mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-17 08:34:26 -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/medium_rss.py
View file

@ -4,9 +4,10 @@ __package__ = 'archivebox.parsers'
from typing import IO, Iterable
from datetime import datetime
from django.db.models import Model
from xml.etree import ElementTree
from ..index.schema import Link
from ..util import (
htmldecode,
enforce_types,
@ -14,8 +15,9 @@ from ..util import (
@enforce_types
def parse_medium_rss_export(rss_file: IO[str], **_kwargs) -> Iterable[Link]:
def parse_medium_rss_export(rss_file: IO[str], **_kwargs) -> Iterable[Model]:
"""Parse Medium RSS feed files into links"""
from core.models import Snapshot
rss_file.seek(0)
root = ElementTree.parse(rss_file).getroot()
@ -26,10 +28,10 @@ def parse_medium_rss_export(rss_file: IO[str], **_kwargs) -> Iterable[Link]:
ts_str = item.find("pubDate").text # type: ignore
time = datetime.strptime(ts_str, "%a, %d %b %Y %H:%M:%S %Z") # type: ignore
yield Link(
yield Snapshot(
url=htmldecode(url),
timestamp=str(time.timestamp()),
title=htmldecode(title) or None,
tags=None,
sources=[rss_file.name],
#tags=None,
#sources=[rss_file.name],
)