CodeMirror/ArchiveBox
1
0
Fork 0
mirror of synced 2024-06-28 11:00:35 +12:00
Code Issues Projects Releases Wiki Activity

refactor: mercury uses snapshot instead of link

Browse source
This commit is contained in:
Cristian 2020-12-29 14:01:04 -05:00
parent 756b7fc76d
commit 48aa9d5c08
1 changed files with 10 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

18
archivebox/extractors/mercury.py
View file

@ -6,6 +6,8 @@ from subprocess import CompletedProcess
from typing import Optional, List from typing import Optional, List
import json import json
from django.db.models import Model
from ..index.schema import Link, ArchiveResult, ArchiveError from ..index.schema import Link, ArchiveResult, ArchiveError
from ..system import run, atomic_write from ..system import run, atomic_write
from ..util import ( from ..util import (
@ -37,20 +39,20 @@ def ShellError(cmd: List[str], result: CompletedProcess, lines: int=20) -> Archi
@enforce_types @enforce_types
def should_save_mercury(link: Link, out_dir: Optional[str]=None) -> bool: def should_save_mercury(snapshot: Model, out_dir: Optional[str]=None) -> bool:
out_dir = out_dir or link.link_dir out_dir = out_dir or snapshot.snapshot_dir
if is_static_file(link.url): if is_static_file(snapshot.url):
return False return False
output = Path(out_dir or link.link_dir) / 'mercury' output = Path(out_dir or snapshot.snapshot_dir) / 'mercury'
return SAVE_MERCURY and MERCURY_VERSION and (not output.exists()) return SAVE_MERCURY and MERCURY_VERSION and (not output.exists())
@enforce_types @enforce_types
def save_mercury(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT) -> ArchiveResult: def save_mercury(snapshot: Model, out_dir: Optional[Path]=None, timeout: int=TIMEOUT) -> ArchiveResult:
"""download reader friendly version using @postlight/mercury-parser""" """download reader friendly version using @postlight/mercury-parser"""
out_dir = Path(out_dir or link.link_dir) out_dir = Path(out_dir or snapshot.snapshot_dir)
output_folder = out_dir.absolute() / "mercury" output_folder = out_dir.absolute() / "mercury"
output = str(output_folder) output = str(output_folder)
@ -60,7 +62,7 @@ def save_mercury(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT)
# Get plain text version of article # Get plain text version of article
cmd = [ cmd = [
DEPENDENCIES['MERCURY_BINARY']['path'], DEPENDENCIES['MERCURY_BINARY']['path'],
link.url, snapshot.url,
"--format=text" "--format=text"
] ]
result = run(cmd, cwd=out_dir, timeout=timeout) result = run(cmd, cwd=out_dir, timeout=timeout)
@ -72,7 +74,7 @@ def save_mercury(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT)
# Get HTML version of article # Get HTML version of article
cmd = [ cmd = [
DEPENDENCIES['MERCURY_BINARY']['path'], DEPENDENCIES['MERCURY_BINARY']['path'],
link.url snapshot.url
] ]
result = run(cmd, cwd=out_dir, timeout=timeout) result = run(cmd, cwd=out_dir, timeout=timeout)
try: try: