mirror of
https://github.com/bellingcat/auto-archiver.git
synced 2026-06-13 05:38:29 +03:00
url auth wall detect
This commit is contained in:
@@ -3,6 +3,7 @@ from dataclasses import dataclass
|
||||
import mimetypes, uuid, os, pathlib
|
||||
from jinja2 import Environment, FileSystemLoader
|
||||
from urllib.parse import quote
|
||||
from loguru import logger
|
||||
|
||||
from ..version import __version__
|
||||
from ..core import Metadata, Media
|
||||
@@ -26,12 +27,17 @@ class HtmlFormatter(Formatter):
|
||||
@staticmethod
|
||||
def configs() -> dict:
|
||||
return {
|
||||
"detect_thumbnails": {"default": True, "help": "if true will group by thumbnails generated by thumbnail enricher by id 'thumbnail_00'"},
|
||||
|
||||
"detect_thumbnails": {"default": True, "help": "if true will group by thumbnails generated by thumbnail enricher by id 'thumbnail_00'"}
|
||||
}
|
||||
|
||||
def format(self, item: Metadata) -> Media:
|
||||
url = item.get_url()
|
||||
if item.is_empty():
|
||||
logger.debug(f"[SKIP] FORMAT there is no media or metadata to format: {url=}")
|
||||
return
|
||||
|
||||
content = self.template.render(
|
||||
url=item.get_url(),
|
||||
url=url,
|
||||
title=item.get_title(),
|
||||
media=item.media,
|
||||
metadata=item.get_clean_metadata(),
|
||||
|
||||
Reference in New Issue
Block a user