[FEATURE] release_date support for date_range (#1231)

Adds the ability to toggle whether to use `upload_date` or `release_date` in the date_range plugin, which is used to cull older videos.

```
date_range:
  type: "upload_date"
```

If using the Only Recent preset, it will be automatically applied based on whether your season/episode ordering uses release date or upload date.

Closes https://github.com/jmbannon/ytdl-sub/issues/1182
This commit is contained in:
Jesse Bannon 2025-06-01 09:28:42 -07:00 committed by GitHub
parent 66fd7cf41c
commit 0bd97dbe96
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
12 changed files with 237 additions and 22 deletions

View file

@ -151,6 +151,8 @@ granularity possible.
date_range:
before: "now"
after: "today-2weeks"
breaks: True
type: "upload_date"
``after``
@ -183,6 +185,14 @@ granularity possible.
is enabled or not via Boolean.
``type``
:expected type: Optional[OverridesFormatter]
:description:
Which type of date to use. Must be either ``upload_date`` or ``release_date``.
Defaults to ``upload_date``.
----------------------------------------------------------------------------------------------------
download

View file

@ -1,5 +1,6 @@
from typing import List
from typing import Optional
from typing import Set
from typing import Tuple
from ytdl_sub.config.plugin.plugin import Plugin
@ -7,6 +8,11 @@ from ytdl_sub.config.validators.options import ToggleableOptionsDictValidator
from ytdl_sub.utils.datetime import to_date_str
from ytdl_sub.validators.string_datetime import StringDatetimeValidator
from ytdl_sub.validators.string_formatter_validators import OverridesBooleanFormatterValidator
from ytdl_sub.validators.string_select_validator import OverridesStringSelectValidator
class DateRangeType(OverridesStringSelectValidator):
_select_values: Set[str] = {"upload_date", "release_date"}
class DateRangeOptions(ToggleableOptionsDictValidator):
@ -30,9 +36,11 @@ class DateRangeOptions(ToggleableOptionsDictValidator):
date_range:
before: "now"
after: "today-2weeks"
breaks: True
type: "upload_date"
"""
_optional_keys = {"enable", "before", "after", "breaks"}
_optional_keys = {"enable", "before", "after", "breaks", "type"}
def __init__(self, name, value):
super().__init__(name, value)
@ -41,6 +49,7 @@ class DateRangeOptions(ToggleableOptionsDictValidator):
self._breaks = self._validate_key_if_present(
"breaks", OverridesBooleanFormatterValidator, default="True"
)
self._type = self._validate_key("type", DateRangeType, default="upload_date")
@property
def before(self) -> Optional[StringDatetimeValidator]:
@ -70,6 +79,16 @@ class DateRangeOptions(ToggleableOptionsDictValidator):
"""
return self._breaks
@property
def type(self) -> DateRangeType:
"""
:expected type: Optional[OverridesFormatter]
:description:
Which type of date to use. Must be either ``upload_date`` or ``release_date``.
Defaults to ``upload_date``.
"""
return self._type
class DateRangePlugin(Plugin[DateRangeOptions]):
plugin_options_type = DateRangeOptions
@ -83,17 +102,18 @@ class DateRangePlugin(Plugin[DateRangeOptions]):
match_filters: List[str] = []
breaking_match_filters: List[str] = []
date_type: str = self.overrides.apply_formatter(formatter=self.plugin_options.type)
if self.plugin_options.before:
before_str = to_date_str(
date_validator=self.plugin_options.before, overrides=self.overrides
)
match_filters.append(f"upload_date < {before_str}")
match_filters.append(f"{date_type} < {before_str}")
if self.plugin_options.after:
after_str = to_date_str(
date_validator=self.plugin_options.after, overrides=self.overrides
)
after_filter = f"upload_date >= {after_str}"
after_filter = f"{date_type} >= {after_str}"
if self.overrides.evaluate_boolean(self.plugin_options.breaks):
breaking_match_filters.append(after_filter)
else:

View file

@ -17,6 +17,9 @@ presets:
chapters:
embed_chapters: True
date_range:
type: "{tv_show_date_range_type}"
overrides:
# MUST DEFINE:
# tv_show_directory
@ -45,3 +48,7 @@ presets:
episode_file_path: "{season_directory_name_sanitized}/{episode_file_name_sanitized}"
thumbnail_file_name: "{episode_file_path}-thumb.jpg"
episode_year: "{%slice(episode_date_standardized, 0, 4)}"
# Used to determine which date type to use if `Only Recent` preset
# is applied
tv_show_date_range_type: "upload_date"

View file

@ -117,6 +117,15 @@ presets:
)
}
tv_show_date_range_type: >-
{
%if(
%contains(tv_show_by_date_season_ordering, "release"),
"release_date",
"upload_date"
)
}
##################################### season/episode pairing validation
"%ordering_pair_eq": >-

View file

@ -71,6 +71,16 @@ presets:
)
}
tv_show_date_range_type: >-
{
%if(
%contains(tv_show_collection_episode_ordering, "release"),
"release_date",
"upload_date"
)
}
### LEGACY PRESETS
season_by_collection__episode_by_year_month_day:

View file

@ -349,6 +349,34 @@ class SubscriptionDownload(BaseSubscription, ABC):
return self.download_archive.get_file_handler_transaction_log()
def get_ytdl_options(
self, plugins: Optional[List[Plugin]], dry_run: bool
) -> SubscriptionYTDLOptions:
"""
Parameters
----------
plugins
Optional. If not provided, will reinitialize them
dry_run
Whether its dry run or not
Returns
-------
SubscriptionYTDLOptions
Both metadata and download ytdl-options
"""
if plugins is None:
plugins = self._initialize_plugins()
return SubscriptionYTDLOptions(
preset=self._preset_options,
plugins=plugins,
enhanced_download_archive=self.download_archive,
overrides=self.overrides,
working_directory=self.working_directory,
dry_run=dry_run,
)
def download(self, dry_run: bool = False) -> FileHandlerTransactionLog:
"""
Performs the subscription download
@ -368,14 +396,7 @@ class SubscriptionDownload(BaseSubscription, ABC):
logging.info("Skipping %s", self.name)
return FileHandlerTransactionLog()
subscription_ytdl_options = SubscriptionYTDLOptions(
preset=self._preset_options,
plugins=plugins,
enhanced_download_archive=self.download_archive,
overrides=self.overrides,
working_directory=self.working_directory,
dry_run=dry_run,
)
subscription_ytdl_options = self.get_ytdl_options(plugins=plugins, dry_run=dry_run)
downloader = MultiUrlDownloader(
options=self.downloader_options,
@ -420,15 +441,7 @@ class SubscriptionDownload(BaseSubscription, ABC):
self.download_archive.reinitialize(dry_run=dry_run)
plugins = self._initialize_plugins()
subscription_ytdl_options = SubscriptionYTDLOptions(
preset=self._preset_options,
plugins=plugins,
enhanced_download_archive=self.download_archive,
overrides=self.overrides,
working_directory=self.working_directory,
dry_run=dry_run,
)
subscription_ytdl_options = self.get_ytdl_options(plugins=plugins, dry_run=dry_run)
# Re-add the original downloader class' plugins
plugins.extend(

View file

@ -1,5 +1,6 @@
from typing import Set
from ytdl_sub.validators.string_formatter_validators import OverridesStringFormatterValidator
from ytdl_sub.validators.validators import StringValidator
@ -17,3 +18,17 @@ class StringSelectValidator(StringValidator):
raise self._validation_exception(
f"Must be one of the following values: {', '.join(self._select_values)}"
)
class OverridesStringSelectValidator(OverridesStringFormatterValidator):
_expected_value_type_name = "overrides select"
_select_values: Set[str] = set()
def post_process(self, resolved: str) -> str:
if resolved not in self._select_values:
raise self._validation_exception(
f"Must be one of the following values: {', '.join(sorted(self._select_values))}"
)
return resolved

View file

@ -12,6 +12,7 @@ from typing import Callable
from typing import Dict
from typing import List
from typing import Optional
from typing import Tuple
from unittest.mock import patch
import pytest
@ -272,3 +273,27 @@ def mock_run_from_cli(args: str) -> List[Subscription]:
args_list = ["ytdl-sub"] + shlex.split(args)
with patch.object(sys, "argv", args_list):
return main()
def get_match_filters(
subscription: Subscription, dry_run: bool, download_filters: bool
) -> Tuple[List[str], List[str]]:
"""
Util function to get match filters from a subscription.
Returns
-------
match_filters, breaking_match_filters
"""
options = subscription.get_ytdl_options(plugins=None, dry_run=dry_run)
options_dict = (
options.download_builder().to_dict()
if download_filters
else options.metadata_builder().to_dict()
)
if "match_filter" not in options_dict:
return [], []
match_filter_str = repr(options_dict["match_filter"])
out = eval(match_filter_str.split("(", maxsplit=1)[-1].split(")")[0])
return out

View file

@ -1,6 +1,5 @@
import contextlib
import os
import shutil
from pathlib import Path
from typing import Callable
from typing import Dict

View file

@ -4,7 +4,6 @@ import pytest
from expected_download import assert_expected_downloads
from expected_transaction_log import assert_transaction_log_matches
from ytdl_sub.entries.entry import ytdl_sub_chapters_from_comments
from ytdl_sub.subscriptions.subscription import Subscription

View file

View file

@ -0,0 +1,108 @@
import re
from typing import Any
from typing import Dict
import pytest
from conftest import get_match_filters
from ytdl_sub.config.config_file import ConfigFile
from ytdl_sub.subscriptions.subscription import Subscription
from ytdl_sub.utils.exceptions import ValidationException
@pytest.fixture
def preset_dict(output_directory) -> Dict[str, Any]:
return {
"download": "https://your.name.here",
"output_options": {"output_directory": output_directory, "file_name": "will_error.mp4"},
}
class TestDateRange:
@pytest.mark.parametrize("date_range_type", ["upload_date", "release_date"])
def test_date_range_type(
self,
default_config: ConfigFile,
preset_dict: Dict[str, Any],
output_directory: str,
date_range_type: str,
):
preset_dict["date_range"] = {
"before": "20250530",
"after": "20250510",
"type": date_range_type,
}
sub = Subscription.from_dict(
config=default_config,
preset_name="test_date_range",
preset_dict=preset_dict,
)
metadata_filter, metadata_breaking_filter = get_match_filters(
subscription=sub, dry_run=False, download_filters=False
)
assert metadata_filter == [
f"!is_live & !is_upcoming & !post_live & {date_range_type} < 20250530"
]
assert metadata_breaking_filter == [f"{date_range_type} >= 20250510"]
download_filter, download_breaking_filter = get_match_filters(
subscription=sub, dry_run=False, download_filters=True
)
assert not download_filter
assert not download_breaking_filter
def test_date_range_breaks_false(
self,
default_config: ConfigFile,
preset_dict: Dict[str, Any],
output_directory: str,
):
preset_dict["date_range"] = {
"before": "20250530",
"after": "20250510",
"breaks": False,
}
sub = Subscription.from_dict(
config=default_config,
preset_name="test_date_range",
preset_dict=preset_dict,
)
metadata_filter, metadata_breaking_filter = get_match_filters(
subscription=sub, dry_run=False, download_filters=False
)
assert metadata_filter == [
f"!is_live & !is_upcoming & !post_live & upload_date < 20250530 & upload_date >= 20250510"
]
assert not metadata_breaking_filter
download_filter, download_breaking_filter = get_match_filters(
subscription=sub, dry_run=False, download_filters=True
)
assert not download_filter
assert not download_breaking_filter
def test_date_range_invalid_type(
self,
default_config: ConfigFile,
preset_dict: Dict[str, Any],
output_directory: str,
):
preset_dict["date_range"] = {
"before": "20250530",
"after": "20250510",
"type": "no",
}
error_msg = (
"Validation error in test_date_range.date_range.type: "
"Must be one of the following values: release_date, upload_date"
)
with pytest.raises(ValidationException, match=re.escape(error_msg)):
Subscription.from_dict(
config=default_config,
preset_name="test_date_range",
preset_dict=preset_dict,
).download(dry_run=False)