Allow a generator to be provided instead of a List

[skinkie]

📒 Description

Considering you are writing a very big tree, you don't want to materialise that tree before it ends up in a file in a list first. Ideally the entire subtree should only be rendered just in time.

🔗 What I've Done

I have allowed a Generator to be handled as a List.

💬 Comments

There might be more places this needs to be changed. Since List[] is used everywhere, type hinting fails.

🛫 Checklist

• [ ] Updated docs
• [ ] Added unit-tests
• [ ] Sample tests pass
• [ ] W3C tests pass

import sqlite3

from xsdata.formats.dataclass.context import XmlContext
from xsdata.formats.dataclass.parsers import XmlParser
from xsdata.formats.dataclass.parsers.config import ParserConfig
from xsdata.formats.dataclass.parsers.handlers import LxmlEventHandler
from xsdata.formats.dataclass.serializers import XmlSerializer
from xsdata.formats.dataclass.serializers.config import SerializerConfig
from xsdata.models.datatype import XmlDateTime

from netex import PublicationDelivery, ParticipantRef, MultilingualString, DataObjectsRelStructure, GeneralFrame, \
    GeneralFrameMembersRelStructure, ServiceJourney

serializer_config = SerializerConfig(ignore_default_attributes=True, xml_declaration=True)
serializer_config.pretty_print = True
serializer_config.ignore_default_attributes = True
serializer = XmlSerializer(config=serializer_config)

context = XmlContext()
config = ParserConfig(fail_on_unknown_properties=False)
parser = XmlParser(context=context, config=config, handler=LxmlEventHandler)

def load_generator(con, clazz, limit=None):
    type = getattr(clazz.Meta, 'name', clazz.__name__)

    cur = con.cursor()
    if limit is None:
        cur.execute(f"SELECT object FROM {type};")
    else:
        cur.execute(f"SELECT object FROM {type} LIMIT {limit};")

    while True:
        xml = cur.fetchone()
        if xml is None:
            break
        yield parser.from_bytes(xml[0], clazz)

with sqlite3.connect("/tmp/netex.sqlite") as con:
    publication_delivery = PublicationDelivery(
                publication_timestamp=XmlDateTime.now(),
                participant_ref=ParticipantRef(value="NDOV"),
                description=MultilingualString(value="Huge XML Serializer test"),
                data_objects=DataObjectsRelStructure(choice=[GeneralFrame(members=GeneralFrameMembersRelStructure(choice=load_generator(con, ServiceJourney, 10)))]),
                version="ntx:1.1",
            )

ns_map = {'': 'http://www.netex.org.uk/netex', 'gml': 'http://www.opengis.net/gml/3.2'}
with open('netex-output/huge.xml', 'w') as out:
    serializer.write(out, publication_delivery, ns_map)

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 100.00%. Comparing base (cce0a16) to head (d4b3e5b).

:exclamation: Current head d4b3e5b differs from pull request most recent head 67d847f

Please upload reports for the commit 67d847f to get more accurate results.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #1030 +/- ## ========================================= Coverage 100.00% 100.00% ========================================= Files 115 115 Lines 9238 9265 +27 Branches 2179 2190 +11 ========================================= + Hits 9238 9265 +27 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[tefra]

We need to properly support the Iterable type annotation in model fields, and of course support this for both xml/json serialization @skinkie.

[skinkie]

We need to properly support the Iterable type annotation in model fields, and of course support this for both xml/json serialization @skinkie.

But would this something you would support from an architecture point of view?

[tefra]

We need to properly support the Iterable type annotation in model fields, and of course support this for both xml/json serialization @skinkie.

But would this something you would support from an architecture point of view?

Yes

[skinkie]

@tefra How would you like to proceed? Materialise List or Tuple in the tests?

[skinkie]

Testing it with my own code results in this error. So it is clearly not done yet.

xsdata.exceptions.XmlContextError: Error on DataObjectsRelStructure::choice: Xml Elements does not support typing `typing.Iterable[typing.Union[netex.general_version_frame_structure.CompositeFrame, netex.mobility_journey_frame.MobilityJourneyFrame, netex.mobility_service_frame.MobilityServiceFrame, netex.sales_transaction_frame.SalesTransactionFrame, netex.fare_frame.FareFrame, netex.driver_schedule_frame.DriverScheduleFrame, netex.vehicle_schedule_frame.VehicleScheduleFrame, netex.service_frame.ServiceFrame, netex.timetable_frame.TimetableFrame, netex.site_frame.SiteFrame, netex.infrastructure_frame.InfrastructureFrame, netex.general_version_frame_structure.GeneralFrame, netex.resource_frame.ResourceFrame, netex.service_calendar_frame.ServiceCalendarFrame]]`

[sonarcloud[bot]]

Quality Gate failed

Failed conditions
1 Security Hotspot

See analysis details on SonarCloud

[skinkie]

Parsing breaks with Iterable.

        if tokens_factory:
            value = value if collections.is_array(value) else value.split()
            return tokens_factory(
                converter.deserialize(val, types, ns_map=ns_map, format=format)
                for val in value

[skinkie]

Parsing solved.

[skinkie]

While it looked like 'working' again, I noticed that Iterable again breaks the parsing.

[sonarcloud[bot]]

Quality Gate passed

Issues
1 New issue
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud