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
  24 from sqlalchemy import Column, Integer, Unicode, UnicodeText, DateTime, \
  25         Boolean, ForeignKey, UniqueConstraint, PrimaryKeyConstraint, \
  26         SmallInteger
  27 from sqlalchemy.orm import relationship, backref
  28 from sqlalchemy.orm.collections import attribute_mapped_collection
  29 from sqlalchemy.sql.expression import desc
  30 from sqlalchemy.ext.associationproxy import association_proxy
  31 from sqlalchemy.util import memoized_property
  32
  33 from mediagoblin.db.extratypes import PathTupleWithSlashes, JSONEncoded
  34 from mediagoblin.db.base import Base, DictReadAttrProxy
  35 from mediagoblin.db.mixin import UserMixin, MediaEntryMixin, MediaCommentMixin, CollectionMixin, CollectionItemMixin
  36 from mediagoblin.tools.files import delete_media_files
  37 from mediagoblin.tools.common import import_component
  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         # Collections get deleted by relationships.
  88
  89         media_entries = MediaEntry.query.filter(MediaEntry.uploader == self.id)
  90         for media in media_entries:
  91             # TODO: Make sure that "MediaEntry.delete()" also deletes
  92             # all related files/Comments
  93             media.delete(del_orphan_tags=False, commit=False)
  94
  95         # Delete now unused tags
  96         # TODO: import here due to cyclic imports!!! This cries for refactoring
  97         from mediagoblin.db.util import clean_orphan_tags
  98         clean_orphan_tags(commit=False)
  99
 100         # Delete user, pass through commit=False/True in kwargs
 101         super(User, self).delete(**kwargs)
 102         _log.info('Deleted user "{0}" account'.format(self.username))
 103
 104
 105 class MediaEntry(Base, MediaEntryMixin):
 106     """
 107     TODO: Consider fetching the media_files using join
 108     """
 109     __tablename__ = "core__media_entries"
 110
 111     id = Column(Integer, primary_key=True)
 112     uploader = Column(Integer, ForeignKey(User.id), nullable=False, index=True)
 113     title = Column(Unicode, nullable=False)
 114     slug = Column(Unicode)
 115     created = Column(DateTime, nullable=False, default=datetime.datetime.now,
 116         index=True)
 117     description = Column(UnicodeText) # ??
 118     media_type = Column(Unicode, nullable=False)
 119     state = Column(Unicode, default=u'unprocessed', nullable=False)
 120         # or use sqlalchemy.types.Enum?
 121     license = Column(Unicode)
 122     collected = Column(Integer, default=0)
 123
 124     fail_error = Column(Unicode)
 125     fail_metadata = Column(JSONEncoded)
 126
 127     transcoding_progress = Column(SmallInteger)
 128
 129     queued_media_file = Column(PathTupleWithSlashes)
 130
 131     queued_task_id = Column(Unicode)
 132
 133     __table_args__ = (
 134         UniqueConstraint('uploader', 'slug'),
 135         {})
 136
 137     get_uploader = relationship(User)
 138
 139     media_files_helper = relationship("MediaFile",
 140         collection_class=attribute_mapped_collection("name"),
 141         cascade="all, delete-orphan"
 142         )
 143     media_files = association_proxy('media_files_helper', 'file_path',
 144         creator=lambda k, v: MediaFile(name=k, file_path=v)
 145         )
 146
 147     attachment_files_helper = relationship("MediaAttachmentFile",
 148         cascade="all, delete-orphan",
 149         order_by="MediaAttachmentFile.created"
 150         )
 151     attachment_files = association_proxy("attachment_files_helper", "dict_view",
 152         creator=lambda v: MediaAttachmentFile(
 153             name=v["name"], filepath=v["filepath"])
 154         )
 155
 156     tags_helper = relationship("MediaTag",
 157         cascade="all, delete-orphan" # should be automatically deleted
 158         )
 159     tags = association_proxy("tags_helper", "dict_view",
 160         creator=lambda v: MediaTag(name=v["name"], slug=v["slug"])
 161         )
 162
 163     collections_helper = relationship("CollectionItem",
 164         cascade="all, delete-orphan"
 165         )
 166     collections = association_proxy("collections_helper", "in_collection")
 167
 168     ## TODO
 169     # fail_error
 170
 171     def get_comments(self, ascending=False):
 172         order_col = MediaComment.created
 173         if not ascending:
 174             order_col = desc(order_col)
 175         return self.all_comments.order_by(order_col)
 176
 177     def url_to_prev(self, urlgen):
 178         """get the next 'newer' entry by this user"""
 179         media = MediaEntry.query.filter(
 180             (MediaEntry.uploader == self.uploader)
 181             & (MediaEntry.state == u'processed')
 182             & (MediaEntry.id > self.id)).order_by(MediaEntry.id).first()
 183
 184         if media is not None:
 185             return media.url_for_self(urlgen)
 186
 187     def url_to_next(self, urlgen):
 188         """get the next 'older' entry by this user"""
 189         media = MediaEntry.query.filter(
 190             (MediaEntry.uploader == self.uploader)
 191             & (MediaEntry.state == u'processed')
 192             & (MediaEntry.id < self.id)).order_by(desc(MediaEntry.id)).first()
 193
 194         if media is not None:
 195             return media.url_for_self(urlgen)
 196
 197     @property
 198     def media_data(self):
 199         return getattr(self, self.media_data_ref)
 200
 201     def media_data_init(self, **kwargs):
 202         """
 203         Initialize or update the contents of a media entry's media_data row
 204         """
 205         media_data = self.media_data
 206
 207         if media_data is None:
 208             # Get the correct table:
 209             table = import_component(self.media_type + '.models:DATA_MODEL')
 210             # No media data, so actually add a new one
 211             media_data = table(**kwargs)
 212             # Get the relationship set up.
 213             media_data.get_media_entry = self
 214         else:
 215             # Update old media data
 216             for field, value in kwargs.iteritems():
 217                 setattr(media_data, field, value)
 218
 219     @memoized_property
 220     def media_data_ref(self):
 221         return import_component(self.media_type + '.models:BACKREF_NAME')
 222
 223     def __repr__(self):
 224         safe_title = self.title.encode('ascii', 'replace')
 225
 226         return '<{classname} {id}: {title}>'.format(
 227                 classname=self.__class__.__name__,
 228                 id=self.id,
 229                 title=safe_title)
 230
 231     def delete(self, del_orphan_tags=True, **kwargs):
 232         """Delete MediaEntry and all related files/attachments/comments
 233
 234         This will *not* automatically delete unused collections, which
 235         can remain empty...
 236
 237         :param del_orphan_tags: True/false if we delete unused Tags too
 238         :param commit: True/False if this should end the db transaction"""
 239         # User's CollectionItems are automatically deleted via "cascade".
 240         # Comments on this Media are deleted by cascade, hopefully.
 241
 242         # Delete all related files/attachments
 243         try:
 244             delete_media_files(self)
 245         except OSError, error:
 246             # Returns list of files we failed to delete
 247             _log.error('No such files from the user "{1}" to delete: '
 248                        '{0}'.format(str(error), self.get_uploader))
 249         _log.info('Deleted Media entry id "{0}"'.format(self.id))
 250         # Related MediaTag's are automatically cleaned, but we might
 251         # want to clean out unused Tag's too.
 252         if del_orphan_tags:
 253             # TODO: Import here due to cyclic imports!!!
 254             #       This cries for refactoring
 255             from mediagoblin.db.util import clean_orphan_tags
 256             clean_orphan_tags(commit=False)
 257         # pass through commit=False/True in kwargs
 258         super(MediaEntry, self).delete(**kwargs)
 259
 260
 261 class FileKeynames(Base):
 262     """
 263     keywords for various places.
 264     currently the MediaFile keys
 265     """
 266     __tablename__ = "core__file_keynames"
 267     id = Column(Integer, primary_key=True)
 268     name = Column(Unicode, unique=True)
 269
 270     def __repr__(self):
 271         return "<FileKeyname %r: %r>" % (self.id, self.name)
 272
 273     @classmethod
 274     def find_or_new(cls, name):
 275         t = cls.query.filter_by(name=name).first()
 276         if t is not None:
 277             return t
 278         return cls(name=name)
 279
 280
 281 class MediaFile(Base):
 282     """
 283     TODO: Highly consider moving "name" into a new table.
 284     TODO: Consider preloading said table in software
 285     """
 286     __tablename__ = "core__mediafiles"
 287
 288     media_entry = Column(
 289         Integer, ForeignKey(MediaEntry.id),
 290         nullable=False)
 291     name_id = Column(SmallInteger, ForeignKey(FileKeynames.id), nullable=False)
 292     file_path = Column(PathTupleWithSlashes)
 293
 294     __table_args__ = (
 295         PrimaryKeyConstraint('media_entry', 'name_id'),
 296         {})
 297
 298     def __repr__(self):
 299         return "<MediaFile %s: %r>" % (self.name, self.file_path)
 300
 301     name_helper = relationship(FileKeynames, lazy="joined", innerjoin=True)
 302     name = association_proxy('name_helper', 'name',
 303         creator=FileKeynames.find_or_new
 304         )
 305
 306
 307 class MediaAttachmentFile(Base):
 308     __tablename__ = "core__attachment_files"
 309
 310     id = Column(Integer, primary_key=True)
 311     media_entry = Column(
 312         Integer, ForeignKey(MediaEntry.id),
 313         nullable=False)
 314     name = Column(Unicode, nullable=False)
 315     filepath = Column(PathTupleWithSlashes)
 316     created = Column(DateTime, nullable=False, default=datetime.datetime.now)
 317
 318     @property
 319     def dict_view(self):
 320         """A dict like view on this object"""
 321         return DictReadAttrProxy(self)
 322
 323
 324 class Tag(Base):
 325     __tablename__ = "core__tags"
 326
 327     id = Column(Integer, primary_key=True)
 328     slug = Column(Unicode, nullable=False, unique=True)
 329
 330     def __repr__(self):
 331         return "<Tag %r: %r>" % (self.id, self.slug)
 332
 333     @classmethod
 334     def find_or_new(cls, slug):
 335         t = cls.query.filter_by(slug=slug).first()
 336         if t is not None:
 337             return t
 338         return cls(slug=slug)
 339
 340
 341 class MediaTag(Base):
 342     __tablename__ = "core__media_tags"
 343
 344     id = Column(Integer, primary_key=True)
 345     media_entry = Column(
 346         Integer, ForeignKey(MediaEntry.id),
 347         nullable=False, index=True)
 348     tag = Column(Integer, ForeignKey(Tag.id), nullable=False, index=True)
 349     name = Column(Unicode)
 350     # created = Column(DateTime, nullable=False, default=datetime.datetime.now)
 351
 352     __table_args__ = (
 353         UniqueConstraint('tag', 'media_entry'),
 354         {})
 355
 356     tag_helper = relationship(Tag)
 357     slug = association_proxy('tag_helper', 'slug',
 358         creator=Tag.find_or_new
 359         )
 360
 361     def __init__(self, name=None, slug=None):
 362         Base.__init__(self)
 363         if name is not None:
 364             self.name = name
 365         if slug is not None:
 366             self.tag_helper = Tag.find_or_new(slug)
 367
 368     @property
 369     def dict_view(self):
 370         """A dict like view on this object"""
 371         return DictReadAttrProxy(self)
 372
 373
 374 class MediaComment(Base, MediaCommentMixin):
 375     __tablename__ = "core__media_comments"
 376
 377     id = Column(Integer, primary_key=True)
 378     media_entry = Column(
 379         Integer, ForeignKey(MediaEntry.id), nullable=False, index=True)
 380     author = Column(Integer, ForeignKey(User.id), nullable=False)
 381     created = Column(DateTime, nullable=False, default=datetime.datetime.now)
 382     content = Column(UnicodeText, nullable=False)
 383
 384     # Cascade: Comments are owned by their creator. So do the full thing.
 385     # lazy=dynamic: People might post a *lot* of comments,
 386     #     so make the "posted_comments" a query-like thing.
 387     get_author = relationship(User,
 388                               backref=backref("posted_comments",
 389                                               lazy="dynamic",
 390                                               cascade="all, delete-orphan"))
 391
 392     # Cascade: Comments are somewhat owned by their MediaEntry.
 393     #     So do the full thing.
 394     # lazy=dynamic: MediaEntries might have many comments,
 395     #     so make the "all_comments" a query-like thing.
 396     get_media_entry = relationship(MediaEntry,
 397                                    backref=backref("all_comments",
 398                                                    lazy="dynamic",
 399                                                    cascade="all, delete-orphan"))
 400
 401
 402 class Collection(Base, CollectionMixin):
 403     """An 'album' or 'set' of media by a user.
 404
 405     On deletion, contained CollectionItems get automatically reaped via
 406     SQL cascade"""
 407     __tablename__ = "core__collections"
 408
 409     id = Column(Integer, primary_key=True)
 410     title = Column(Unicode, nullable=False)
 411     slug = Column(Unicode)
 412     created = Column(DateTime, nullable=False, default=datetime.datetime.now,
 413                      index=True)
 414     description = Column(UnicodeText)
 415     creator = Column(Integer, ForeignKey(User.id), nullable=False)
 416     # TODO: No of items in Collection. Badly named, can we migrate to num_items?
 417     items = Column(Integer, default=0)
 418
 419     # Cascade: Collections are owned by their creator. So do the full thing.
 420     get_creator = relationship(User,
 421                                backref=backref("collections",
 422                                                cascade="all, delete-orphan"))
 423
 424     __table_args__ = (
 425         UniqueConstraint('creator', 'slug'),
 426         {})
 427
 428     def get_collection_items(self, ascending=False):
 429         #TODO, is this still needed with self.collection_items being available?
 430         order_col = CollectionItem.position
 431         if not ascending:
 432             order_col = desc(order_col)
 433         return CollectionItem.query.filter_by(
 434             collection=self.id).order_by(order_col)
 435
 436
 437 class CollectionItem(Base, CollectionItemMixin):
 438     __tablename__ = "core__collection_items"
 439
 440     id = Column(Integer, primary_key=True)
 441     media_entry = Column(
 442         Integer, ForeignKey(MediaEntry.id), nullable=False, index=True)
 443     collection = Column(Integer, ForeignKey(Collection.id), nullable=False)
 444     note = Column(UnicodeText, nullable=True)
 445     added = Column(DateTime, nullable=False, default=datetime.datetime.now)
 446     position = Column(Integer)
 447
 448     # Cascade: CollectionItems are owned by their Collection. So do the full thing.
 449     in_collection = relationship(Collection,
 450                                  backref=backref(
 451                                      "collection_items",
 452                                      cascade="all, delete-orphan"))
 453
 454     get_media_entry = relationship(MediaEntry)
 455
 456     __table_args__ = (
 457         UniqueConstraint('collection', 'media_entry'),
 458         {})
 459
 460     @property
 461     def dict_view(self):
 462         """A dict like view on this object"""
 463         return DictReadAttrProxy(self)
 464
 465
 466 class ProcessingMetaData(Base):
 467     __tablename__ = 'core__processing_metadata'
 468
 469     id = Column(Integer, primary_key=True)
 470     media_entry_id = Column(Integer, ForeignKey(MediaEntry.id), nullable=False,
 471             index=True)
 472     media_entry = relationship(MediaEntry,
 473             backref=backref('processing_metadata',
 474                 cascade='all, delete-orphan'))
 475     callback_url = Column(Unicode)
 476
 477     @property
 478     def dict_view(self):
 479         """A dict like view on this object"""
 480         return DictReadAttrProxy(self)
 481
 482
 483 MODELS = [
 484     User, MediaEntry, Tag, MediaTag, MediaComment, Collection, CollectionItem, MediaFile, FileKeynames,
 485     MediaAttachmentFile, ProcessingMetaData]
 486
 487
 488 ######################################################
 489 # Special, migrations-tracking table
 490 #
 491 # Not listed in MODELS because this is special and not
 492 # really migrated, but used for migrations (for now)
 493 ######################################################
 494
 495 class MigrationData(Base):
 496     __tablename__ = "core__migrations"
 497
 498     name = Column(Unicode, primary_key=True)
 499     version = Column(Integer, nullable=False, default=0)
 500
 501 ######################################################
 502
 503
 504 def show_table_init(engine_uri):
 505     if engine_uri is None:
 506         engine_uri = 'sqlite:///:memory:'
 507     from sqlalchemy import create_engine
 508     engine = create_engine(engine_uri, echo=True)
 509
 510     Base.metadata.create_all(engine)
 511
 512
 513 if __name__ == '__main__':
 514     from sys import argv
 515     print repr(argv)
 516     if len(argv) == 2:
 517         uri = argv[1]
 518     else:
 519         uri = None
 520     show_table_init(uri)