin.db.util import clean_orphan_tags
 100         clean_orphan_tags(commit=False)
 101
 102         # Delete user, pass through commit=False/True in kwargs
 103         super(User, self).delete(**kwargs)
 104         _log.info('Deleted user "{0}" account'.format(self.username))
 105
 106
 107 class MediaEntry(Base, MediaEntryMixin):
 108     """
 109     TODO: Consider fetching the media_files using join
 110     """
 111     __tablename__ = "core__media_entries"
 112
 113     id = Column(Integer, primary_key=True)
 114     uploader = Column(Integer, ForeignKey(User.id), nullable=False, index=True)
 115     title = Column(Unicode, nullable=False)
 116     slug = Column(Unicode)
 117     created = Column(DateTime, nullable=False, default=datetime.datetime.now,
 118         index=True)
 119     description = Column(UnicodeText) # ??
 120     media_type = Column(Unicode, nullable=False)
 121     state = Column(Unicode, default=u'unprocessed', nullable=False)
 122         # or use sqlalchemy.types.Enum?
 123     license = Column(Unicode)
 124     collected = Column(Integer, default=0)
 125
 126     fail_error = Column(Unicode)
 127     fail_metadata = Column(JSONEncoded)
 128
 129     transcoding_progress = Column(SmallInteger)
 130
 131     queued_media_file = Column(PathTupleWithSlashes)
 132
 133     queued_task_id = Column(Unicode)
 134
 135     __table_args__ = (
 136         UniqueConstraint('uploader', 'slug'),
 137         {})
 138
 139     get_uploader = relationship(User)
 140
 141     media_files_helper = relationship("MediaFile",
 142         collection_class=attribute_mapped_collection("name"),
 143         cascade="all, delete-orphan"
 144         )
 145     media_files = association_proxy('media_files_helper', 'file_path',
 146         creator=lambda k, v: MediaFile(name=k, file_path=v)
 147         )
 148
 149     attachment_files_helper = relationship("MediaAttachmentFile",
 150         order_by="MediaAttachmentFile.created"
 151         )
 152     attachment_files = association_proxy("attachment_files_helper", "dict_view",
 153         creator=lambda v: MediaAttachmentFile(
 154             name=v["name"], filepath=v["filepath"])
 155         )
 156
 157     tags_helper = relationship("MediaTag",
 158         cascade="all, delete-orphan" # should be automatically deleted
 159         )
 160     tags = association_proxy("tags_helper", "dict_view",
 161         creator=lambda v: MediaTag(name=v["name"], slug=v["slug"])
 162         )
 163
 164     collections_helper = relationship("CollectionItem",
 165         cascade="all, delete-orphan"
 166         )
 167     collections = association_proxy("collections_helper", "in_collection")
 168
 169     ## TODO
 170     # media_data
 171     # fail_error
 172
 173     def get_comments(self, ascending=False):
 174         order_col = MediaComment.created
 175         if not ascending:
 176             order_col = desc(order_col)
 177         return MediaComment.query.filter_by(
 178             media_entry=self.id).order_by(order_col)
 179
 180     def url_to_prev(self, urlgen):
 181         """get the next 'newer' entry by this user"""
 182         media = MediaEntry.query.filter(
 183             (MediaEntry.uploader == self.uploader)
 184             & (MediaEntry.state == u'processed')
 185             & (MediaEntry.id > self.id)).order_by(MediaEntry.id).first()
 186
 187         if media is not None:
 188             return media.url_for_self(urlgen)
 189
 190     def url_to_next(self, urlgen):
 191         """get the next 'older' entry by this user"""
 192         media = MediaEntry.query.filter(
 193             (MediaEntry.uploader == self.uploader)
 194             & (MediaEntry.state == u'processed')
 195             & (MediaEntry.id < self.id)).order_by(desc(MediaEntry.id)).first()
 196
 197         if media is not None:
 198             return media.url_for_self(urlgen)
 199
 200     #@memoized_property
 201     @property
 202     def media_data(self):
 203         session = Session()
 204
 205         return session.query(self.media_data_table).filter_by(
 206             media_entry=self.id).first()
 207
 208     def media_data_init(self, **kwargs):
 209         """
 210         Initialize or update the contents of a media entry's media_data row
 211         """
 212         session = Session()
 213
 214         media_data = session.query(self.media_data_table).filter_by(
 215             media_entry=self.id).first()
 216
 217         # No media data, so actually add a new one
 218         if media_data is None:
 219             media_data = self.media_data_table(
 220                 media_entry=self.id,
 221                 **kwargs)
 222             session.add(media_data)
 223         # Update old media data
 224         else:
 225             for field, value in kwargs.iteritems():
 226                 setattr(media_data, field, value)
 227
 228     @memoized_property
 229     def media_data_table(self):
 230         # TODO: memoize this
 231         models_module = self.media_type + '.models'
 232         __import__(models_module)
 233         return sys.modules[models_module].DATA_MODEL
 234
 235     def __repr__(self):
 236         safe_title = self.title.encode('ascii', 'replace')
 237
 238         return '<{classname} {id}: {title}>'.format(
 239                 classname=self.__class__.__name__,
 240                 id=self.id,
 241                 title=safe_title)
 242
 243     def delete(self, del_orphan_tags=True, **kwargs):
 244         """Delete MediaEntry and all related files/attachments/comments
 245
 246         This will *not* automatically delete unused collections, which
 247         can remain empty...
 248
 249         :param del_orphan_tags: True/false if we delete unused Tags too
 250         :param commit: True/False if this should end the db transaction"""
 251         # User's CollectionItems are automatically deleted via "cascade".
 252         # Delete all the associated comments
 253         for comment in self.get_comments():
 254             comment.delete(commit=False)
 255
 256         # Delete all related files/attachments
 257         try:
 258             delete_media_files(self)
 259         except OSError, error:
 260             # Returns list of files we failed to delete
 261             _log.error('No such files from the user "{1}" to delete: '
 262                        '{0}'.format(str(error), self.get_uploader))
 263         _log.info('Deleted Media entry id "{0}"'.format(self.id))
 264         # Related MediaTag's are automatically cleaned, but we might
 265         # want to clean out unused Tag's too.
 266         if del_orphan_tags:
 267             # TODO: Import here due to cyclic imports!!!
 268             #       This cries for refactoring
 269             from mediagoblin.db.util import clean_orphan_tags
 270             clean_orphan_tags(commit=False)
 271         # pass through commit=False/True in kwargs
 272         super(MediaEntry, self).delete(**kwargs)
 273
 274
 275 class FileKeynames(Base):
 276     """
 277     keywords for various places.
 278     currently the MediaFile keys
 279     """
 280     __tablename__ = "core__file_keynames"
 281     id = Column(Integer, primary_key=True)
 282     name = Column(Unicode, unique=True)
 283
 284     def __repr__(self):
 285         return "<FileKeyname %r: %r>" % (self.id, self.name)
 286
 287     @classmethod
 288     def find_or_new(cls, name):
 289         t = cls.query.filter_by(name=name).first()
 290         if t is not None:
 291             return t
 292         return cls(name=name)
 293
 294
 295 class MediaFile(Base):
 296     """
 297     TODO: Highly consider moving "name" into a new table.
 298     TODO: Consider preloading said table in software
 299     """
 300     __tablename__ = "core__mediafiles"
 301
 302     media_entry = Column(
 303         Integer, ForeignKey(MediaEntry.id),
 304         nullable=False)
 305     name_id = Column(SmallInteger, ForeignKey(FileKeynames.id), nullable=False)
 306     file_path = Column(PathTupleWithSlashes)
 307
 308     __table_args__ = (
 309         PrimaryKeyConstraint('media_entry', 'name_id'),
 310         {})
 311
 312     def __repr__(self):
 313         return "<MediaFile %s: %r>" % (self.name, self.file_path)
 314
 315     name_helper = relationship(FileKeynames, lazy="joined", innerjoin=True)
 316     name = association_proxy('name_helper', 'name',
 317         creator=FileKeynames.find_or_new
 318         )
 319
 320
 321 class MediaAttachmentFile(Base):
 322     __tablename__ = "core__attachment_files"
 323
 324     id = Column(Integer, primary_key=True)
 325     media_entry = Column(
 326         Integer, ForeignKey(MediaEntry.id),
 327         nullable=False)
 328     name = Column(Unicode, nullable=False)
 329     filepath = Column(PathTupleWithSlashes)
 330     created = Column(DateTime, nullable=False, default=datetime.datetime.now)
 331
 332     @property
 333     def dict_view(self):
 334         """A dict like view on this object"""
 335         return DictReadAttrProxy(self)
 336
 337
 338 class Tag(Base):
 339     __tablename__ = "core__tags"
 340
 341     id = Column(Integer, primary_key=True)
 342     slug = Column(Unicode, nullable=False, unique=True)
 343
 344     def __repr__(self):
 345         return "<Tag %r: %r>" % (self.id, self.slug)
 346
 347     @classmethod
 348     def find_or_new(cls, slug):
 349         t = cls.query.filter_by(slug=slug).first()
 350         if t is not None:
 351             return t
 352         return cls(slug=slug)
 353
 354
 355 class MediaTag(Base):
 356     __tablename__ = "core__media_tags"
 357
 358     id = Column(Integer, primary_key=True)
 359     media_entry = Column(
 360         Integer, ForeignKey(MediaEntry.id),
 361         nullable=False, index=True)
 362     tag = Column(Integer, ForeignKey(Tag.id), nullable=False, index=True)
 363     name = Column(Unicode)
 364     # created = Column(DateTime, nullable=False, default=datetime.datetime.now)
 365
 366     __table_args__ = (
 367         UniqueConstraint('tag', 'media_entry'),
 368         {})
 369
 370     tag_helper = relationship(Tag)
 371     slug = association_proxy('tag_helper', 'slug',
 372         creator=Tag.find_or_new
 373         )
 374
 375     def __init__(self, name=None, slug=None):
 376         Base.__init__(self)
 377         if name is not None:
 378             self.name = name
 379         if slug is not None:
 380             self.tag_helper = Tag.find_or_new(slug)
 381
 382     @property
 383     def dict_view(self):
 384         """A dict like view on this object"""
 385         return DictReadAttrProxy(self)
 386
 387
 388 class MediaComment(Base, MediaCommentMixin):
 389     __tablename__ = "core__media_comments"
 390
 391     id = Column(Integer, primary_key=True)
 392     media_entry = Column(
 393         Integer, ForeignKey(MediaEntry.id), nullable=False, index=True)
 394     author = Column(Integer, ForeignKey(User.id), nullable=False)
 395     created = Column(DateTime, nullable=False, default=datetime.datetime.now)
 396     content = Column(UnicodeText, nullable=False)
 397
 398     get_author = relationship(User)
 399
 400
 401 class Collection(Base, CollectionMixin):
 402     """An 'album' or 'set' of media by a user.
 403
 404     On deletion, contained CollectionItems get automatically reaped via
 405     SQL cascade"""
 406     __tablename__ = "core__collections"
 407
 408     id = Column(Integer, primary_key=True)
 409     title = Column(Unicode, nullable=False)
 410     slug = Column(Unicode)
 411     created = Column(DateTime, nullable=False, default=datetime.datetime.now,
 412         index=True)
 413     description = Column(UnicodeText)
 414     creator = Column(Integer, ForeignKey(User.id), nullable=False)
 415     # TODO: No of items in Collection. Badly named, can we migrate to num_items?
 416     items = Column(Integer, default=0)
 417
 418     get_creator = relationship(User, backref="collections")
 419
 420     def get_collection_items(self, ascending=False):
 421         #TODO, is this still needed with self.collection_items being available?
 422         order_col = CollectionItem.position
 423         if not ascending:
 424             order_col = desc(order_col)
 425         return CollectionItem.query.filter_by(
 426             collection=self.id).order_by(order_col)
 427
 428
 429 class CollectionItem(Base, CollectionItemMixin):
 430     __tablename__ = "core__collection_items"
 431
 432     id = Column(Integer, primary_key=True)
 433     media_entry = Column(
 434         Integer, ForeignKey(MediaEntry.id), nullable=False, index=True)
 435     collection = Column(Integer, ForeignKey(Collection.id), nullable=False)
 436     note = Column(UnicodeText, nullable=True)
 437     added = Column(DateTime, nullable=False, default=datetime.datetime.now)
 438     position = Column(Integer)
 439     in_collection = relationship("Collection",
 440                                  backref=backref(
 441                                      "collection_items",
 442                                      cascade="all, delete-orphan"))
 443
 444     get_media_entry = relationship(MediaEntry)
 445
 446     __table_args__ = (
 447         UniqueConstraint('collection', 'media_entry'),
 448         {})
 449
 450     @property
 451     def dict_view(self):
 452         """A dict like view on this object"""
 453         return DictReadAttrProxy(self)
 454
 455
 456 class ProcessingMetaData(Base):
 457     __tablename__ = 'core__processing_metadata'
 458
 459     id = Column(Integer, primary_key=True)
 460     media_entry_id = Column(Integer, ForeignKey(MediaEntry.id), nullable=False,
 461             index=True)
 462     media_entry = relationship(MediaEntry,
 463             backref=backref('processing_metadata',
 464                 cascade='all, delete-orphan'))
 465     callback_url = Column(Unicode)
 466
 467     @property
 468     def dict_view(self):
 469         """A dict like view on this object"""
 470         return DictReadAttrProxy(self)
 471
 472
 473 MODELS = [
 474     User, MediaEntry, Tag, MediaTag, MediaComment, Collection, CollectionItem, MediaFile, FileKeynames,
 475     MediaAttachmentFile, ProcessingMetaData]
 476
 477
 478 ######################################################
 479 # Special, migrations-tracking table
 480 #
 481 # Not listed in MODELS because this is special and not
 482 # really migrated, but used for migrations (for now)
 483 ######################################################
 484
 485 class MigrationData(Base):
 486     __tablename__ = "core__migrations"
 487
 488     name = Column(Unicode, primary_key=True)
 489     version = Column(Integer, nullable=False, default=0)
 490
 491 ######################################################
 492
 493
 494 def show_table_init(engine_uri):
 495     if engine_uri is None:
 496         engine_uri = 'sqlite:///:memory:'
 497     from sqlalchemy import create_engine
 498     engine = create_engine(engine_uri, echo=True)
 499
 500     Base.metadata.create_all(engine)
 501
 502
 503 if __name__ == '__main__':
 504     from sys import argv
 505     print repr(argv)
 506     if len(argv) == 2:
 507         uri = argv[1]
 508     else:
 509         uri = None
 510     show_table_init(uri)