Home
Diffusion Pillar (master) Browse

pillar/pillar/cli/maintenance.py
151484dee39dmaster

Last Change
Raw File

maintenance.py
View Options

import copy
import datetime
import logging
from pathlib import PurePosixPath
import re
import typing
import bson.tz_util
from bson import ObjectId
from bson.errors import InvalidId
from flask_script import Manager
import pymongo
from pillar import current_app
# Collections to skip when finding file references (during orphan file detection).
# This collection can be added to from PillarExtension.setup_app().
ORPHAN_FINDER_SKIP_COLLECTIONS = {
# Skipping the files collection under the assumption that we have no files
# referencing other files.
'files',
# Authentication tokens never refer to files, and it's a big collection so
# good to skip.
'tokens',
}
log = logging.getLogger(__name__)
manager_maintenance = Manager(
current_app, usage="Maintenance scripts, to update user groups")
def _single_logger(*args, level=logging.INFO, **kwargs):
"""Construct a logger function that's only logging once."""
shown = False
def log_once():
nonlocal shown
if shown:
return
log.log(level, *args, **kwargs)
shown = True
return log_once
@manager_maintenance.command
def find_duplicate_users():
"""Finds users that have the same BlenderID user_id."""
from collections import defaultdict
users_coll = current_app.data.driver.db['users']
nodes_coll = current_app.data.driver.db['nodes']
projects_coll = current_app.data.driver.db['projects']
found_users = defaultdict(list)
for user in users_coll.find():
blender_ids = [auth['user_id'] for auth in user['auth']
if auth['provider'] == 'blender-id']
if not blender_ids:
continue
blender_id = blender_ids[0]
found_users[blender_id].append(user)
for blender_id, users in found_users.items():
if len(users) == 1:
continue
usernames = ', '.join(user['username'] for user in users)
print('Blender ID: %5s has %i users: %s' % (
blender_id, len(users), usernames))
for user in users:
print(' %s owns %i nodes and %i projects' % (
user['username'],
nodes_coll.count({'user': user['_id']}),
projects_coll.count({'user': user['_id']}),
))
@manager_maintenance.command
def sync_role_groups(do_revoke_groups):
"""For each user, synchronizes roles and group membership.
This ensures that everybody with the 'subscriber' role is also member of the 'subscriber'
group, and people without the 'subscriber' role are not member of that group. Same for
admin and demo groups.
When do_revoke_groups=False (the default), people are only added to groups.
when do_revoke_groups=True, people are also removed from groups.
"""
from pillar.api import service
if do_revoke_groups not in {'true', 'false'}:
print('Use either "true" or "false" as first argument.')
print('When passing "false", people are only added to groups.')
print('when passing "true", people are also removed from groups.')
raise SystemExit()
do_revoke_groups = do_revoke_groups == 'true'
service.fetch_role_to_group_id_map()
users_coll = current_app.data.driver.db['users']
groups_coll = current_app.data.driver.db['groups']
group_names = {}
def gname(gid):
try:
return group_names[gid]
except KeyError:
name = groups_coll.find_one(gid, projection={'name': 1})['name']
name = str(name)
group_names[gid] = name
return name
ok_users = bad_users = 0
for user in users_coll.find():
grant_groups = set()
revoke_groups = set()
current_groups = set(user.get('groups', []))
user_roles = user.get('roles', set())
for role in service.ROLES_WITH_GROUPS:
action = 'grant' if role in user_roles else 'revoke'
groups = service.manage_user_group_membership(user, role, action)
if groups is None:
# No changes required
continue
if groups == current_groups:
continue
grant_groups.update(groups.difference(current_groups))
revoke_groups.update(current_groups.difference(groups))
if grant_groups or revoke_groups:
bad_users += 1
expected_groups = current_groups.union(grant_groups).difference(revoke_groups)
print('Discrepancy for user %s/%s:' % (user['_id'], user['full_name'].encode('utf8')))
print(' - actual groups :', sorted(gname(gid) for gid in user.get('groups')))
print(' - expected groups:', sorted(gname(gid) for gid in expected_groups))
print(' - will grant :', sorted(gname(gid) for gid in grant_groups))
if do_revoke_groups:
label = 'WILL REVOKE '
else:
label = 'could revoke'
print(' - %s :' % label, sorted(gname(gid) for gid in revoke_groups))
if grant_groups and revoke_groups:
print(' ------ CAREFUL this one has BOTH grant AND revoke -----')
# Determine which changes we'll apply
final_groups = current_groups.union(grant_groups)
if do_revoke_groups:
final_groups.difference_update(revoke_groups)
print(' - final groups :', sorted(gname(gid) for gid in final_groups))
# Perform the actual update
users_coll.update_one({'_id': user['_id']},
{'$set': {'groups': list(final_groups)}})
else:
ok_users += 1
print('%i bad and %i ok users seen.' % (bad_users, ok_users))
@manager_maintenance.command
def sync_project_groups(user_email, fix):
"""Gives the user access to their self-created projects."""
if fix.lower() not in {'true', 'false'}:
print('Use either "true" or "false" as second argument.')
print('When passing "false", only a report is produced.')
print('when passing "true", group membership is fixed.')
raise SystemExit()
fix = fix.lower() == 'true'
users_coll = current_app.data.driver.db['users']
proj_coll = current_app.data.driver.db['projects']
groups_coll = current_app.data.driver.db['groups']
# Find by email or by user ID
if '@' in user_email:
where = {'email': user_email}
else:
try:
where = {'_id': ObjectId(user_email)}
except InvalidId:
log.warning('Invalid ObjectID: %s', user_email)
return
user = users_coll.find_one(where, projection={'_id': 1, 'groups': 1})
if user is None:
log.error('User %s not found', where)
raise SystemExit()
user_groups = set(user['groups'])
user_id = user['_id']
log.info('Updating projects for user %s', user_id)
ok_groups = missing_groups = 0
for proj in proj_coll.find({'user': user_id}):
project_id = proj['_id']
log.info('Investigating project %s (%s)', project_id, proj['name'])
# Find the admin group
admin_group = groups_coll.find_one({'name': str(project_id)}, projection={'_id': 1})
if admin_group is None:
log.warning('No admin group for project %s', project_id)
continue
group_id = admin_group['_id']
# Check membership
if group_id not in user_groups:
log.info('Missing group membership')
missing_groups += 1
user_groups.add(group_id)
else:
ok_groups += 1
log.info('User %s was missing %i group memberships; %i projects were ok.',
user_id, missing_groups, ok_groups)
if missing_groups > 0 and fix:
log.info('Updating database.')
result = users_coll.update_one({'_id': user_id},
{'$set': {'groups': list(user_groups)}})
log.info('Updated %i user.', result.modified_count)
@manager_maintenance.command
def check_home_project_groups():
"""Checks all users' group membership of their home project admin group."""
users_coll = current_app.data.driver.db['users']
proj_coll = current_app.data.driver.db['projects']
good = bad = 0
for proj in proj_coll.find({'category': 'home'}):
pid = proj['_id']
try:
admin_group_perms = proj['permissions']['groups'][0]
except IndexError:
log.error('Project %s has no admin group', pid)
return 255
except KeyError:
log.error('Project %s has no group permissions at all', pid)
return 255
user = users_coll.find_one({'_id': proj['user']},
projection={'groups': 1})
if user is None:
log.error('Project %s has non-existing owner %s', pid, proj['user'])
return 255
user_groups = set(user['groups'])
admin_group_id = admin_group_perms['group']
if admin_group_id in user_groups:
# All is fine!
good += 1
continue
log.warning('User %s has no admin rights to home project %s -- needs group %s',
proj['user'], pid, admin_group_id)
bad += 1
log.info('%i projects OK, %i projects in error', good, bad)
return bad
@manager_maintenance.option('-g', '--go', dest='go',
action='store_true', default=False,
help='Actually go and perform the changes, without this just '
'shows differences.')
def purge_home_projects(go=False):
"""Deletes all home projects that have no owner."""
from pillar.api.utils.authentication import force_cli_user
force_cli_user()
users_coll = current_app.data.driver.db['users']
proj_coll = current_app.data.driver.db['projects']
good = bad = 0
def bad_projects():
nonlocal good, bad
for proj in proj_coll.find({'category': 'home', '_deleted': {'$ne': True}}):
pid = proj['_id']
uid = proj.get('user')
if not uid:
log.info('Project %s has no user assigned', uid)
bad += 1
yield pid
continue
if users_coll.find({'_id': uid, '_deleted': {'$ne': True}}).count() == 0:
log.info('Project %s has non-existing owner %s', pid, uid)
bad += 1
yield pid
continue
good += 1
if not go:
log.info('Dry run, use --go to actually perform the changes.')
for project_id in bad_projects():
log.info('Soft-deleting project %s', project_id)
if go:
r, _, _, status = current_app.delete_internal('projects', _id=project_id)
if status != 204:
raise ValueError(f'Error {status} deleting {project_id}: {r}')
log.info('%i projects OK, %i projects deleted', good, bad)
return bad
@manager_maintenance.command
@manager_maintenance.option('-c', '--chunk', dest='chunk_size', default=50,
help='Number of links to update, use 0 to update all.')
@manager_maintenance.option('-q', '--quiet', dest='quiet', action='store_true', default=False)
@manager_maintenance.option('-w', '--window', dest='window', default=12,
help='Refresh links that expire in this many hours.')
def refresh_backend_links(backend_name, chunk_size=50, quiet=False, window=12):
"""Refreshes all file links that are using a certain storage backend.
Use `--chunk 0` to refresh all links.
"""
loglevel = logging.WARNING if quiet else logging.DEBUG
logging.getLogger('pillar.api.file_storage').setLevel(loglevel)
# CLI parameters are passed as strings
chunk_size = int(chunk_size)
window = int(window)
from pillar.api import file_storage
file_storage.refresh_links_for_backend(backend_name, chunk_size, window * 3600)
@manager_maintenance.command
@manager_maintenance.option('-c', '--chunk', dest='chunk_size', default=50,
help='Number of links to update, use 0 to update all.')
def refresh_backend_links_celery(backend_name, chunk_size=50):
"""Starts a Celery task that refreshes all file links that are using a certain storage backend.
"""
from pillar.celery import file_link_tasks
chunk_size = int(chunk_size) # CLI parameters are passed as strings
file_link_tasks.regenerate_all_expired_links.delay(backend_name, chunk_size)
log.info('File link regeneration task has been queued for execution.')
_var_type_re = re.compile(r'-[a-z0-9A-Z]+$')
def _fix_variation(fdoc, variation, nice_name):
from pillar.api.file_storage_backends import Bucket
# See if we can reuse the bucket we already had.
backend = fdoc['backend']
pid_str = str(fdoc['project'])
bucket_cls = Bucket.for_backend(backend)
bucket = bucket_cls(pid_str)
var_path = PurePosixPath(variation["file_path"])
# NOTE: this breaks for variations with double extensions
var_stem = var_path.stem
m = _var_type_re.search(var_stem)
var_type = m.group(0) if m else ''
var_name = f'{nice_name}{var_type}{var_path.suffix}'
log.info(f' - %s → %s', variation["file_path"], var_name)
blob = bucket.blob(variation['file_path'])
if not blob.exists():
log.warning('Blob %s does not exist', blob)
return
try:
blob.update_filename(var_name)
except Exception:
log.warning('Unable to update blob %s filename to %r', blob, var_name, exc_info=True)
@manager_maintenance.option('-p', '--project', dest='proj_url', nargs='?',
help='Project URL')
@manager_maintenance.option('-a', '--all', dest='all_projects', action='store_true', default=False,
help='Replace on all projects.')
@manager_maintenance.option('-c', '--chunk', dest='chunk_size', default=50,
help='Number of links to update, use 0 to update all.')
def refresh_content_disposition(proj_url=None, all_projects=False, chunk_size=0):
"""Refreshes the filename as mentioned in the Content Disposition header.
Works on all files of a specific project, or on all files in general.
Only works on variations, as this is intended to fix the database after
T51477 is fixed, and that issue doesn't affect the original files.
"""
from concurrent.futures import ProcessPoolExecutor as Executor
if bool(proj_url) == all_projects:
log.error('Use either --project or --all.')
return 1
# CLI parameters are passed as strings
chunk_size = int(chunk_size)
# Main implementation in separate function so that we're sure that
# fix_variation() doesn't accidentally use nonlocal variables.
def go():
query = {'_deleted': {'$ne': False}}
if proj_url:
from pillar.api.projects.utils import get_project
proj = get_project(proj_url)
query['project'] = proj['_id']
files_coll = current_app.db('files')
cursor = files_coll.find(query)
if all_projects:
cursor = cursor.sort([('project', pymongo.ASCENDING)])
cursor = cursor.limit(chunk_size)
with Executor(max_workers=15) as exe:
futures = []
for fdoc in cursor:
nice_name = PurePosixPath(fdoc['filename']).stem
variations = fdoc.get('variations') or []
futures.extend(exe.submit(_fix_variation, fdoc, variation, nice_name)
for variation in variations)
for future in futures:
future.result()
go()
@manager_maintenance.command
def expire_all_project_links(project_uuid):
"""Expires all file links for a certain project without refreshing.
This is just for testing.
"""
import datetime
from pillar.api.utils import utcnow
files_collection = current_app.data.driver.db['files']
expires = utcnow() - datetime.timedelta(days=1)
result = files_collection.update_many(
{'project': ObjectId(project_uuid)},
{'$set': {'link_expires': expires}}
)
print('Expired %i links' % result.matched_count)
@manager_maintenance.option('-u', '--url', dest='project_url', nargs='?',
help='Project URL')
@manager_maintenance.option('-a', '--all', dest='all_projects', action='store_true', default=False,
help='Replace on all projects.')
@manager_maintenance.option('-m', '--missing', dest='missing',
action='store_true', default=False,
help='Add missing node types. Note that this may add unwanted ones.')
@manager_maintenance.option('-g', '--go', dest='go',
action='store_true', default=False,
help='Actually go and perform the changes, without this just '
'shows differences.')
@manager_maintenance.option('-i', '--id', dest='project_id', nargs='?',
help='Project ID')
def replace_pillar_node_type_schemas(project_url=None, all_projects=False, missing=False, go=False,
project_id=None):
"""Replaces the project's node type schemas with the standard Pillar ones.
Non-standard node types are left alone.
"""
from pillar.api.utils.authentication import force_cli_user
force_cli_user()
from pillar.api.node_types import PILLAR_NAMED_NODE_TYPES
from pillar.api.utils import remove_private_keys, doc_diff
will_would = 'Will' if go else 'Would'
projects_changed = projects_seen = 0
for proj in _db_projects(project_url, all_projects, project_id, go=go):
projects_seen += 1
orig_proj = copy.deepcopy(proj)
proj_id = proj['_id']
if 'url' not in proj:
log.warning('Project %s has no URL!', proj_id)
proj_url = proj.get('url', f'-no URL id {proj_id}')
log.debug('Handling project %s', proj_url)
for proj_nt in proj['node_types']:
nt_name = proj_nt['name']
try:
pillar_nt = PILLAR_NAMED_NODE_TYPES[nt_name]
except KeyError:
log.debug(' - skipping non-standard node type "%s"', nt_name)
continue
log.debug(' - replacing schema on node type "%s"', nt_name)
# This leaves node type keys intact that aren't in Pillar's node_type_xxx definitions,
# such as permissions. It also keeps form schemas as-is.
pillar_nt.pop('form_schema', None)
proj_nt.update(copy.deepcopy(pillar_nt))
# Find new node types that aren't in the project yet.
if missing:
project_ntnames = set(nt['name'] for nt in proj['node_types'])
for nt_name in set(PILLAR_NAMED_NODE_TYPES.keys()) - project_ntnames:
log.info(' - Adding node type "%s"', nt_name)
pillar_nt = PILLAR_NAMED_NODE_TYPES[nt_name]
proj['node_types'].append(copy.deepcopy(pillar_nt))
proj_has_difference = False
for key, val1, val2 in doc_diff(orig_proj, proj, falsey_is_equal=False):
if not proj_has_difference:
if proj.get('_deleted', False):
deleted = ' (deleted)'
else:
deleted = ''
log.info('%s change project %s%s', will_would, proj_url, deleted)
proj_has_difference = True
log.info(' %30r: %r → %r', key, val1, val2)
projects_changed += proj_has_difference
if go and proj_has_difference:
# Use Eve to PUT, so we have schema checking.
db_proj = remove_private_keys(proj)
try:
r, _, _, status = current_app.put_internal('projects', db_proj, _id=proj_id)
except Exception:
log.exception('Error saving project %s (url=%s)', proj_id, proj_url)
raise SystemExit(5)
if status != 200:
log.error('Error %i storing altered project %s %s', status, proj['_id'], r)
raise SystemExit('Error storing project, see log.')
log.debug('Project saved succesfully.')
log.info('%s %d of %d projects',
'Changed' if go else 'Would change',
projects_changed, projects_seen)
@manager_maintenance.command
def remarkdown_comments():
"""Retranslates all Markdown to HTML for all comment nodes.
"""
from pillar.api.nodes import convert_markdown
nodes_collection = current_app.db()['nodes']
comments = nodes_collection.find({'node_type': 'comment'},
projection={'properties.content': 1,
'node_type': 1})
updated = identical = skipped = errors = 0
for node in comments:
convert_markdown(node)
node_id = node['_id']
try:
content_html = node['properties']['content_html']
except KeyError:
log.warning('Node %s has no content_html', node_id)
skipped += 1
continue
result = nodes_collection.update_one(
{'_id': node_id},
{'$set': {'properties.content_html': content_html}}
)
if result.matched_count != 1:
log.error('Unable to update node %s', node_id)
errors += 1
continue
if result.modified_count:
updated += 1
else:
identical += 1
log.info('updated : %i', updated)
log.info('identical: %i', identical)
log.info('skipped : %i', skipped)
log.info('errors : %i', errors)
@manager_maintenance.option('-p', '--project', dest='proj_url', nargs='?',
help='Project URL')
@manager_maintenance.option('-a', '--all', dest='all_projects', action='store_true', default=False,
help='Replace on all projects.')
@manager_maintenance.option('-g', '--go', dest='go', action='store_true', default=False,
help='Actually perform the changes (otherwise just show as dry-run).')
def upgrade_attachment_schema(proj_url=None, all_projects=False, go=False):
"""Replaces the project's attachments with the new schema.
Updates both the schema definition and the nodes with attachments (asset, page, post).
"""
if bool(proj_url) == all_projects:
log.error('Use either --project or --all.')
return 1
from pillar.api.utils.authentication import force_cli_user
force_cli_user()
from pillar.api.node_types.asset import node_type_asset
from pillar.api.node_types.page import node_type_page
from pillar.api.node_types.post import node_type_post
from pillar.api.utils import remove_private_keys, doc_diff
# Node types that support attachments
node_types = (node_type_asset, node_type_page, node_type_post)
nts_by_name = {nt['name']: nt for nt in node_types}
nodes_coll = current_app.db('nodes')
def replace_schemas(project):
project_url = project.get('url', '-no-url-')
log_proj = _single_logger('Upgrading schema project %s (%s)', project_url, project['_id'])
orig_proj = copy.deepcopy(project)
for proj_nt in project['node_types']:
nt_name = proj_nt['name']
if nt_name not in nts_by_name:
continue
pillar_nt = nts_by_name[nt_name]
pillar_dyn_schema = pillar_nt['dyn_schema']
if proj_nt['dyn_schema'] == pillar_dyn_schema:
# Schema already up to date.
continue
log_proj()
log.info(' - replacing dyn_schema on node type "%s"', nt_name)
proj_nt['dyn_schema'] = copy.deepcopy(pillar_dyn_schema)
seen_changes = False
for key, val1, val2 in doc_diff(orig_proj, project):
if not seen_changes:
log.info('Schema changes to project %s (%s):', project_url, project['_id'])
seen_changes = True
log.info(' - %30s: %s → %s', key, val1, val2)
if go:
# Use Eve to PUT, so we have schema checking.
db_proj = remove_private_keys(project)
r, _, _, status = current_app.put_internal('projects', db_proj, _id=project['_id'])
if status != 200:
log.error('Error %i storing altered project %s %s', status, project['_id'], r)
raise SystemExit('Error storing project, see log.')
log.debug('Project saved succesfully.')
def replace_attachments(project):
project_url = project.get('url', '-no-url-')
log_proj = _single_logger('Upgrading nodes for project %s (%s)',
project_url, project['_id'])
# Remove empty attachments
if go:
res = nodes_coll.update_many(
{'properties.attachments': {},
'project': project['_id']},
{'$unset': {'properties.attachments': 1}},
)
if res.matched_count > 0:
log_proj()
log.info('Removed %d empty attachment dicts', res.modified_count)
else:
to_remove = nodes_coll.count({'properties.attachments': {},
'project': project['_id']})
if to_remove:
log_proj()
log.info('Would remove %d empty attachment dicts', to_remove)
# Convert attachments.
nodes = nodes_coll.find({
'_deleted': False,
'project': project['_id'],
'node_type': {'$in': list(nts_by_name)},
'properties.attachments': {'$exists': True},
})
for node in nodes:
attachments = node['properties']['attachments']
if not attachments:
# If we're not modifying the database (e.g. go=False),
# any attachments={} will not be filtered out earlier.
if go or attachments != {}:
log_proj()
log.info(' - Node %s (%s) still has empty attachments %r',
node['_id'], node.get('name'), attachments)
continue
if isinstance(attachments, dict):
# This node has already been upgraded.
continue
# Upgrade from list [{'slug': 'xxx', 'oid': 'yyy'}, ...]
# to dict {'xxx': {'oid': 'yyy'}, ...}
log_proj()
log.info(' - Updating schema on node %s (%s)', node['_id'], node.get('name'))
new_atts = {}
for field_info in attachments:
for attachment in field_info.get('files', []):
new_atts[attachment['slug']] = {'oid': attachment['file']}
node['properties']['attachments'] = new_atts
log.info(' from %s to %s', attachments, new_atts)
if go:
# Use Eve to PUT, so we have schema checking.
db_node = remove_private_keys(node)
r, _, _, status = current_app.put_internal('nodes', db_node, _id=node['_id'])
if status != 200:
log.error('Error %i storing altered node %s %s', status, node['_id'], r)
raise SystemExit('Error storing node; see log.')
for proj in _db_projects(proj_url, all_projects, go=go):
replace_schemas(proj)
replace_attachments(proj)
def iter_markdown(proj_node_types: dict, some_node: dict, callback: typing.Callable[[str], str]):
"""Calls the callback for each MarkDown value in the node.
Replaces the value in-place with the return value of the callback.
"""
from collections import deque
from pillar.api.eve_settings import nodes_schema
my_log = log.getChild('iter_markdown')
# Inspect the node type to find properties containing Markdown.
node_type_name = some_node['node_type']
try:
node_type = proj_node_types[node_type_name]
except KeyError:
raise KeyError(f'Project has no node type {node_type_name}')
to_visit = deque([
(some_node, nodes_schema),
(some_node['properties'], node_type['dyn_schema'])])
while to_visit:
doc, doc_schema = to_visit.popleft()
for key, definition in doc_schema.items():
if definition.get('type') == 'dict' and definition.get('schema'):
# This is a subdocument with its own schema, visit it later.
subdoc = doc.get(key)
if not subdoc:
continue
to_visit.append((subdoc, definition['schema']))
continue
if definition.get('coerce') != 'markdown':
continue
my_log.debug('I have to change %r of %s', key, doc)
old_value = doc.get(key)
if not old_value:
continue
new_value = callback(old_value)
doc[key] = new_value
@manager_maintenance.option('-p', '--project', dest='proj_url', nargs='?',
help='Project URL')
@manager_maintenance.option('-a', '--all', dest='all_projects', action='store_true', default=False,
help='Replace on all projects.')
@manager_maintenance.option('-g', '--go', dest='go', action='store_true', default=False,
help='Actually perform the changes (otherwise just show as dry-run).')
def upgrade_attachment_usage(proj_url=None, all_projects=False, go=False):
"""Replaces '@[slug]' with '{attachment slug}'.
Also moves links from the attachment dict to the attachment shortcode.
"""
if bool(proj_url) == all_projects:
log.error('Use either --project or --all.')
return 1
import html
from pillar.api.projects.utils import node_type_dict
from pillar.api.utils import remove_private_keys
from pillar.api.utils.authentication import force_cli_user
force_cli_user()
nodes_coll = current_app.db('nodes')
total_nodes = 0
failed_node_ids = set()
# Use a mixture of the old slug RE that still allowes spaces in the slug
# name and the new RE that allows dashes.
old_slug_re = re.compile(r'@\[([a-zA-Z0-9_\- ]+)\]')
for proj in _db_projects(proj_url, all_projects, go=go):
proj_id = proj['_id']
proj_url = proj.get('url', '-no-url-')
nodes = nodes_coll.find({
'_deleted': {'$ne': True},
'project': proj_id,
'properties.attachments': {'$exists': True},
})
node_count = nodes.count()
if node_count == 0:
log.debug('Skipping project %s (%s)', proj_url, proj_id)
continue
proj_node_types = node_type_dict(proj)
for node in nodes:
attachments = node['properties']['attachments']
replaced = False
# Inner functions because of access to the node's attachments.
def replace(match):
nonlocal replaced
slug = match.group(1)
log.debug(' - OLD STYLE attachment slug %r', slug)
try:
att = attachments[slug]
except KeyError:
log.info("Attachment %r not found for node %s", slug, node['_id'])
link = ''
else:
link = att.get('link', '')
if link == 'self':
link = " link='self'"
elif link == 'custom':
url = att.get('link_custom')
if url:
link = " link='%s'" % html.escape(url)
replaced = True
return '{attachment %r%s}' % (slug.replace(' ', '-'), link)
def update_markdown(value: str) -> str:
return old_slug_re.sub(replace, value)
iter_markdown(proj_node_types, node, update_markdown)
# Remove no longer used properties from attachments
new_attachments = {}
for slug, attachment in attachments.items():
replaced |= 'link' in attachment # link_custom implies link
attachment.pop('link', None)
attachment.pop('link_custom', None)
new_attachments[slug.replace(' ', '-')] = attachment
node['properties']['attachments'] = new_attachments
if replaced:
total_nodes += 1
else:
# Nothing got replaced,
continue
if go:
# Use Eve to PUT, so we have schema checking.
db_node = remove_private_keys(node)
r, _, _, status = current_app.put_internal('nodes', db_node, _id=node['_id'])
if status != 200:
log.error('Error %i storing altered node %s %s', status, node['_id'], r)
failed_node_ids.add(node['_id'])
# raise SystemExit('Error storing node; see log.')
log.debug('Updated node %s: %s', node['_id'], r)
log.info('Project %s (%s) has %d nodes with attachments',
proj_url, proj_id, node_count)
log.info('%s %d nodes', 'Updated' if go else 'Would update', total_nodes)
if failed_node_ids:
log.warning('Failed to update %d of %d nodes: %s', len(failed_node_ids), total_nodes,
', '.join(str(nid) for nid in failed_node_ids))
def _db_projects(proj_url: str, all_projects: bool, project_id='', *, go: bool) \
-> typing.Iterable[dict]:
"""Yields a subset of the projects in the database.
:param all_projects: when True, yields all projects.
:param proj_url: when all_projects is False, this denotes the project
to yield.
Handles soft-deleted projects as non-existing. This ensures that
the receiver can actually modify and save the project without any
issues.
Also shows duration and a note about dry-running when go=False.
"""
if sum([bool(proj_url), all_projects, bool(project_id)]) != 1:
log.error('Only use one way to specify a project / all projects')
raise SystemExit(1)
projects_coll = current_app.db('projects')
start = datetime.datetime.now()
if all_projects:
yield from projects_coll.find({'_deleted': {'$ne': True}})
else:
if proj_url:
q = {'url': proj_url}
else:
q = {'_id': bson.ObjectId(project_id)}
proj = projects_coll.find_one({**q, '_deleted': {'$ne': True}})
if not proj:
log.error('Project %s not found', q)
raise SystemExit(3)
yield proj
if not go:
log.info('Dry run, use --go to perform the change.')
duration = datetime.datetime.now() - start
log.info('Command took %s', duration)
def _find_orphan_files() -> typing.Set[bson.ObjectId]:
"""Finds all non-referenced files for the given project.
Returns an iterable of all orphan file IDs.
"""
log.debug('Finding orphan files')
# Get all file IDs that belong to this project.
files_coll = current_app.db('files')
cursor = files_coll.find({'_deleted': {'$ne': True}}, projection={'_id': 1})
file_ids = {doc['_id'] for doc in cursor}
if not file_ids:
log.debug('No files found')
return set()
total_file_count = len(file_ids)
log.debug('Found %d files in total', total_file_count)
def find_object_ids(something: typing.Any) -> typing.Iterable[bson.ObjectId]:
if isinstance(something, bson.ObjectId):
yield something
elif isinstance(something, str) and len(something) == 24:
try:
yield bson.ObjectId(something)
except (bson.objectid.InvalidId, TypeError):
# It apparently wasn't an ObjectID after all.
pass
elif isinstance(something, (list, set, tuple)):
for item in something:
yield from find_object_ids(item)
elif isinstance(something, dict):
for item in something.values():
yield from find_object_ids(item)
# Find all references by iterating through the project itself and every document that has a
# 'project' key set to this ObjectId.
db = current_app.db()
for coll_name in sorted(db.collection_names(include_system_collections=False)):
if coll_name in ORPHAN_FINDER_SKIP_COLLECTIONS:
continue
doc_filter = {'_deleted': {'$ne': True}}
log.debug(' - inspecting collection %r with filter %r', coll_name, doc_filter)
coll = db[coll_name]
for doc in coll.find(doc_filter):
for obj_id in find_object_ids(doc):
# Found an Object ID that is in use, so discard it from our set of file IDs.
file_ids.discard(obj_id)
orphan_count = len(file_ids)
log.info('Found %d files or which %d are orphaned (%d%%)',
total_file_count, orphan_count, 100 * orphan_count / total_file_count)
return file_ids
@manager_maintenance.command
def find_orphan_files():
"""Finds unused files in the given project.
This is a heavy operation that inspects *everything* in MongoDB. Use with care.
"""
from jinja2.filters import do_filesizeformat
from pathlib import Path
output_fpath = Path(current_app.config['STORAGE_DIR']) / 'orphan-files.txt'
if output_fpath.exists():
log.error('Output filename %s already exists, remove it first.', output_fpath)
return 1
start_timestamp = datetime.datetime.now()
orphans = _find_orphan_files()
if not orphans:
log.info('No orphan files found, congratulations.')
return 0
files_coll = current_app.db('files')
aggr = files_coll.aggregate([
{'$match': {'_id': {'$in': list(orphans)}}},
{'$group': {
'_id': None,
'size': {'$sum': '$length_aggregate_in_bytes'},
}}
])
total_size = list(aggr)[0]['size']
log.info('Total orphan file size: %s', do_filesizeformat(total_size, binary=True))
orphan_count = len(orphans)
total_count = files_coll.count()
log.info('Total nr of orphan files: %d', orphan_count)
log.info('Total nr of files : %d', total_count)
log.info('Orphan percentage : %d%%', 100 * orphan_count / total_count)
end_timestamp = datetime.datetime.now()
duration = end_timestamp - start_timestamp
log.info('Finding orphans took %s', duration)
log.info('Writing Object IDs to %s', output_fpath)
with output_fpath.open('w', encoding='ascii') as outfile:
outfile.write('\n'.join(str(oid) for oid in sorted(orphans)) + '\n')
@manager_maintenance.command
def delete_orphan_files():
"""Deletes orphan files mentioned in orphan-files.txt
Use 'find_orphan_files' first to generate orphan-files.txt.
"""
import pymongo.results
from pathlib import Path
output_fpath = Path(current_app.config['STORAGE_DIR']) / 'orphan-files.txt'
with output_fpath.open('r', encoding='ascii') as infile:
oids = [bson.ObjectId(oid.strip()) for oid in infile]
log.info('Found %d Object IDs to remove', len(oids))
# Ensure that the list of Object IDs actually matches files.
# I hope this works as a security measure against deleting from obsolete orphan-files.txt files.
files_coll = current_app.db('files')
oid_filter = {'_id': {'$in': oids},
'_deleted': {'$ne': True}}
file_count = files_coll.count(oid_filter)
if file_count == len(oids):
log.info('Found %d matching files', file_count)
else:
log.warning("Found %d matching files, which doesn't match the number of Object IDs. "
"Refusing to continue.", file_count)
return 1
res: pymongo.results.UpdateResult = files_coll.update_many(
oid_filter,
{'$set': {'_deleted': True}}
)
if res.matched_count != file_count:
log.warning('Soft-deletion matched %d of %d files', res.matched_count, file_count)
elif res.modified_count != file_count:
log.warning('Soft-deletion modified %d of %d files', res.modified_count, file_count)
log.info('%d files have been soft-deleted', res.modified_count)