mirror of
https://github.com/bellingcat/auto-archiver-api.git
synced 2026-06-11 04:58:33 +03:00
WIP introduces new DB model for users,groups,tags
This commit is contained in:
@@ -1,22 +1,24 @@
|
||||
from functools import cache
|
||||
from sqlalchemy.orm import Session, load_only
|
||||
from loguru import logger
|
||||
from . import models, schemas
|
||||
import yaml
|
||||
|
||||
|
||||
def get_task(db: Session, task_id: str):
|
||||
return base_query(db).filter(models.Task.id == task_id).first()
|
||||
return base_query(db).filter(models.Archive.id == task_id).first()
|
||||
|
||||
def get_tasks(db: Session, skip: int = 0, limit: int = 100):
|
||||
return base_query(db).offset(skip).limit(limit).all()
|
||||
|
||||
def search_tasks_by_url(db: Session, url:str, skip: int = 0, limit: int = 100):
|
||||
return base_query(db).filter(models.Task.url.like(f'%{url}%')).offset(skip).limit(limit).all()
|
||||
return base_query(db).filter(models.Archive.url.like(f'%{url}%')).offset(skip).limit(limit).all()
|
||||
|
||||
def search_tasks_by_email(db: Session, email:str, skip: int = 0, limit: int = 100):
|
||||
return base_query(db).filter(models.Task.author==email).offset(skip).limit(limit).all()
|
||||
return base_query(db).filter(models.Archive.author.has(email=email)).offset(skip).limit(limit).all()
|
||||
|
||||
def create_task(db: Session, task: schemas.TaskCreate):
|
||||
db_task = models.Task(id=task.id, url=task.url, author=task.author, result=task.result)
|
||||
db_task = models.Archive(id=task.id, url=task.url, author=task.author, result=task.result)
|
||||
db.add(db_task)
|
||||
db.commit()
|
||||
db.refresh(db_task)
|
||||
@@ -30,7 +32,7 @@ def create_task(db: Session, task: schemas.TaskCreate):
|
||||
# return db_task is not None
|
||||
|
||||
def soft_delete_task(db: Session, task_id: str, email:str)->bool:
|
||||
db_task = db.query(models.Task).filter(models.Task.id == task_id, models.Task.author==email, models.Task.deleted==False).first()
|
||||
db_task = db.query(models.Archive).filter(models.Archive.id == task_id, models.Archive.author==email, models.Archive.deleted==False).first()
|
||||
if db_task:
|
||||
db_task.deleted = True
|
||||
db.commit()
|
||||
@@ -38,6 +40,51 @@ def soft_delete_task(db: Session, task_id: str, email:str)->bool:
|
||||
|
||||
def base_query(db:Session):
|
||||
# allow only some fields to be returned, for example author should remain hidden
|
||||
return db.query(models.Task)\
|
||||
.options(load_only(models.Task.id, models.Task.created_at, models.Task.url, models.Task.result))\
|
||||
.filter(models.Task.deleted == False)
|
||||
return db.query(models.Archive)\
|
||||
.options(load_only(models.Archive.id, models.Archive.created_at, models.Archive.url, models.Archive.result))\
|
||||
.filter(models.Archive.deleted == False)
|
||||
|
||||
@cache
|
||||
def get_group(db:Session, group_name:str)->models.Group:
|
||||
db_group = db.query(models.Group).filter(models.Group.id==group_name).first()
|
||||
if db_group is None:
|
||||
db_group = models.Group(id=group_name)
|
||||
db.add(db_group)
|
||||
return db_group
|
||||
|
||||
|
||||
def upsert_user_groups(db:Session, filename:str):
|
||||
"""
|
||||
reads the user_groups yaml file and inserts any new users, groups,
|
||||
along with new participation of users in groups
|
||||
"""
|
||||
logger.debug("Updating user-groups configuration.")
|
||||
|
||||
# read yaml safely
|
||||
with open(filename) as inf:
|
||||
try:
|
||||
user_groups_yaml = yaml.safe_load(inf)
|
||||
except yaml.YAMLError as e:
|
||||
logger.error(f"could not open user groups filename {filename}: {e}")
|
||||
raise e
|
||||
|
||||
# upserting in DB
|
||||
user_groups = user_groups_yaml.get("users", {})
|
||||
logger.debug(f"Found {len(user_groups)} users.")
|
||||
db.query(models.association_table_user_groups).delete()
|
||||
|
||||
for user_email, groups in user_groups.items():
|
||||
assert '@' in user_email, f'Invalid user email {user_email}'
|
||||
logger.info(f"email='{user_email[0:3]}...{user_email[-8:]}', {groups=}")
|
||||
db_user = db.query(models.User).filter(models.User.email==user_email).first()
|
||||
if db_user is None:
|
||||
db_user = models.User(email=user_email)
|
||||
db.add(db_user)
|
||||
if not groups: continue # avoid hanging in for x in None:
|
||||
for group in groups:
|
||||
db_group = get_group(db, group)
|
||||
db_group.users.append(db_user)
|
||||
|
||||
db.commit()
|
||||
count_user_groups = db.query(models.association_table_user_groups).count()
|
||||
logger.success(f"Completed refresh, now: {count_user_groups} user-groups relationships.")
|
||||
@@ -1,18 +1,85 @@
|
||||
from sqlalchemy import Column, String, JSON, DateTime, Boolean
|
||||
from sqlalchemy import Column, String, JSON, DateTime, Boolean, Table, ForeignKey
|
||||
from sqlalchemy.sql import func
|
||||
from sqlalchemy.orm import relationship
|
||||
import uuid
|
||||
from .database import Base
|
||||
|
||||
|
||||
class Task(Base):
|
||||
__tablename__ = "tasks"
|
||||
# class Task(Base):
|
||||
# __tablename__ = "tasks"
|
||||
|
||||
# id = Column(String, primary_key=True, index=True)
|
||||
# url = Column(String, index=True)
|
||||
# author = Column(String, index=True)
|
||||
# result = Column(JSON, default=None)
|
||||
# created_at = Column(DateTime(timezone=True), server_default=func.now())
|
||||
# # updated_at = Column(DateTime(timezone=True), onupdate=func.now())
|
||||
# deleted = Column(Boolean, default=False)
|
||||
|
||||
|
||||
# # items = relationship("Item", back_populates="owner")
|
||||
# # tags = relationship("Tag", back_populates="owner")
|
||||
|
||||
def generate_uuid():
|
||||
return str(uuid.uuid4())
|
||||
### new data model below
|
||||
|
||||
# many to many association tables
|
||||
association_table_archive_tags = Table(
|
||||
"mtm_archives_tags",
|
||||
Base.metadata,
|
||||
Column("archive_id", ForeignKey("archives.id")),
|
||||
Column("tag_id", ForeignKey("tags.id")),
|
||||
)
|
||||
association_table_user_groups = Table(
|
||||
"mtm_users_groups",
|
||||
Base.metadata,
|
||||
Column("user_id", ForeignKey("users.email")),
|
||||
Column("group_id", ForeignKey("groups.id")),
|
||||
)
|
||||
|
||||
# data model tables
|
||||
class Archive(Base):
|
||||
__tablename__ = "archives"
|
||||
|
||||
id = Column(String, primary_key=True, index=True)
|
||||
url = Column(String, index=True)
|
||||
author = Column(String, index=True)
|
||||
result = Column(JSON, default=None)
|
||||
created_at = Column(DateTime(timezone=True), server_default=func.now())
|
||||
# updated_at = Column(DateTime(timezone=True), onupdate=func.now())
|
||||
updated_at = Column(DateTime(timezone=True), onupdate=func.now())
|
||||
deleted = Column(Boolean, default=False)
|
||||
public = Column(Boolean, default=True) # if public=false, access to group and author
|
||||
|
||||
group_id = Column(String, ForeignKey("groups.id"), default=None)
|
||||
author_id = Column(String, ForeignKey("users.email"))
|
||||
|
||||
group = relationship("Group", back_populates="archives")
|
||||
tags = relationship("Tag", back_populates="archives", secondary=association_table_archive_tags)
|
||||
author = relationship("User", back_populates="archives")
|
||||
|
||||
class Tag(Base):
|
||||
__tablename__ = "tags"
|
||||
|
||||
id = Column(String, primary_key=True, index=True, default=generate_uuid)
|
||||
name = Column(String, unique=True, index=True)
|
||||
created_at = Column(DateTime(timezone=True), server_default=func.now())
|
||||
|
||||
archives = relationship("Archive", back_populates="tags", secondary=association_table_archive_tags)
|
||||
|
||||
|
||||
# items = relationship("Item", back_populates="owner")
|
||||
class User(Base):
|
||||
__tablename__ = "users"
|
||||
|
||||
email = Column(String, primary_key=True, index=True)
|
||||
is_active = Column(Boolean, default=True)
|
||||
|
||||
archives = relationship("Archive", back_populates="author")
|
||||
groups = relationship("Group", back_populates="users", secondary=association_table_user_groups)
|
||||
|
||||
class Group(Base):
|
||||
__tablename__ = "groups"
|
||||
|
||||
id = Column(String, primary_key=True, index=True)
|
||||
|
||||
archives = relationship("Archive", back_populates="group")
|
||||
users = relationship("User", back_populates="groups", secondary=association_table_user_groups)
|
||||
Reference in New Issue
Block a user