mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-17 00:24:26 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

refactor: All parsers return snapshot instead of link

Browse source
This commit is contained in:
Cristian 2021-01-04 09:31:14 -05:00
parent 15d88be229
commit b8efaa5b6a
11 changed files with 72 additions and 53 deletions
Download patch file Download diff file Expand all files Collapse all files

12
archivebox/parsers/generic_json.py
View file

@ -5,7 +5,8 @@ import json
from typing import IO, Iterable
from datetime import datetime
from ..index.schema import Link
from django.db.models import Model
from ..util import (
htmldecode,
enforce_types,
@ -13,8 +14,9 @@ from ..util import (
@enforce_types
def parse_generic_json_export(json_file: IO[str], **_kwargs) -> Iterable[Link]:
def parse_generic_json_export(json_file: IO[str], **_kwargs) -> Iterable[Model]:
"""Parse JSON-format bookmarks export files (produced by pinboard.in/export/, or wallabag)"""
from core.models import Snapshot
json_file.seek(0)
links = json.load(json_file)
@ -56,10 +58,10 @@ def parse_generic_json_export(json_file: IO[str], **_kwargs) -> Iterable[Link]:
elif link.get('name'):
title = link['name'].strip()
yield Link(
yield Snapshot(
url=htmldecode(url),
timestamp=ts_str,
title=htmldecode(title) or None,
tags=htmldecode(link.get('tags')) or '',
sources=[json_file.name],
#tags=htmldecode(link.get('tags')) or '',
#sources=[json_file.name],
)