new vastly simplified plugin spec without pydantic
Some checks are pending
Build Debian package / build (push) Waiting to run
Build Docker image / buildx (push) Waiting to run
Build Homebrew package / build (push) Waiting to run
Run linters / lint (push) Waiting to run
Build Pip package / build (push) Waiting to run
Run tests / python_tests (ubuntu-22.04, 3.11) (push) Waiting to run
Run tests / docker_tests (push) Waiting to run

This commit is contained in:
Nick Sweeting 2024-10-14 21:50:47 -07:00
parent abf75f49f4
commit 01ba6d49d3
No known key found for this signature in database
115 changed files with 2466 additions and 2301 deletions

View file

@ -0,0 +1,51 @@
__package__ = 'plugins_extractor.singlefile'
__label__ = 'singlefile'
__version__ = '2024.10.14'
__author__ = 'Nick Sweeting'
__homepage__ = 'https://github.com/gildas-lormeau/singlefile'
__dependencies__ = ['npm']
import abx
@abx.hookimpl
def get_PLUGIN():
return {
'singlefile': {
'PACKAGE': __package__,
'LABEL': __label__,
'VERSION': __version__,
'AUTHOR': __author__,
'HOMEPAGE': __homepage__,
'DEPENDENCIES': __dependencies__,
}
}
@abx.hookimpl
def get_CONFIG():
from .config import SINGLEFILE_CONFIG
return {
'singlefile': SINGLEFILE_CONFIG
}
@abx.hookimpl
def get_BINARIES():
from .binaries import SINGLEFILE_BINARY
return {
'singlefile': SINGLEFILE_BINARY,
}
@abx.hookimpl
def get_EXTRACTORS():
from .extractors import SINGLEFILE_EXTRACTOR
return {
'singlefile': SINGLEFILE_EXTRACTOR,
}
# @abx.hookimpl
# def get_INSTALLED_APPS():
# # needed to load ./models.py
# return [__package__]

View file

@ -1,110 +0,0 @@
__package__ = 'archivebox.plugins_extractor.singlefile'
from pathlib import Path
from typing import List, Optional
# from typing_extensions import Self
# Depends on other PyPI/vendor packages:
from pydantic import InstanceOf, Field
from pydantic_pkgr import BinProvider, BinaryOverrides, BinName, bin_abspath
# Depends on other Django apps:
from abx.archivebox.base_plugin import BasePlugin
from abx.archivebox.base_configset import BaseConfigSet
from abx.archivebox.base_binary import BaseBinary, env
from abx.archivebox.base_extractor import BaseExtractor
from abx.archivebox.base_queue import BaseQueue
from abx.archivebox.base_hook import BaseHook
# Depends on Other Plugins:
from archivebox.config.common import ARCHIVING_CONFIG
from plugins_pkg.npm.apps import SYS_NPM_BINPROVIDER, LIB_NPM_BINPROVIDER
###################### Config ##########################
class SinglefileConfig(BaseConfigSet):
SAVE_SINGLEFILE: bool = True
SINGLEFILE_USER_AGENT: str = Field(default=lambda: ARCHIVING_CONFIG.USER_AGENT)
SINGLEFILE_TIMEOUT: int = Field(default=lambda: ARCHIVING_CONFIG.TIMEOUT)
SINGLEFILE_CHECK_SSL_VALIDITY: bool = Field(default=lambda: ARCHIVING_CONFIG.CHECK_SSL_VALIDITY)
SINGLEFILE_COOKIES_FILE: Optional[Path] = Field(default=lambda: ARCHIVING_CONFIG.COOKIES_FILE)
SINGLEFILE_BINARY: str = Field(default='single-file')
SINGLEFILE_EXTRA_ARGS: List[str] = []
SINGLEFILE_CONFIG = SinglefileConfig()
SINGLEFILE_MIN_VERSION = '1.1.54'
SINGLEFILE_MAX_VERSION = '1.1.60'
class SinglefileBinary(BaseBinary):
name: BinName = SINGLEFILE_CONFIG.SINGLEFILE_BINARY
binproviders_supported: List[InstanceOf[BinProvider]] = [LIB_NPM_BINPROVIDER, SYS_NPM_BINPROVIDER, env]
overrides: BinaryOverrides = {
LIB_NPM_BINPROVIDER.name: {
"abspath": lambda:
bin_abspath(SINGLEFILE_CONFIG.SINGLEFILE_BINARY, PATH=LIB_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file", PATH=LIB_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file-node.js", PATH=LIB_NPM_BINPROVIDER.PATH),
"packages": [f"single-file-cli@>={SINGLEFILE_MIN_VERSION} <{SINGLEFILE_MAX_VERSION}"],
},
SYS_NPM_BINPROVIDER.name: {
"abspath": lambda:
bin_abspath(SINGLEFILE_CONFIG.SINGLEFILE_BINARY, PATH=SYS_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file", PATH=SYS_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file-node.js", PATH=SYS_NPM_BINPROVIDER.PATH),
"packages": [f"single-file-cli@>={SINGLEFILE_MIN_VERSION} <{SINGLEFILE_MAX_VERSION}"],
"install": lambda: None,
},
env.name: {
'abspath': lambda:
bin_abspath(SINGLEFILE_CONFIG.SINGLEFILE_BINARY, PATH=env.PATH)
or bin_abspath('single-file', PATH=env.PATH)
or bin_abspath('single-file-node.js', PATH=env.PATH),
},
}
SINGLEFILE_BINARY = SinglefileBinary()
PLUGIN_BINARIES = [SINGLEFILE_BINARY]
class SinglefileExtractor(BaseExtractor):
name: str = 'singlefile'
binary: BinName = SINGLEFILE_BINARY.name
def get_output_path(self, snapshot) -> Path:
return Path(snapshot.link_dir) / 'singlefile.html'
SINGLEFILE_BINARY = SinglefileBinary()
SINGLEFILE_EXTRACTOR = SinglefileExtractor()
class SinglefileQueue(BaseQueue):
name: str = 'singlefile'
binaries: List[InstanceOf[BaseBinary]] = [SINGLEFILE_BINARY]
SINGLEFILE_QUEUE = SinglefileQueue()
class SinglefilePlugin(BasePlugin):
app_label: str ='singlefile'
verbose_name: str = 'SingleFile'
hooks: List[InstanceOf[BaseHook]] = [
SINGLEFILE_CONFIG,
SINGLEFILE_BINARY,
SINGLEFILE_EXTRACTOR,
SINGLEFILE_QUEUE,
]
PLUGIN = SinglefilePlugin()
# PLUGIN.register(settings)
DJANGO_APP = PLUGIN.AppConfig

View file

@ -0,0 +1,48 @@
__package__ = 'plugins_extractor.singlefile'
from typing import List
from pydantic import InstanceOf
from pydantic_pkgr import BinProvider, BinaryOverrides, BinName, bin_abspath
from abx.archivebox.base_binary import BaseBinary, env
from plugins_pkg.npm.binproviders import SYS_NPM_BINPROVIDER, LIB_NPM_BINPROVIDER
from .config import SINGLEFILE_CONFIG
SINGLEFILE_MIN_VERSION = '1.1.54'
SINGLEFILE_MAX_VERSION = '1.1.60'
class SinglefileBinary(BaseBinary):
name: BinName = SINGLEFILE_CONFIG.SINGLEFILE_BINARY
binproviders_supported: List[InstanceOf[BinProvider]] = [LIB_NPM_BINPROVIDER, SYS_NPM_BINPROVIDER, env]
overrides: BinaryOverrides = {
LIB_NPM_BINPROVIDER.name: {
"abspath": lambda:
bin_abspath(SINGLEFILE_CONFIG.SINGLEFILE_BINARY, PATH=LIB_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file", PATH=LIB_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file-node.js", PATH=LIB_NPM_BINPROVIDER.PATH),
"packages": [f"single-file-cli@>={SINGLEFILE_MIN_VERSION} <{SINGLEFILE_MAX_VERSION}"],
},
SYS_NPM_BINPROVIDER.name: {
"abspath": lambda:
bin_abspath(SINGLEFILE_CONFIG.SINGLEFILE_BINARY, PATH=SYS_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file", PATH=SYS_NPM_BINPROVIDER.PATH)
or bin_abspath("single-file-node.js", PATH=SYS_NPM_BINPROVIDER.PATH),
"packages": [f"single-file-cli@>={SINGLEFILE_MIN_VERSION} <{SINGLEFILE_MAX_VERSION}"],
"install": lambda: None,
},
env.name: {
'abspath': lambda:
bin_abspath(SINGLEFILE_CONFIG.SINGLEFILE_BINARY, PATH=env.PATH)
or bin_abspath('single-file', PATH=env.PATH)
or bin_abspath('single-file-node.js', PATH=env.PATH),
},
}
SINGLEFILE_BINARY = SinglefileBinary()

View file

@ -0,0 +1,25 @@
__package__ = 'plugins_extractor.singlefile'
from pathlib import Path
from typing import List, Optional
from pydantic import Field
from abx.archivebox.base_configset import BaseConfigSet
from archivebox.config.common import ARCHIVING_CONFIG
class SinglefileConfig(BaseConfigSet):
SAVE_SINGLEFILE: bool = True
SINGLEFILE_USER_AGENT: str = Field(default=lambda: ARCHIVING_CONFIG.USER_AGENT)
SINGLEFILE_TIMEOUT: int = Field(default=lambda: ARCHIVING_CONFIG.TIMEOUT)
SINGLEFILE_CHECK_SSL_VALIDITY: bool = Field(default=lambda: ARCHIVING_CONFIG.CHECK_SSL_VALIDITY)
SINGLEFILE_COOKIES_FILE: Optional[Path] = Field(default=lambda: ARCHIVING_CONFIG.COOKIES_FILE)
SINGLEFILE_BINARY: str = Field(default='single-file')
SINGLEFILE_EXTRA_ARGS: List[str] = []
SINGLEFILE_CONFIG = SinglefileConfig()

View file

@ -0,0 +1,19 @@
__package__ = 'plugins_extractor.singlefile'
from pathlib import Path
from pydantic_pkgr import BinName
from abx.archivebox.base_extractor import BaseExtractor
from .binaries import SINGLEFILE_BINARY
class SinglefileExtractor(BaseExtractor):
name: str = 'singlefile'
binary: BinName = SINGLEFILE_BINARY.name
def get_output_path(self, snapshot) -> Path:
return Path(snapshot.link_dir) / 'singlefile.html'
SINGLEFILE_EXTRACTOR = SinglefileExtractor()

View file

@ -1,26 +0,0 @@
# Generated by Django 5.1.1 on 2024-09-10 05:05
from django.db import migrations
class Migration(migrations.Migration):
initial = True
dependencies = [
('core', '0074_alter_snapshot_downloaded_at'),
]
operations = [
migrations.CreateModel(
name='SinglefileResult',
fields=[
],
options={
'proxy': True,
'indexes': [],
'constraints': [],
},
bases=('core.archiveresult',),
),
]

View file

@ -1,40 +0,0 @@
__package__ = 'archivebox.queues'
import time
from django.core.cache import cache
from huey import crontab
from django_huey import db_task, on_startup, db_periodic_task
from huey_monitor.models import TaskModel
from huey_monitor.tqdm import ProcessInfo
@db_task(queue="singlefile", context=True)
def extract(url, out_dir, config, task=None, parent_task_id=None):
if task and parent_task_id:
TaskModel.objects.set_parent_task(main_task_id=parent_task_id, sub_task_id=task.id)
process_info = ProcessInfo(task, desc="extract_singlefile", parent_task_id=parent_task_id, total=1)
time.sleep(5)
process_info.update(n=1)
return {'output': 'singlefile.html', 'status': 'succeeded'}
# @on_startup(queue='singlefile')
# def start_singlefile_queue():
# print("[+] Starting singlefile worker...")
# update_version.call_local()
# @db_periodic_task(crontab(minute='*/5'), queue='singlefile')
# def update_version():
# print('[*] Updating singlefile version... 5 minute interval')
# from django.conf import settings
# bin = settings.BINARIES.SinglefileBinary.load()
# if bin.version:
# cache.set(f"bin:abspath:{bin.name}", bin.abspath)
# cache.set(f"bin:version:{bin.name}:{bin.abspath}", bin.version)
# print('[√] Updated singlefile version:', bin.version, bin.abspath)