nopaque/app/models.py

495 lines
16 KiB
Python
Raw Normal View History

from flask import current_app
from flask_login import UserMixin, AnonymousUserMixin
2019-08-22 07:35:23 +00:00
from itsdangerous import BadSignature, TimedJSONWebSignatureSerializer
from werkzeug.security import generate_password_hash, check_password_hash
from . import db
from . import login_manager
2019-08-06 12:26:22 +00:00
from datetime import datetime
2019-09-12 14:45:05 +00:00
import os
import shutil
import logging
class Permission:
2019-07-11 13:33:48 +00:00
"""
Defines User permissions as integers by the power of 2. User permission
can be evaluated using the bitwise operator &. 3 equals to CREATE_JOB and
DELETE_JOB and so on.
"""
CREATE_JOB = 1
DELETE_JOB = 2
# WRITE = 4
# MODERATE = 8
ADMIN = 16
class Role(db.Model):
2019-07-11 13:33:48 +00:00
"""
2019-08-22 07:35:23 +00:00
Model for the different roles Users can have. Is a one-to-many
relationship. A Role can be associated with many User rows.
2019-07-11 13:33:48 +00:00
"""
__tablename__ = 'roles'
2019-08-06 09:47:04 +00:00
# Primary key
id = db.Column(db.Integer, primary_key=True)
default = db.Column(db.Boolean, default=False, index=True)
2019-08-06 09:47:04 +00:00
name = db.Column(db.String(64), unique=True)
permissions = db.Column(db.Integer)
2019-08-06 09:47:04 +00:00
# Relationships
users = db.relationship('User', backref='role', lazy='dynamic')
def __init__(self, **kwargs):
super(Role, self).__init__(**kwargs)
if self.permissions is None:
self.permissions = 0
def __repr__(self):
2019-07-11 13:33:48 +00:00
"""
String representation of the Role. For human readability.
"""
return '<Role %r>' % self.name
def add_permission(self, perm):
2019-07-11 13:33:48 +00:00
"""
Add new permission to Role. Input is a Permission.
"""
if not self.has_permission(perm):
self.permissions += perm
def remove_permission(self, perm):
2019-07-11 13:33:48 +00:00
"""
Removes permission from a Role. Input a Permission.
"""
if self.has_permission(perm):
self.permissions -= perm
def reset_permissions(self):
2019-07-11 13:33:48 +00:00
"""
Resets permissions to zero. Zero equals no permissions at all.
"""
self.permissions = 0
def has_permission(self, perm):
2019-07-11 13:33:48 +00:00
"""
Checks if a Role has a specific Permission. Does this with the bitwise
2019-07-11 13:33:48 +00:00
operator.
"""
return self.permissions & perm == perm
@staticmethod
def insert_roles():
2019-07-11 13:33:48 +00:00
"""
Inserts roles into the databes. This has to be executed befor Users are
added to the database. Otherwiese Users will not have a Role assigned
to them. Order of the roles dictionary determines the ID of each role.
User hast the ID 1 and Administrator has the ID 2.
"""
roles = {
'User': [Permission.CREATE_JOB],
'Administrator': [Permission.ADMIN,
Permission.CREATE_JOB,
Permission.DELETE_JOB]
}
default_role = 'User'
for r in roles:
role = Role.query.filter_by(name=r).first()
if role is None:
role = Role(name=r)
role.reset_permissions()
for perm in roles[r]:
role.add_permission(perm)
role.default = (role.name == default_role)
db.session.add(role)
db.session.commit()
class User(UserMixin, db.Model):
2019-07-11 13:33:48 +00:00
"""
Model for Users that are registered to Opaque.
"""
__tablename__ = 'users'
2019-08-06 09:47:04 +00:00
# Primary key
id = db.Column(db.Integer, primary_key=True)
2019-08-06 09:47:04 +00:00
confirmed = db.Column(db.Boolean, default=False)
email = db.Column(db.String(254), unique=True, index=True)
password_hash = db.Column(db.String(128))
2019-08-06 13:39:09 +00:00
registration_date = db.Column(db.DateTime(), default=datetime.utcnow)
role_id = db.Column(db.Integer, db.ForeignKey('roles.id'))
2019-08-06 09:47:04 +00:00
username = db.Column(db.String(64), unique=True, index=True)
# Relationships
corpora = db.relationship('Corpus', backref='creator', lazy='dynamic',
cascade='save-update, merge, delete')
jobs = db.relationship('Job', backref='creator', lazy='dynamic',
cascade='save-update, merge, delete')
2019-10-09 14:10:30 +00:00
is_dark = db.Column(db.Boolean, default=False)
def __repr__(self):
2019-07-11 13:33:48 +00:00
"""
String representation of the User. For human readability.
"""
return '<User %r>' % self.username
def __init__(self, **kwargs):
super(User, self).__init__(**kwargs)
if self.role is None:
if self.email == current_app.config['OPAQUE_ADMIN']:
self.role = Role.query.filter_by(name='Administrator').first()
if self.role is None:
self.role = Role.query.filter_by(default=True).first()
2019-07-08 13:59:15 +00:00
def generate_confirmation_token(self, expiration=3600):
2019-07-11 13:33:48 +00:00
"""
Generates a confirmation token for user confirmation via email.
"""
2019-08-22 07:35:23 +00:00
s = TimedJSONWebSignatureSerializer(current_app.config['SECRET_KEY'],
expiration)
2019-07-08 13:59:15 +00:00
return s.dumps({'confirm': self.id}).decode('utf-8')
def generate_reset_token(self, expiration=3600):
2019-07-11 13:33:48 +00:00
"""
Generates a reset token for password reset via email.
"""
2019-08-22 07:35:23 +00:00
s = TimedJSONWebSignatureSerializer(current_app.config['SECRET_KEY'],
expiration)
return s.dumps({'reset': self.id}).decode('utf-8')
2019-07-08 13:59:15 +00:00
def confirm(self, token):
2019-07-11 13:33:48 +00:00
"""
Confirms User if the given token is valid and not expired.
"""
2019-08-22 07:35:23 +00:00
s = TimedJSONWebSignatureSerializer(current_app.config['SECRET_KEY'])
2019-07-08 13:59:15 +00:00
try:
data = s.loads(token.encode('utf-8'))
2019-08-22 07:35:23 +00:00
except BadSignature:
2019-07-08 13:59:15 +00:00
return False
if data.get('confirm') != self.id:
return False
self.confirmed = True
db.session.add(self)
return True
2019-07-08 13:13:32 +00:00
@staticmethod
def reset_password(token, new_password):
2019-07-11 13:33:48 +00:00
"""
Resets password for User if the given token is valid and not expired.
"""
2019-08-22 07:35:23 +00:00
s = TimedJSONWebSignatureSerializer(current_app.config['SECRET_KEY'])
2019-07-08 13:13:32 +00:00
try:
data = s.loads(token.encode('utf-8'))
2019-08-22 07:35:23 +00:00
except BadSignature:
2019-07-08 13:13:32 +00:00
return False
user = User.query.get(data.get('reset'))
if user is None:
return False
user.password = new_password
db.session.add(user)
return True
@property
def password(self):
raise AttributeError('password is not a readable attribute')
@password.setter
def password(self, password):
self.password_hash = generate_password_hash(password)
def verify_password(self, password):
return check_password_hash(self.password_hash, password)
def can(self, perm):
2019-07-11 13:33:48 +00:00
"""
Checks if a User with its current role can doe something. Checks if the
associated role actually has the needed Permission.
"""
return self.role is not None and self.role.has_permission(perm)
def is_administrator(self):
2019-07-11 13:33:48 +00:00
"""
Checks if User has Admin permissions.
"""
return self.can(Permission.ADMIN)
2019-08-30 11:31:00 +00:00
def corpora_as_dict(self):
corpora = {}
for corpus in self.corpora:
corpora[str(corpus.id)] = corpus.to_dict()
return corpora
def jobs_as_dict(self):
jobs = {}
for job in self.jobs:
jobs[str(job.id)] = job.to_dict()
return jobs
def delete_user(self):
"""
Delete user from database. Also delete all associated jobs and corpora
files.
"""
logger = logging.getLogger(__name__)
delete_path = os.path.join('/mnt/opaque/', str(self.id))
logger.warning('Delete path for user is: {}'.format(delete_path))
while os.path.exists(delete_path):
try:
shutil.rmtree(delete_path, ignore_errors=True)
logger.warning('Path does still exist.')
except OSError:
pass
db.session.delete(self)
db.session.commit()
2019-09-09 14:17:59 +00:00
class AnonymousUser(AnonymousUserMixin):
2019-07-11 13:33:48 +00:00
"""
Model replaces the default AnonymousUser.
"""
2019-08-06 12:26:22 +00:00
def can(self, permissions):
return False
def is_administrator(self):
return False
2019-10-16 14:52:05 +00:00
class JobInput(db.Model):
"""
2019-10-17 11:26:20 +00:00
Class to define JobInputs.
2019-10-16 14:52:05 +00:00
"""
__tablename__ = 'job_inputs'
# Primary key
id = db.Column(db.Integer, primary_key=True)
filename = db.Column(db.String(255))
dir = db.Column(db.String(255))
2019-10-16 14:52:05 +00:00
job_id = db.Column(db.Integer, db.ForeignKey('jobs.id'))
2019-10-17 11:32:22 +00:00
# Relationships
results = db.relationship('JobResult',
backref='job_input',
lazy='dynamic',
cascade='save-update, merge, delete')
2019-10-16 14:52:05 +00:00
2019-10-17 11:26:20 +00:00
def __repr__(self):
"""
String representation of the JobInput. For human readability.
"""
return '<JobInput %r>' % self.filename
def to_dict(self):
return {'id': self.id,
2019-10-24 11:29:02 +00:00
'dir': self.dir,
2019-10-17 11:26:20 +00:00
'filename': self.filename,
'job_id': self.job_id,
'results': [result.to_dict() for result in self.results]}
2019-10-17 11:26:20 +00:00
2019-10-16 14:52:05 +00:00
class JobResult(db.Model):
"""
2019-10-17 11:26:20 +00:00
Class to define JobResults.
2019-10-16 14:52:05 +00:00
"""
__tablename__ = 'job_results'
# Primary key
id = db.Column(db.Integer, primary_key=True)
filename = db.Column(db.String(255))
dir = db.Column(db.String(255))
2019-10-16 14:52:05 +00:00
job_id = db.Column(db.Integer, db.ForeignKey('jobs.id'))
2019-10-17 11:32:22 +00:00
job_input_id = db.Column(db.Integer, db.ForeignKey('job_inputs.id'))
2019-10-16 14:52:05 +00:00
2019-10-17 11:26:20 +00:00
def __repr__(self):
"""
String representation of the JobResult. For human readability.
"""
return '<JobResult %r>' % self.filename
def to_dict(self):
return {'id': self.id,
2019-10-24 11:29:02 +00:00
'dir': self.dir,
2019-10-17 11:26:20 +00:00
'filename': self.filename,
2019-10-24 11:29:02 +00:00
'job_id': self.job_id,
'job_input_id': self.job_input_id}
2019-10-17 11:26:20 +00:00
2019-10-16 14:52:05 +00:00
2019-08-06 09:47:04 +00:00
class Job(db.Model):
2019-08-05 14:45:38 +00:00
"""
Class to define Jobs.
"""
__tablename__ = 'jobs'
2019-08-06 09:47:04 +00:00
# Primary key
2019-08-05 14:45:38 +00:00
id = db.Column(db.Integer, primary_key=True)
2019-08-06 12:26:22 +00:00
creation_date = db.Column(db.DateTime(), default=datetime.utcnow)
2019-08-09 09:48:43 +00:00
description = db.Column(db.String(255))
end_date = db.Column(db.DateTime())
2019-08-09 09:48:43 +00:00
mem_mb = db.Column(db.Integer)
n_cores = db.Column(db.Integer)
2019-08-06 09:47:04 +00:00
service = db.Column(db.String(64))
'''
2019-08-09 09:48:43 +00:00
' Service specific arguments as string list.
' Example: ["-l eng", "--keep-intermediates", "--skip-binarization"]
2019-08-06 09:47:04 +00:00
'''
service_args = db.Column(db.String(255))
2019-08-09 09:48:43 +00:00
service_version = db.Column(db.String(16))
status = db.Column(db.String(16))
2019-08-06 09:47:04 +00:00
title = db.Column(db.String(32))
user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
2019-10-16 14:52:05 +00:00
# Relationships
inputs = db.relationship('JobInput',
backref='job',
lazy='dynamic',
cascade='save-update, merge, delete')
results = db.relationship('JobResult',
backref='job',
lazy='dynamic',
cascade='save-update, merge, delete')
2019-08-05 14:45:38 +00:00
def __init__(self, **kwargs):
super(Job, self).__init__(**kwargs)
def __repr__(self):
"""
String representation of the Job. For human readability.
"""
return '<Job %r>' % self.title
def to_dict(self):
return {'id': self.id,
'creation_date': self.creation_date.timestamp(),
'description': self.description,
'end_date': (self.end_date.timestamp() if self.end_date else
None),
2019-10-17 11:26:20 +00:00
'inputs': [input.to_dict() for input in self.inputs],
'mem_mb': self.mem_mb,
'n_cores': self.n_cores,
2019-10-17 11:26:20 +00:00
'results': [result.to_dict() for result in self.results],
'service': self.service,
'service_args': self.service_args,
'service_version': self.service_version,
'status': self.status,
'title': self.title,
'user_id': self.user_id}
def flag_for_stop(self):
"""
Flag running or failed job (anything that is not completed) with
stopping. Opaque daemon will end services flaged with 'stopping'.
"""
self.status = 'stopping'
db.session.commit()
2019-09-12 14:45:05 +00:00
def delete_job(self):
"""
Delete job with given job id from database. Also delete associated job
files. Contianers are still running for a few seconds after
the associated service has been removed. This is the reason for the
while loop. The loop checks if the file path to all the job files still
exists and removes it again and again till the container did shutdown
for good.
2019-09-16 12:09:07 +00:00
See: https://docs.docker.com/engine/swarm/swarm-tutorial/delete-service/
"""
2019-09-12 14:45:05 +00:00
logger = logging.getLogger(__name__)
delete_path = os.path.join('/mnt/opaque/', str(self.user_id), 'jobs',
str(self.id))
logger.warning('Delete path is: {}'.format(delete_path))
while os.path.exists(delete_path):
try:
shutil.rmtree(delete_path, ignore_errors=True)
logger.warning('Path does still exist.')
except OSError:
pass
2019-09-12 14:45:05 +00:00
db.session.delete(self)
db.session.commit()
2019-08-05 14:45:38 +00:00
2019-10-16 14:52:05 +00:00
class CorpusFile(db.Model):
"""
Class to define Files.
"""
__tablename__ = 'corpus_files'
# Primary key
id = db.Column(db.Integer, primary_key=True)
author = db.Column(db.String(64))
dir = db.Column(db.String(255))
filename = db.Column(db.String(255))
publishing_year = db.Column(db.Integer)
title = db.Column(db.String(64))
2019-10-16 14:52:05 +00:00
corpus_id = db.Column(db.Integer, db.ForeignKey('corpora.id'))
2019-10-30 07:28:52 +00:00
def delete(self):
logger = logging.getLogger(__name__)
2019-10-30 08:02:45 +00:00
logger.warning('Called CorpusFile.delete')
2019-10-30 07:28:52 +00:00
path = os.path.join(current_app.config['OPAQUE_STORAGE_DIRECTORY'],
self.dir,
self.filename)
try:
os.remove(path)
except:
logger.warning('[ERROR] CorpusFile.delete')
return
db.session.delete(self)
db.session.commit()
2019-10-16 14:52:05 +00:00
2019-08-06 10:06:41 +00:00
class Corpus(db.Model):
"""
Class to define a corpus.
"""
__tablename__ = 'corpora'
# Primary key
id = db.Column(db.Integer, primary_key=True)
2019-08-06 13:41:07 +00:00
creation_date = db.Column(db.DateTime(), default=datetime.utcnow)
description = db.Column(db.String(255))
2019-08-06 10:06:41 +00:00
title = db.Column(db.String(32))
user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
2019-10-16 14:52:05 +00:00
# Relationships
files = db.relationship('CorpusFile',
2019-10-17 11:32:22 +00:00
backref='corpus',
2019-10-16 14:52:05 +00:00
lazy='dynamic',
cascade='save-update, merge, delete')
2019-08-06 10:06:41 +00:00
def __init__(self, **kwargs):
2019-08-06 11:25:27 +00:00
super(Corpus, self).__init__(**kwargs)
2019-08-06 10:06:41 +00:00
def __repr__(self):
"""
String representation of the corpus. For human readability.
"""
return '<Corpus %r>' % self.title
def to_dict(self):
return {'id': self.id,
2019-08-23 13:05:01 +00:00
'creation_date': self.creation_date.timestamp(),
'description': self.description,
'title': self.title,
'user_id': self.user_id}
2019-08-06 10:06:41 +00:00
2019-10-30 07:28:52 +00:00
def delete(self):
logger = logging.getLogger(__name__)
2019-10-30 08:02:45 +00:00
logger.warning('Called Corpus.delete')
2019-10-30 07:28:52 +00:00
for corpus_file in self.files:
corpus_file.delete()
2019-10-30 08:02:45 +00:00
logger.warning('bis hierhin und nicht weiter')
logger.warning('base_dir: {}'.format(current_app.config['OPAQUE_STORAGE_DIRECTORY']))
logger.warning('user_id: {}'.format(self.user_id))
logger.warning('id: {}'.format(self.id))
2019-10-30 07:28:52 +00:00
path = os.path.join(current_app.config['OPAQUE_STORAGE_DIRECTORY'],
2019-10-30 12:12:31 +00:00
str(self.user_id),
2019-10-30 07:28:52 +00:00
'corpora',
2019-10-30 12:12:31 +00:00
str(self.id))
2019-10-30 08:02:45 +00:00
logger.warning(path)
2019-10-30 07:28:52 +00:00
try:
2019-10-30 08:02:45 +00:00
logger.warning('Try to remove {}'.format(path))
2019-10-30 07:28:52 +00:00
shutil.rmtree(path)
except:
logger.warning('[ERROR] Corpus.delete')
return
db.session.delete(self)
db.session.commit()
2019-08-22 07:35:23 +00:00
2019-08-06 09:47:04 +00:00
'''
' Flask-Login is told to use the applications custom anonymous user by setting
' its class in the login_manager.anonymous_user attribute.
'''
login_manager.anonymous_user = AnonymousUser
@login_manager.user_loader
def load_user(user_id):
return User.query.get(int(user_id))