Import forum profile pictures and host them directly

This commit is contained in:
rubenwardy 2023-05-05 17:14:23 +01:00
parent a026e2c2bb
commit 3ccb165522
6 changed files with 96 additions and 55 deletions

@ -27,9 +27,9 @@ from app.logic.game_support import GameSupportResolver
from app.models import PackageRelease, db, Package, PackageState, PackageScreenshot, MetaPackage, User, \
NotificationType, PackageUpdateConfig, License, UserRank, PackageType, ThreadReply
from app.tasks.emails import send_pending_digests
from app.tasks.forumtasks import importTopicList, checkAllForumAccounts
from app.tasks.forumtasks import importTopicList, checkAllForumAccounts, checkForumAccount
from app.tasks.importtasks import importRepoScreenshot, checkZipRelease, check_for_updates, updateAllGameSupport
from app.tasks.usertasks import upgrade_new_members
from app.tasks.usertasks import upgrade_new_members, set_profile_picture_from_url
from app.utils import addNotification, get_system_user
from app.utils.image import get_image_size
@ -102,6 +102,13 @@ def check_all_forum_accounts():
return redirect(url_for("tasks.check", id=task.id, r=url_for("admin.admin_page")))
@action("Import forum profile pics")
def import_forum_profile_pics():
users = User.query.filter(and_(User.forums_username.isnot(None), User.profile_pic.ilike("https://forum.minetest.net/%"))).all()
for user in users:
checkForumAccount.delay(user.forums_username)
@action("Import screenshots from Git")
def import_screenshots():
packages = Package.query \
@ -129,8 +136,9 @@ def clean_uploads():
release_urls = get_filenames_from_column(PackageRelease.url)
screenshot_urls = get_filenames_from_column(PackageScreenshot.url)
pp_urls = get_filenames_from_column(User.profile_pic)
db_urls = release_urls.union(screenshot_urls)
db_urls = release_urls.union(screenshot_urls).union(pp_urls)
unreachable = existing_uploads.difference(db_urls)
import sys
@ -343,3 +351,12 @@ def set_new_members():
task_id = uuid()
upgrade_new_members.apply_async((), task_id=task_id)
return redirect(url_for("tasks.check", id=task_id, r=url_for("admin.admin_page")))
@action("Import profile pictures from forums")
def import_forum_pp():
users = User.query.filter(User.profile_pic.ilike("https://forum.minetest.net/%")).all()
for user in users:
set_profile_picture_from_url.delay(user.username, user.profile_pic)
flash(f"Importing {len(users)} profile pictures", "success")

@ -235,25 +235,6 @@ def profile(username):
medals_unlocked=unlocked, medals_locked=locked)
@bp.route("/users/<username>/check/", methods=["POST"])
@login_required
def user_check(username):
user = User.query.filter_by(username=username).first()
if user is None:
abort(404)
if current_user != user and not current_user.rank.atLeast(UserRank.MODERATOR):
abort(403)
if user.forums_username is None:
abort(404)
task = checkForumAccount.delay(user.forums_username)
next_url = url_for("users.profile", username=username)
return redirect(url_for("tasks.check", id=task.id, r=next_url))
@bp.route("/user/stats/")
@login_required
def statistics_redirect():

@ -21,25 +21,29 @@ from app.tasks import celery
from app.utils import is_username_valid
from app.utils.phpbbparser import getProfile, getTopicsFromForum
import urllib.request
from urllib.parse import urljoin
from .usertasks import set_profile_picture_from_url
@celery.task()
def checkForumAccount(username, forceNoSave=False):
print("Checking " + username)
def checkForumAccount(forums_username):
print("### Checking " + forums_username, file=sys.stderr)
try:
profile = getProfile("https://forum.minetest.net", username)
except OSError:
profile = getProfile("https://forum.minetest.net", forums_username)
except OSError as e:
print(e, file=sys.stderr)
return
if profile is None:
return
user = User.query.filter_by(forums_username=username).first()
user = User.query.filter_by(forums_username=forums_username).first()
# Create user
needsSaving = False
if user is None:
user = User(username)
user.forums_username = username
user = User(forums_username)
user.forums_username = forums_username
db.session.add(user)
# Get github username
@ -50,33 +54,32 @@ def checkForumAccount(username, forceNoSave=False):
needsSaving = True
pic = profile.avatar
if pic and "http" in pic:
if pic and pic.startswith("http"):
pic = None
needsSaving = needsSaving or pic != user.profile_pic
if pic:
user.profile_pic = "https://forum.minetest.net/" + pic
else:
user.profile_pic = None
# Save
if needsSaving and not forceNoSave:
if needsSaving:
db.session.commit()
if pic:
pic = urljoin("https://forum.minetest.net/", pic)
print(f"####### Picture: {pic}", file=sys.stderr)
print(f"####### User pp {user.profile_pic}", file=sys.stderr)
pic_needs_replacing = user.profile_pic is None or user.profile_pic == "" or \
user.profile_pic.startswith("https://forum.minetest.net")
if pic_needs_replacing and pic.startswith("https://forum.minetest.net"):
print(f"####### Queueing", file=sys.stderr)
set_profile_picture_from_url.delay(user.username, pic)
return needsSaving
@celery.task()
def checkAllForumAccounts(forceNoSave=False):
needsSaving = False
def checkAllForumAccounts():
query = User.query.filter(User.forums_username.isnot(None))
for user in query.all():
needsSaving = checkForumAccount(user.username) or needsSaving
if needsSaving and not forceNoSave:
db.session.commit()
return needsSaving
checkForumAccount(user.forums_username)
regex_tag = re.compile(r"\[([a-z0-9_]+)\]")

@ -1,5 +1,5 @@
# ContentDB
# Copyright (C) 2021 rubenwardy
# Copyright (C) 2021-23 rubenwardy
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published by
@ -15,13 +15,17 @@
# along with this program. If not, see <https://www.gnu.org/licenses/>.
import datetime
import datetime, requests
import os
import sys
from sqlalchemy import or_, and_
from app import app
from app.models import User, db, UserRank, ThreadReply, Package
from app.utils import randomString
from app.utils.models import create_session
from app.tasks import celery
from app.tasks import celery, TaskError
@celery.task()
@ -46,3 +50,45 @@ def upgrade_new_members():
User.packages.any(Package.approved_at < threshold)))).update({"rank": UserRank.MEMBER}, synchronize_session=False)
session.commit()
@celery.task()
def set_profile_picture_from_url(username: str, url: str):
print("### Setting pp for " + username + " to " + url, file=sys.stderr)
user = User.query.filter_by(username=username).first()
if user is None:
raise TaskError(f"Unable to find user {username}")
headers = {"Accept": "image/jpeg, image/png, image/gif"}
resp = requests.get(url, stream=True, headers=headers, timeout=15)
if resp.status_code != 200:
raise TaskError(f"Failed to download {url}: {resp.status_code}: {resp.reason}")
content_type = resp.headers["content-type"]
if content_type is None:
raise TaskError("Content-Type needed")
elif content_type == "image/jpeg":
ext = "jpg"
elif content_type == "image/png":
ext = "png"
elif content_type == "image/gif":
ext = "gif"
else:
raise TaskError(f"Unacceptable content-type: {content_type}")
filename = randomString(10) + "." + ext
filepath = os.path.join(app.config["UPLOAD_DIR"], filename)
with open(filepath, "wb") as f:
size = 0
for chunk in resp.iter_content(chunk_size=1024):
if chunk: # filter out keep-alive new chunks
size += len(chunk)
if size > 3 * 1000 * 1000: # 3 MB
raise TaskError(f"File too large to download {url}")
f.write(chunk)
user.profile_pic = "/uploads/" + filename
db.session.commit()
return filepath

@ -21,12 +21,6 @@
{% endif %}
</div>
<div class="col">
{% if user.forums_username %}
<form method="post" action="{{ url_for('users.user_check', username=user.username) }}" class="" style="display:inline-block;">
<input type="hidden" name="csrf_token" value="{{ csrf_token() }}" />
<input type="submit" class="btn btn-primary" value="{{ _('Sync with Forums') }}" />
</form>
{% endif %}
{% if user.email %}
<a class="btn btn-primary" href="https://en.gravatar.com/">
Gravatar

@ -92,7 +92,7 @@ def getProfile(url, username):
url = getProfileURL(url, username)
try:
req = urllib.request.urlopen(url, timeout=5)
req = urllib.request.urlopen(url, timeout=15)
except urllib.error.HTTPError as e:
if e.code == 404:
return None