This commit is contained in:
Nick Sweeting 2024-10-25 01:06:12 -07:00
parent 4b6f08b0fe
commit 5d9a32c364
No known key found for this signature in database
178 changed files with 2982 additions and 1322 deletions

View file

@ -0,0 +1,51 @@
__package__ = 'plugins_extractor.singlefile'
__label__ = 'singlefile'
__version__ = '2024.10.14'
__author__ = 'ArchiveBox'
__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

@ -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

@ -0,0 +1,14 @@
from django.db import models
from core.models import ArchiveResult
class SinglefileResultManager(models.Manager):
def get_queryset(self):
return super().get_queryset().filter(extractor='singlefile')
class SinglefileResult(ArchiveResult):
objects = SinglefileResultManager()
class Meta:
proxy = True

View file

@ -0,0 +1,7 @@
[project]
name = "abx-singlefile-extractor"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.10"
dependencies = []