Source code for panoptes.utils.database.base
import abc
from loguru import logger
from panoptes.utils.library import load_module
from panoptes.utils.time import current_time
[docs]def create_storage_obj(collection, data, obj_id):
"""Wrap the data in a dict along with the id and a timestamp."""
return dict(_id=obj_id, data=data, type=collection, date=current_time(datetime=True))
[docs]def get_db_class(module_name='file'):
"""Load the main DB class for the module of the given name.
.. note::
This is used by the `PanDB` constructor to determine the
correct database type. Normal DB instantiation should be done
via the `PanDB()` class with the desired `db_type` parameter
set. See example in `PanDB` below.
Args:
module_name (str): Name of module, one of: `file` (default), 'memory'.
Returns:
`panoptes.utils.database.PanDB`: An instance of the db class for the correct database type.
Raises:
Exception: If an unsupported database type string is passed.
"""
class_map = {
'file': 'PanFileDB',
'memory': 'PanMemoryDB',
}
full_module_name = f'panoptes.utils.database.{module_name}'
try:
db_module = load_module(full_module_name)
return getattr(db_module, class_map[module_name])
except Exception as e:
raise Exception(f'Unsupported database type: {full_module_name}: {e!r}')
[docs]class AbstractPanDB(metaclass=abc.ABCMeta):
def __init__(self, db_name=None, **kwargs):
"""
Init base class for db instances.
Args:
db_name: Name of the database, typically 'panoptes' or 'panoptes_testing'.
"""
self.db_name = db_name
logger.info(f'Creating PanDB {self.db_name}')
[docs] @abc.abstractmethod
def insert_current(self, collection, obj, store_permanently=True): # pragma: no cover
"""Insert an object into both the `current` collection and the collection provided.
Args:
collection (str): Name of valid collection within the db.
obj (dict or str): Object to be inserted.
store_permanently (bool): Whether to also update the collection,
defaults to True.
Returns:
str: identifier of inserted record. If `store_permanently` is True, will
be the identifier of the object in the `collection`, otherwise will be the
identifier of object in the `current` collection. These may or
may not be the same.
Returns None if unable to insert into the collection.
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def insert(self, collection, obj): # pragma: no cover
"""Insert an object into the collection provided.
The `obj` to be stored in a collection should include the `type`
and `date` metadata as well as a `data` key that contains the actual
object data. If these keys are not provided then `obj` will be wrapped
in a corresponding object that does contain the metadata.
Args:
collection (str): Name of valid collection within the db.
obj (dict or str): Object to be inserted.
Returns:
str: identifier of inserted record in `collection`.
Returns None if unable to insert into the collection.
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def get_current(self, collection): # pragma: no cover
"""Returns the most current record for the given collection
Args:
collection (str): Name of the collection to get most current from
Returns:
dict|None: Current object of the collection or None.
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def find(self, collection, obj_id): # pragma: no cover
"""Find an object by it's identifier.
Args:
collection (str): Collection to search for object.
obj_id (ObjectID|str): Record identifier returned earlier by insert
or insert_current.
Returns:
dict|None: Object matching identifier or None.
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def clear_current(self, type): # pragma: no cover
"""Clear the current record of a certain type
Args:
type (str): The type of entry in the current collection that
should be cleared.
"""
raise NotImplementedError()
[docs]class PanDB(object):
"""Simple class to load the appropriate DB type based on the config.
We don't actually create instances of this class, but instead create
an instance of the 'correct' type of db.
.. doctest:
>>> from panoptes.utils.database import PanDB
>>> type(PanDB(db_type='file'))
<class 'panoptes.utils.database.file.PanFileDB'>
>>> type(PanDB('memory'))
<class 'panoptes.utils.database.memory.PanMemoryDB'>
>>> # Make a new instance of the db.
>>> memory_db = PanDB(db_type='memory')
>>> # Returns None if noe record.
>>> memory_db.get_current('safety')
>>> insert_id = memory_db.insert_current('safety', True)
>>> record = memory_db.get_current('safety')
>>> record['data']
True
"""
def __new__(cls, db_type='memory', db_name=None, *args, **kwargs):
"""Create an instance based on db_type."""
# Load the correct DB module
DatabaseModule = get_db_class(db_type)
if db_type == 'memory':
# The memory type has special setup
db_instance = DatabaseModule.get_or_create(**kwargs)
else:
db_instance = DatabaseModule(**kwargs)
return db_instance
[docs] @classmethod
def permanently_erase_database(cls,
db_type,
db_name,
storage_dir=None,
really=False,
dangerous=False,
*args, **kwargs):
"""Permanently delete the contents of the identified database."""
if not isinstance(db_name, str) or 'test' not in db_name:
raise ValueError(f'permanently_erase_database() called for non-test database {db_name!r}')
if really != 'Yes' or dangerous != 'Totally':
raise Exception('PanDB.permanently_erase_database called with invalid args!')
# Load the correct DB module and do the deletion.
get_db_class(db_type).permanently_erase_database(db_name, storage_dir=storage_dir, *args, **kwargs)