cubicweb/dataimport/importer.py
author Denis Laxalde <denis.laxalde@logilab.fr>
Tue, 25 Jun 2019 16:42:58 +0200
branch3.26
changeset 12665 112f3c2b6f1e
parent 12625 ba5231e1aa45
child 12626 32ee89340e59
permissions -rw-r--r--
[debian] Install cubicweb-ctl from python2 build Previously, the binary package was empty. Now it is bound to the python2 package as support for python3 is "fresh". In order python2 and python3 packages to be co-installable, we drop the cubicweb-ctl file from both packages.
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
11139
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
     1
# copyright 2015-2016 LOGILAB S.A. (Paris, FRANCE), all rights reserved.
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     2
# contact http://www.logilab.fr -- mailto:contact@logilab.fr
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     3
#
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     4
# This program is free software: you can redistribute it and/or modify it under
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     5
# the terms of the GNU Lesser General Public License as published by the Free
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     6
# Software Foundation, either version 2.1 of the License, or (at your option)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     7
# any later version.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     8
#
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
     9
# This program is distributed in the hope that it will be useful, but WITHOUT
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    10
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    11
# FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    12
# details.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    13
#
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    14
# You should have received a copy of the GNU Lesser General Public License along
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    15
# with this program. If not, see <http://www.gnu.org/licenses/>.
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
    16
"""Data import of external entities.
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    17
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
    18
Main entry points:
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    19
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
    20
.. autoclass:: ExtEntitiesImporter
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
    21
.. autoclass:: ExtEntity
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    22
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
    23
Utilities:
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    24
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
    25
.. autofunction:: cwuri2eid
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
    26
.. autoclass:: RelationMapping
10514
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    27
.. autofunction:: cubicweb.dataimport.importer.use_extid_as_cwuri
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    28
"""
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    29
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    30
from collections import defaultdict
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    31
import logging
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    32
12625
ba5231e1aa45 [dataimport] Fix case when extid is text in use_extid_as_cwuri()
julien tayon <julien.tayon@logilab.fr>
parents: 12173
diff changeset
    33
import six
ba5231e1aa45 [dataimport] Fix case when extid is text in use_extid_as_cwuri()
julien tayon <julien.tayon@logilab.fr>
parents: 12173
diff changeset
    34
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    35
from logilab.mtconverter import xml_escape
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    36
11342
62a7100d774b [dataimport] Handle Bytes data in ExtEntity and convert them to Binary
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11313
diff changeset
    37
from cubicweb import Binary
62a7100d774b [dataimport] Handle Bytes data in ExtEntity and convert them to Binary
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11313
diff changeset
    38
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    39
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    40
def cwuri2eid(cnx, etypes, source_eid=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    41
    """Return a dictionary mapping cwuri to eid for entities of the given entity types and / or
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    42
    source.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    43
    """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    44
    assert source_eid or etypes, 'no entity types nor source specified'
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    45
    rql = 'Any U, X WHERE X cwuri U'
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    46
    args = {}
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    47
    if len(etypes) == 1:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    48
        rql += ', X is %s' % etypes[0]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    49
    elif etypes:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    50
        rql += ', X is IN (%s)' % ','.join(etypes)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    51
    if source_eid is not None:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    52
        rql += ', X cw_source S, S eid %(s)s'
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    53
        args['s'] = source_eid
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    54
    return dict(cnx.execute(rql, args))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    55
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
    56
10514
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    57
def use_extid_as_cwuri(extid2eid):
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    58
    """Return a generator of :class:`ExtEntity` objects that will set `cwuri`
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    59
    using entity's extid if the entity does not exist yet and has no `cwuri`
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    60
    defined.
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    61
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    62
    `extid2eid` is an extid to eid dictionary coming from an
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    63
    :class:`ExtEntitiesImporter` instance.
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    64
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    65
    Example usage:
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    66
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    67
    .. code-block:: python
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    68
11128
9b4de34ad394 [dataimport] update example to use standard importer
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11057
diff changeset
    69
        importer = ExtEntitiesImporter(cnx, store, import_log)
10514
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    70
        set_cwuri = use_extid_as_cwuri(importer.extid2eid)
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    71
        importer.import_entities(set_cwuri(extentities))
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    72
    """
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    73
    def use_extid_as_cwuri_filter(extentities):
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    74
        for extentity in extentities:
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    75
            if extentity.extid not in extid2eid:
12625
ba5231e1aa45 [dataimport] Fix case when extid is text in use_extid_as_cwuri()
julien tayon <julien.tayon@logilab.fr>
parents: 12173
diff changeset
    76
                cwuri = extentity.extid
ba5231e1aa45 [dataimport] Fix case when extid is text in use_extid_as_cwuri()
julien tayon <julien.tayon@logilab.fr>
parents: 12173
diff changeset
    77
                if isinstance(cwuri, six.binary_type):
ba5231e1aa45 [dataimport] Fix case when extid is text in use_extid_as_cwuri()
julien tayon <julien.tayon@logilab.fr>
parents: 12173
diff changeset
    78
                    cwuri = cwuri.decode('utf-8')
ba5231e1aa45 [dataimport] Fix case when extid is text in use_extid_as_cwuri()
julien tayon <julien.tayon@logilab.fr>
parents: 12173
diff changeset
    79
                extentity.values.setdefault('cwuri', set([cwuri]))
10514
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    80
            yield extentity
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    81
    return use_extid_as_cwuri_filter
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    82
b29d9904482e add use_extid_as_cwuri ext entity transform, that will be often necessary and not so easy to write at once
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 10461
diff changeset
    83
11139
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    84
def drop_extra_values(extentities, schema, import_log):
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    85
    """Return a generator of :class:`ExtEntity` objects that will ensure their attributes and
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    86
    inlined relations have a single value. When it's not the case, a warning will be recorded in
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    87
    the import log and one value among other will be kept (randomly).
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    88
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    89
    `schema` is the instance's schema, `import_log` is an instance of a class implementing the
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    90
    :class:`SimpleImportLog` interface.
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    91
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    92
    Example usage:
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    93
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    94
    .. code-block:: python
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    95
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    96
        importer = ExtEntitiesImporter(schema, store, import_log)
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    97
        importer.import_entities(drop_extra_values(extentities, schema, import_log))
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    98
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
    99
    """
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   100
    _get_rschema = schema.rschema
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   101
    for extentity in extentities:
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   102
        entity_dict = extentity.values
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   103
        for key, rtype, role in extentity.iter_rdefs():
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   104
            rschema = _get_rschema(rtype)
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   105
            if (rschema.final or (rschema.inlined and role == 'subject')) \
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   106
               and len(entity_dict[key]) > 1:
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   107
                values = ', '.join(repr(v) for v in entity_dict[key])
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   108
                import_log.record_warning(
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   109
                    "more than one value for attribute %r, only one will be kept: %s"
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   110
                    % (rtype, values), path=extentity.extid)
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   111
                entity_dict[key] = set([entity_dict[key].pop()])
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   112
        yield extentity
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   113
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   114
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   115
class RelationMapping(object):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   116
    """Read-only mapping from relation type to set of related (subject, object) eids.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   117
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   118
    If `source` is specified, only returns relations implying entities from
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   119
    this source.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   120
    """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   121
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   122
    def __init__(self, cnx, source=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   123
        self.cnx = cnx
10939
b30c2f49da57 [dataimport] Format strings with % instead of .format()
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10809
diff changeset
   124
        self._rql_template = 'Any S,O WHERE S %s O'
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   125
        self._kwargs = {}
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   126
        if source is not None:
10939
b30c2f49da57 [dataimport] Format strings with % instead of .format()
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10809
diff changeset
   127
            self._rql_template += ', S cw_source SO, O cw_source SO, SO eid %%(s)s'
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   128
            self._kwargs['s'] = source.eid
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   129
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   130
    def __getitem__(self, rtype):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   131
        """Return a set of (subject, object) eids already related by `rtype`"""
10939
b30c2f49da57 [dataimport] Format strings with % instead of .format()
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10809
diff changeset
   132
        rql = self._rql_template % rtype
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   133
        return set(tuple(x) for x in self.cnx.execute(rql, self._kwargs))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   134
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   135
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   136
class ExtEntity(object):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   137
    """Transitional representation of an entity for use in data importer.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   138
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   139
    An external entity has the following properties:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   140
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   141
    * ``extid`` (external id), an identifier for the ext entity,
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   142
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   143
    * ``etype`` (entity type), a string which must be the name of one entity type in the schema
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   144
      (eg. ``'Person'``, ``'Animal'``, ...),
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   145
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   146
    * ``values``, a dictionary whose keys are attribute or relation names from the schema (eg.
11342
62a7100d774b [dataimport] Handle Bytes data in ExtEntity and convert them to Binary
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11313
diff changeset
   147
      ``'first_name'``, ``'friend'``), and whose values are *sets*. For
62a7100d774b [dataimport] Handle Bytes data in ExtEntity and convert them to Binary
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11313
diff changeset
   148
      attributes of type Bytes, byte strings should be inserted in `values`.
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   149
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   150
    For instance:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   151
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   152
    .. code-block:: python
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   153
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   154
        ext_entity.extid = 'http://example.org/person/debby'
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   155
        ext_entity.etype = 'Person'
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   156
        ext_entity.values = {'first_name': set([u"Deborah", u"Debby"]),
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   157
                            'friend': set(['http://example.org/person/john'])}
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   158
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   159
    """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   160
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   161
    def __init__(self, etype, extid, values=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   162
        self.etype = etype
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   163
        self.extid = extid
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   164
        if values is None:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   165
            values = {}
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   166
        self.values = values
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   167
        self._schema = None
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   168
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   169
    def __repr__(self):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   170
        return '<%s %s %s>' % (self.etype, self.extid, self.values)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   171
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   172
    def iter_rdefs(self):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   173
        """Yield (key, rtype, role) defined in `.values` dict, with:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   174
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   175
        * `key` is the original key in `.values` (i.e. the relation type or a 2-uple (relation type,
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   176
          role))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   177
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   178
        * `rtype` is a yams relation type, expected to be found in the schema (attribute or
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   179
          relation)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   180
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   181
        * `role` is the role of the entity in the relation, 'subject' or 'object'
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   182
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   183
        Iteration is done on a copy of the keys so values may be inserted/deleted during it.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   184
        """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   185
        for key in list(self.values):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   186
            if isinstance(key, tuple):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   187
                rtype, role = key
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   188
                assert role in ('subject', 'object'), key
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   189
                yield key, rtype, role
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   190
            else:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   191
                yield key, key, 'subject'
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   192
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   193
    def prepare(self, schema):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   194
        """Prepare an external entity for later insertion:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   195
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   196
        * ensure attributes and inlined relations have a single value
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   197
        * turn set([value]) into value and remove key associated to empty set
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   198
        * remove non inlined relations and return them as a [(e1key, relation, e2key)] list
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   199
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   200
        Return a list of non inlined relations that may be inserted later, each relations defined by
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   201
        a 3-tuple (subject extid, relation type, object extid).
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   202
11139
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   203
        The instance's schema is given as argument.
df928a3a94e3 [dataimport] add a filter function to not fail if some extentity has several values for an attribute of final relation
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11128
diff changeset
   204
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   205
        Take care the importer may call this method several times.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   206
        """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   207
        assert self._schema is None, 'prepare() has already been called for %s' % self
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   208
        self._schema = schema
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   209
        eschema = schema.eschema(self.etype)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   210
        deferred = []
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   211
        entity_dict = self.values
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   212
        for key, rtype, role in self.iter_rdefs():
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   213
            rschema = schema.rschema(rtype)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   214
            if rschema.final or (rschema.inlined and role == 'subject'):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   215
                assert len(entity_dict[key]) <= 1, \
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   216
                    "more than one value for %s: %s (%s)" % (rtype, entity_dict[key], self.extid)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   217
                if entity_dict[key]:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   218
                    entity_dict[rtype] = entity_dict[key].pop()
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   219
                    if key != rtype:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   220
                        del entity_dict[key]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   221
                    if (rschema.final and eschema.has_metadata(rtype, 'format')
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   222
                            and not rtype + '_format' in entity_dict):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   223
                        entity_dict[rtype + '_format'] = u'text/plain'
11393
e148b384a782 [dataimport] Restore handling of Binary in ExtEntity's values
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11353
diff changeset
   224
                    if (rschema.final
e148b384a782 [dataimport] Restore handling of Binary in ExtEntity's values
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11353
diff changeset
   225
                            and eschema.rdef(rtype).object.type == 'Bytes'
e148b384a782 [dataimport] Restore handling of Binary in ExtEntity's values
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11353
diff changeset
   226
                            and not isinstance(entity_dict[rtype], Binary)):
11342
62a7100d774b [dataimport] Handle Bytes data in ExtEntity and convert them to Binary
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11313
diff changeset
   227
                        entity_dict[rtype] = Binary(entity_dict[rtype])
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   228
                else:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   229
                    del entity_dict[key]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   230
            else:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   231
                for target_extid in entity_dict.pop(key):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   232
                    if role == 'subject':
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   233
                        deferred.append((self.extid, rtype, target_extid))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   234
                    else:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   235
                        deferred.append((target_extid, rtype, self.extid))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   236
        return deferred
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   237
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   238
    def is_ready(self, extid2eid):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   239
        """Return True if the ext entity is ready, i.e. has all the URIs used in inlined relations
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   240
        currently existing.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   241
        """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   242
        assert self._schema, 'prepare() method should be called first on %s' % self
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   243
        # as .prepare has been called, we know that .values only contains subject relation *type* as
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   244
        # key (no more (rtype, role) tuple)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   245
        schema = self._schema
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   246
        entity_dict = self.values
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   247
        for rtype in entity_dict:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   248
            rschema = schema.rschema(rtype)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   249
            if not rschema.final:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   250
                # .prepare() should drop other cases from the entity dict
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   251
                assert rschema.inlined
12167
1ca864397424 [cleanup] Fix undetected pep8 error
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11943
diff changeset
   252
                if entity_dict[rtype] not in extid2eid:
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   253
                    return False
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   254
        # entity is ready, replace all relation's extid by eids
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   255
        for rtype in entity_dict:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   256
            rschema = schema.rschema(rtype)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   257
            if rschema.inlined:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   258
                entity_dict[rtype] = extid2eid[entity_dict[rtype]]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   259
        return True
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   260
12173
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   261
    def why_not_ready(self, extid2eid):
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   262
        """Return some text explaining why this ext entity is not ready.
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   263
        """
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   264
        assert self._schema, 'prepare() method should be called first on %s' % self
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   265
        # as .prepare has been called, we know that .values only contains subject relation *type* as
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   266
        # key (no more (rtype, role) tuple)
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   267
        schema = self._schema
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   268
        entity_dict = self.values
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   269
        for rtype in entity_dict:
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   270
            rschema = schema.rschema(rtype)
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   271
            if not rschema.final:
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   272
                if entity_dict[rtype] not in extid2eid:
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   273
                    return u'inlined relation %s is not present (%s)' % (rtype, entity_dict[rtype])
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   274
        raise AssertionError('this external entity seems actually ready for insertion')
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   275
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   276
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   277
class ExtEntitiesImporter(object):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   278
    """This class is responsible for importing externals entities, that is instances of
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   279
    :class:`ExtEntity`, into CubicWeb entities.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   280
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   281
    :param schema: the CubicWeb's instance schema
11313
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   282
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   283
    :param store: a CubicWeb `Store`
11313
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   284
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   285
    :param extid2eid: optional {extid: eid} dictionary giving information on existing entities. It
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   286
        will be completed during import. You may want to use :func:`cwuri2eid` to build it.
11313
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   287
11943
760deab5413e [dataimport] Fix "existing_relations" parameter name in ExtEntitiesImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 11393
diff changeset
   288
    :param existing_relations: optional {rtype: set((subj eid, obj eid))} mapping giving information
11313
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   289
        on existing relations of a given type. You may want to use :class:`RelationMapping` to build
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   290
        it.
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   291
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   292
    :param etypes_order_hint: optional ordered iterable on entity types, giving an hint on the
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   293
        order in which they should be attempted to be imported
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   294
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   295
    :param import_log: optional object implementing the :class:`SimpleImportLog` interface to
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   296
        record events occuring during the import
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   297
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   298
    :param raise_on_error: optional boolean flag - default to false, indicating whether errors
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   299
        should be raised or logged. You usually want them to be raised during test but to be logged
682b15eb2dd2 [dataimport] flake8
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 11279
diff changeset
   300
        in production.
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   301
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   302
    Instances of this class are meant to import external entities through :meth:`import_entities`
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   303
    which handles a stream of :class:`ExtEntity`. One may then plug arbitrary filters into the
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   304
    external entities stream.
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   305
10461
37644c518705 [doc] Add a tutorial and extend documentation for ExtEntityImporter
Denis Laxalde <denis.laxalde@logilab.fr>
parents: 10460
diff changeset
   306
    .. automethod:: import_entities
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   307
    """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   308
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   309
    def __init__(self, schema, store, extid2eid=None, existing_relations=None,
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   310
                 etypes_order_hint=(), import_log=None, raise_on_error=False):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   311
        self.schema = schema
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   312
        self.store = store
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   313
        self.extid2eid = extid2eid if extid2eid is not None else {}
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   314
        self.existing_relations = (existing_relations if existing_relations is not None
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   315
                                   else defaultdict(set))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   316
        self.etypes_order_hint = etypes_order_hint
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   317
        if import_log is None:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   318
            import_log = SimpleImportLog('<unspecified>')
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   319
        self.import_log = import_log
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   320
        self.raise_on_error = raise_on_error
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   321
        # set of created/updated eids
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   322
        self.created = set()
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   323
        self.updated = set()
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   324
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   325
    def import_entities(self, ext_entities):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   326
        """Import given external entities (:class:`ExtEntity`) stream (usually a generator)."""
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   327
        # {etype: [etype dict]} of entities that are in the import queue
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   328
        queue = {}
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   329
        # order entity dictionaries then create/update them
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   330
        deferred = self._import_entities(ext_entities, queue)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   331
        # create deferred relations that don't exist already
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   332
        missing_relations = self.prepare_insert_deferred_relations(deferred)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   333
        self._warn_about_missing_work(queue, missing_relations)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   334
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   335
    def _import_entities(self, ext_entities, queue):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   336
        extid2eid = self.extid2eid
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   337
        deferred = {}  # non inlined relations that may be deferred
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   338
        self.import_log.record_debug('importing entities')
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   339
        for ext_entity in self.iter_ext_entities(ext_entities, deferred, queue):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   340
            try:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   341
                eid = extid2eid[ext_entity.extid]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   342
            except KeyError:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   343
                self.prepare_insert_entity(ext_entity)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   344
            else:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   345
                if ext_entity.values:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   346
                    self.prepare_update_entity(ext_entity, eid)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   347
        return deferred
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   348
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   349
    def iter_ext_entities(self, ext_entities, deferred, queue):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   350
        """Yield external entities in an order which attempts to satisfy
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   351
        schema constraints (inlined / cardinality) and to optimize the import.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   352
        """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   353
        schema = self.schema
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   354
        extid2eid = self.extid2eid
11275
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   355
        order_hint = list(self.etypes_order_hint)
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   356
        for ext_entity in ext_entities:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   357
            # check data in the transitional representation and prepare it for
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   358
            # later insertion in the database
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   359
            for subject_uri, rtype, object_uri in ext_entity.prepare(schema):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   360
                deferred.setdefault(rtype, set()).add((subject_uri, object_uri))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   361
            if not ext_entity.is_ready(extid2eid):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   362
                queue.setdefault(ext_entity.etype, []).append(ext_entity)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   363
                continue
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   364
            yield ext_entity
11275
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   365
            if not queue:
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   366
                continue
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   367
            # check for some entities in the queue that may now be ready. We'll have to restart
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   368
            # search for ready entities until no one is generated
11275
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   369
            for etype in queue:
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   370
                if etype not in order_hint:
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   371
                    order_hint.append(etype)
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   372
            new = True
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   373
            while new:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   374
                new = False
11275
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   375
                for etype in order_hint:
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   376
                    if etype in queue:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   377
                        new_queue = []
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   378
                        for ext_entity in queue[etype]:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   379
                            if ext_entity.is_ready(extid2eid):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   380
                                yield ext_entity
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   381
                                # may unlock entity previously handled within this loop
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   382
                                new = True
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   383
                            else:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   384
                                new_queue.append(ext_entity)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   385
                        if new_queue:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   386
                            queue[etype][:] = new_queue
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   387
                        else:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   388
                            del queue[etype]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   389
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   390
    def prepare_insert_entity(self, ext_entity):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   391
        """Call the store to prepare insertion of the given external entity"""
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   392
        eid = self.store.prepare_insert_entity(ext_entity.etype, **ext_entity.values)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   393
        self.extid2eid[ext_entity.extid] = eid
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   394
        self.created.add(eid)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   395
        return eid
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   396
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   397
    def prepare_update_entity(self, ext_entity, eid):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   398
        """Call the store to prepare update of the given external entity"""
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   399
        self.store.prepare_update_entity(ext_entity.etype, eid, **ext_entity.values)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   400
        self.updated.add(eid)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   401
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   402
    def prepare_insert_deferred_relations(self, deferred):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   403
        """Call the store to insert deferred relations (not handled during insertion/update for
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   404
        entities). Return a list of relations `[(subj ext id, obj ext id)]` that may not be inserted
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   405
        because the target entities don't exists yet.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   406
        """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   407
        prepare_insert_relation = self.store.prepare_insert_relation
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   408
        rschema = self.schema.rschema
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   409
        extid2eid = self.extid2eid
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   410
        missing_relations = []
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   411
        for rtype, relations in deferred.items():
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   412
            self.import_log.record_debug('importing %s %s relations' % (len(relations), rtype))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   413
            symmetric = rschema(rtype).symmetric
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   414
            existing = self.existing_relations[rtype]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   415
            for subject_uri, object_uri in relations:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   416
                try:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   417
                    subject_eid = extid2eid[subject_uri]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   418
                    object_eid = extid2eid[object_uri]
11275
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   419
                except KeyError as exc:
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   420
                    missing_relations.append((subject_uri, rtype, object_uri, exc))
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   421
                    continue
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   422
                if (subject_eid, object_eid) not in existing:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   423
                    prepare_insert_relation(subject_eid, rtype, object_eid)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   424
                    existing.add((subject_eid, object_eid))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   425
                    if symmetric:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   426
                        existing.add((object_eid, subject_eid))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   427
        return missing_relations
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   428
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   429
    def _warn_about_missing_work(self, queue, missing_relations):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   430
        error = self.import_log.record_error
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   431
        if queue:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   432
            msgs = ["can't create some entities, is there some cycle or "
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   433
                    "missing data?"]
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   434
            for ext_entities in queue.values():
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   435
                for ext_entity in ext_entities:
12173
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   436
                    msg = '{}: {}'.format(ext_entity, ext_entity.why_not_ready(self.extid2eid))
d13fc09301bd [dataimport] Add explanation about why external entities can't be inserted
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 12167
diff changeset
   437
                    msgs.append(msg)
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   438
            map(error, msgs)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   439
            if self.raise_on_error:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   440
                raise Exception('\n'.join(msgs))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   441
        if missing_relations:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   442
            msgs = ["can't create some relations, is there missing data?"]
11275
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   443
            for subject_uri, rtype, object_uri, exc in missing_relations:
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   444
                msgs.append("Could not find %s when trying to insert (%s, %s, %s)"
814f54d6183b [dataimport] order of ExtEntities should be irrelevant (closes #13117472)
Nicolas Chauvat <nicolas.chauvat@logilab.fr>
parents: 10939
diff changeset
   445
                            % (exc, subject_uri, rtype, object_uri))
10460
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   446
            map(error, msgs)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   447
            if self.raise_on_error:
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   448
                raise Exception('\n'.join(msgs))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   449
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   450
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   451
class SimpleImportLog(object):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   452
    """Fake CWDataImport log using a simple text format.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   453
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   454
    Useful to display logs in the UI instead of storing them to the
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   455
    database.
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   456
    """
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   457
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   458
    def __init__(self, filename):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   459
        self.logs = []
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   460
        self.filename = filename
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   461
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   462
    def record_debug(self, msg, path=None, line=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   463
        self._log(logging.DEBUG, msg, path, line)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   464
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   465
    def record_info(self, msg, path=None, line=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   466
        self._log(logging.INFO, msg, path, line)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   467
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   468
    def record_warning(self, msg, path=None, line=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   469
        self._log(logging.WARNING, msg, path, line)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   470
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   471
    def record_error(self, msg, path=None, line=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   472
        self._log(logging.ERROR, msg, path, line)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   473
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   474
    def record_fatal(self, msg, path=None, line=None):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   475
        self._log(logging.FATAL, msg, path, line)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   476
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   477
    def _log(self, severity, msg, path, line):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   478
        encodedmsg = u'%s\t%s\t%s\t%s' % (severity, self.filename,
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   479
                                          line or u'', msg)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   480
        self.logs.append(encodedmsg)
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   481
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   482
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   483
class HTMLImportLog(SimpleImportLog):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   484
    """Fake CWDataImport log using a simple HTML format."""
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   485
    def __init__(self, filename):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   486
        super(HTMLImportLog, self).__init__(xml_escape(filename))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   487
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   488
    def _log(self, severity, msg, path, line):
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   489
        encodedmsg = u'%s\t%s\t%s\t%s<br/>' % (severity, self.filename,
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   490
                                               line or u'', xml_escape(msg))
d260722f2453 [dataimport] introduce the importer and extentity classes
Yann Voté <yann.vote@logilab.fr>
parents:
diff changeset
   491
        self.logs.append(encodedmsg)