wallabag-tools/wallabag_mass_import.py

86 lines
2.3 KiB
Python
Raw Normal View History

2020-03-12 22:15:20 +00:00
import asyncio
import configparser
import logging
import json
import os
from typing import Any, Dict, List
import aiohttp
from wallabag_api.wallabag import Wallabag
logging.basicConfig(level=logging.DEBUG)
ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
CONFIG_PATH = os.path.join(ROOT_DIR, "config.ini")
ARTICLES_PATH = os.path.join(ROOT_DIR, "articles.json")
def load_configuration(path):
config = configparser.ConfigParser()
config.read(path)
return config["Wallabag Configuration"]
async def connect_to_wallabag(
configuration: Dict[str, str], session: aiohttp.ClientSession
) -> Wallabag:
configuration["access_token"] = await Wallabag.get_token(**configuration)
return Wallabag(
host=configuration["host"],
token=configuration["access_token"],
client_id=configuration["client_id"],
client_secret=configuration["client_secret"],
aio_sess=session,
)
def read_urls_file(f):
while True:
line = f.readline()
if line:
yield line
else:
break
async def import_articles(w_api: Wallabag, path: str, limit=0):
entries_id = []
with open(path) as f:
for index, item in enumerate(json.load(f)):
if limit > 0 and index == limit:
break
entry = await w_api.post_entries(
url=item["url"],
archive=item["is_archived"],
starred=item["is_starred"],
original_url=item["origin_url"],
tags=",".join(item["tags"]),
)
logging.info(
f"Entry url {item['url']} posted to wallabag with id {entry['id']}"
)
entries_id.append(entry["id"])
return entries_id
async def delete_all_entries(ids: List[int], w_api: Wallabag):
for id in ids:
entry = await w_api.delete_entries(id)
logging.info(f"Deleted wallabag entry with id {entry['id']}")
async def main():
async with aiohttp.ClientSession(loop=asyncio.get_event_loop()) as session:
configuration = dict(load_configuration(CONFIG_PATH))
w_api = await connect_to_wallabag(configuration, session)
entries_id = await import_articles(w_api, "articles.json", limit=1)
await delete_all_entries(entries_id, w_api)
if __name__ == "__main__":
asyncio.run(main())