mirror of
https://github.com/ArchiveBox/ArchiveBox.git
synced 2025-05-15 07:34:27 -04:00
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
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:
parent
abf75f49f4
commit
01ba6d49d3
115 changed files with 2466 additions and 2301 deletions
46
archivebox/plugins_extractor/mercury/__init__.py
Normal file
46
archivebox/plugins_extractor/mercury/__init__.py
Normal file
|
@ -0,0 +1,46 @@
|
|||
__package__ = 'plugins_extractor.mercury'
|
||||
__label__ = 'mercury'
|
||||
__version__ = '2024.10.14'
|
||||
__author__ = 'Nick Sweeting'
|
||||
__homepage__ = 'https://github.com/postlight/mercury-parser'
|
||||
__dependencies__ = ['npm']
|
||||
|
||||
import abx
|
||||
|
||||
|
||||
@abx.hookimpl
|
||||
def get_PLUGIN():
|
||||
return {
|
||||
'mercury': {
|
||||
'PACKAGE': __package__,
|
||||
'LABEL': __label__,
|
||||
'VERSION': __version__,
|
||||
'AUTHOR': __author__,
|
||||
'HOMEPAGE': __homepage__,
|
||||
'DEPENDENCIES': __dependencies__,
|
||||
}
|
||||
}
|
||||
|
||||
@abx.hookimpl
|
||||
def get_CONFIG():
|
||||
from .config import MERCURY_CONFIG
|
||||
|
||||
return {
|
||||
'mercury': MERCURY_CONFIG
|
||||
}
|
||||
|
||||
@abx.hookimpl
|
||||
def get_BINARIES():
|
||||
from .binaries import MERCURY_BINARY
|
||||
|
||||
return {
|
||||
'mercury': MERCURY_BINARY,
|
||||
}
|
||||
|
||||
@abx.hookimpl
|
||||
def get_EXTRACTORS():
|
||||
from .extractors import MERCURY_EXTRACTOR
|
||||
|
||||
return {
|
||||
'mercury': MERCURY_EXTRACTOR,
|
||||
}
|
|
@ -1,80 +0,0 @@
|
|||
__package__ = 'plugins_extractor.mercury'
|
||||
|
||||
from typing import List, Optional
|
||||
from pathlib import Path
|
||||
|
||||
from pydantic import InstanceOf, Field
|
||||
from pydantic_pkgr import BinProvider, BinName, BinaryOverrides, bin_abspath
|
||||
|
||||
from abx.archivebox.base_plugin import BasePlugin, BaseHook
|
||||
from abx.archivebox.base_configset import BaseConfigSet
|
||||
from abx.archivebox.base_binary import BaseBinary, env
|
||||
from abx.archivebox.base_extractor import BaseExtractor, ExtractorName
|
||||
|
||||
from archivebox.config.common import ARCHIVING_CONFIG, STORAGE_CONFIG
|
||||
from archivebox.plugins_pkg.npm.apps import SYS_NPM_BINPROVIDER, LIB_NPM_BINPROVIDER
|
||||
|
||||
class MercuryConfig(BaseConfigSet):
|
||||
|
||||
SAVE_MERCURY: bool = Field(default=True, alias='USE_MERCURY')
|
||||
|
||||
MERCURY_BINARY: str = Field(default='postlight-parser')
|
||||
MERCURY_EXTRA_ARGS: List[str] = []
|
||||
|
||||
SAVE_MERCURY_REQUISITES: bool = Field(default=True)
|
||||
MERCURY_RESTRICT_FILE_NAMES: str = Field(default=lambda: STORAGE_CONFIG.RESTRICT_FILE_NAMES)
|
||||
|
||||
MERCURY_TIMEOUT: int = Field(default=lambda: ARCHIVING_CONFIG.TIMEOUT)
|
||||
MERCURY_CHECK_SSL_VALIDITY: bool = Field(default=lambda: ARCHIVING_CONFIG.CHECK_SSL_VALIDITY)
|
||||
MERCURY_USER_AGENT: str = Field(default=lambda: ARCHIVING_CONFIG.USER_AGENT)
|
||||
MERCURY_COOKIES_FILE: Optional[Path] = Field(default=lambda: ARCHIVING_CONFIG.COOKIES_FILE)
|
||||
|
||||
|
||||
|
||||
MERCURY_CONFIG = MercuryConfig()
|
||||
|
||||
|
||||
class MercuryBinary(BaseBinary):
|
||||
name: BinName = MERCURY_CONFIG.MERCURY_BINARY
|
||||
binproviders_supported: List[InstanceOf[BinProvider]] = [LIB_NPM_BINPROVIDER, SYS_NPM_BINPROVIDER, env]
|
||||
|
||||
overrides: BinaryOverrides = {
|
||||
LIB_NPM_BINPROVIDER.name: {
|
||||
'packages': ['@postlight/parser@^2.2.3'],
|
||||
},
|
||||
SYS_NPM_BINPROVIDER.name: {
|
||||
'packages': ['@postlight/parser@^2.2.3'],
|
||||
'install': lambda: None, # never try to install things into global prefix
|
||||
},
|
||||
env.name: {
|
||||
'version': lambda: '999.999.999' if bin_abspath('postlight-parser', PATH=env.PATH) else None,
|
||||
},
|
||||
}
|
||||
|
||||
MERCURY_BINARY = MercuryBinary()
|
||||
|
||||
|
||||
class MercuryExtractor(BaseExtractor):
|
||||
name: ExtractorName = 'mercury'
|
||||
binary: str = MERCURY_BINARY.name
|
||||
|
||||
def get_output_path(self, snapshot) -> Path | None:
|
||||
return snapshot.link_dir / 'mercury' / 'content.html'
|
||||
|
||||
MERCURY_EXTRACTOR = MercuryExtractor()
|
||||
|
||||
|
||||
|
||||
class MercuryPlugin(BasePlugin):
|
||||
app_label: str = 'mercury'
|
||||
verbose_name: str = 'MERCURY'
|
||||
|
||||
hooks: List[InstanceOf[BaseHook]] = [
|
||||
MERCURY_CONFIG,
|
||||
MERCURY_BINARY,
|
||||
MERCURY_EXTRACTOR,
|
||||
]
|
||||
|
||||
|
||||
PLUGIN = MercuryPlugin()
|
||||
DJANGO_APP = PLUGIN.AppConfig
|
32
archivebox/plugins_extractor/mercury/binaries.py
Normal file
32
archivebox/plugins_extractor/mercury/binaries.py
Normal file
|
@ -0,0 +1,32 @@
|
|||
__package__ = 'plugins_extractor.mercury'
|
||||
|
||||
from typing import List
|
||||
|
||||
from pydantic import InstanceOf
|
||||
from pydantic_pkgr import BinProvider, BinName, BinaryOverrides, bin_abspath
|
||||
|
||||
from abx.archivebox.base_binary import BaseBinary, env
|
||||
|
||||
from archivebox.plugins_pkg.npm.binproviders import SYS_NPM_BINPROVIDER, LIB_NPM_BINPROVIDER
|
||||
|
||||
from .config import MERCURY_CONFIG
|
||||
|
||||
|
||||
class MercuryBinary(BaseBinary):
|
||||
name: BinName = MERCURY_CONFIG.MERCURY_BINARY
|
||||
binproviders_supported: List[InstanceOf[BinProvider]] = [LIB_NPM_BINPROVIDER, SYS_NPM_BINPROVIDER, env]
|
||||
|
||||
overrides: BinaryOverrides = {
|
||||
LIB_NPM_BINPROVIDER.name: {
|
||||
'packages': ['@postlight/parser@^2.2.3'],
|
||||
},
|
||||
SYS_NPM_BINPROVIDER.name: {
|
||||
'packages': ['@postlight/parser@^2.2.3'],
|
||||
'install': lambda: None, # never try to install things into global prefix
|
||||
},
|
||||
env.name: {
|
||||
'version': lambda: '999.999.999' if bin_abspath('postlight-parser', PATH=env.PATH) else None,
|
||||
},
|
||||
}
|
||||
|
||||
MERCURY_BINARY = MercuryBinary()
|
31
archivebox/plugins_extractor/mercury/config.py
Normal file
31
archivebox/plugins_extractor/mercury/config.py
Normal file
|
@ -0,0 +1,31 @@
|
|||
__package__ = 'plugins_extractor.mercury'
|
||||
|
||||
from typing import List, Optional
|
||||
from pathlib import Path
|
||||
|
||||
from pydantic import Field
|
||||
|
||||
from abx.archivebox.base_configset import BaseConfigSet
|
||||
|
||||
from archivebox.config.common import ARCHIVING_CONFIG, STORAGE_CONFIG
|
||||
|
||||
|
||||
|
||||
class MercuryConfig(BaseConfigSet):
|
||||
|
||||
SAVE_MERCURY: bool = Field(default=True, alias='USE_MERCURY')
|
||||
|
||||
MERCURY_BINARY: str = Field(default='postlight-parser')
|
||||
MERCURY_EXTRA_ARGS: List[str] = []
|
||||
|
||||
SAVE_MERCURY_REQUISITES: bool = Field(default=True)
|
||||
MERCURY_RESTRICT_FILE_NAMES: str = Field(default=lambda: STORAGE_CONFIG.RESTRICT_FILE_NAMES)
|
||||
|
||||
MERCURY_TIMEOUT: int = Field(default=lambda: ARCHIVING_CONFIG.TIMEOUT)
|
||||
MERCURY_CHECK_SSL_VALIDITY: bool = Field(default=lambda: ARCHIVING_CONFIG.CHECK_SSL_VALIDITY)
|
||||
MERCURY_USER_AGENT: str = Field(default=lambda: ARCHIVING_CONFIG.USER_AGENT)
|
||||
MERCURY_COOKIES_FILE: Optional[Path] = Field(default=lambda: ARCHIVING_CONFIG.COOKIES_FILE)
|
||||
|
||||
|
||||
|
||||
MERCURY_CONFIG = MercuryConfig()
|
19
archivebox/plugins_extractor/mercury/extractors.py
Normal file
19
archivebox/plugins_extractor/mercury/extractors.py
Normal file
|
@ -0,0 +1,19 @@
|
|||
__package__ = 'plugins_extractor.mercury'
|
||||
|
||||
from pathlib import Path
|
||||
|
||||
from abx.archivebox.base_extractor import BaseExtractor, ExtractorName
|
||||
|
||||
from .binaries import MERCURY_BINARY
|
||||
|
||||
|
||||
|
||||
class MercuryExtractor(BaseExtractor):
|
||||
name: ExtractorName = 'mercury'
|
||||
binary: str = MERCURY_BINARY.name
|
||||
|
||||
def get_output_path(self, snapshot) -> Path | None:
|
||||
return snapshot.link_dir / 'mercury' / 'content.html'
|
||||
|
||||
|
||||
MERCURY_EXTRACTOR = MercuryExtractor()
|
Loading…
Add table
Add a link
Reference in a new issue