mirror of
https://github.com/lrnselfreliance/wrolpi.git
synced 2026-10-26 10:51:36 -04:00
Python 3.11 is now required, old versions were causing unpredictability in tests. (Sanic does not yet support 3.12) Sanic has been upgraded to 23.6.0, which is the latest version that avoids this bug: https://github.com/sanic-org/sanic/issues/2921 New strategy for multiprocessing is to create all multiprocessing tools in one process, then fork to other processes. The previous strategy was to declare multiprocessing tools at the top of every file, or wherever they were needed at import/creation. Now all multiprocessing tools are attached to the app.shared_ctx. This means `api_app` is imported in many, many places. This forced a change in how the DownloadManager works. Previously, it would continually run download workers which would pull downloads from a multiprocessing.Queue. Now, a single worker checks for new downloads and sends a Sanic signal. Flags have been reworked to use the `api_app`. I removed the `which` flag functionality because the `which` are called at import and needed their own multiprocessing.Event.
210 lines
6.3 KiB
Python
210 lines
6.3 KiB
Python
from decimal import Decimal
|
|
from functools import partial
|
|
from pathlib import Path
|
|
from typing import List, Tuple
|
|
|
|
from pint import Quantity
|
|
|
|
from wrolpi import before_startup
|
|
from wrolpi.common import logger, Base, ConfigFile
|
|
from wrolpi.db import get_db_session
|
|
from .errors import NoInventories, InventoriesVersionMismatch
|
|
from .inventory import get_items, get_inventories, increment_inventories_version, \
|
|
get_inventories_version, get_unit_registry
|
|
from .models import Inventory, Item
|
|
|
|
MY_DIR: Path = Path(__file__).parent
|
|
|
|
logger = logger.getChild(__name__)
|
|
|
|
|
|
def sum_by_key(items: List[Base], key: callable):
|
|
"""
|
|
Sum the total size of each item by the provided key function. Returns a dict containing the key, and the total_size
|
|
for that key. Combine total of like units. This means ounces and pounds will be in the same total.
|
|
"""
|
|
unit_registry = get_unit_registry()
|
|
|
|
summed = dict()
|
|
for item in items:
|
|
k = key(item)
|
|
item_size, count, unit = item.item_size, item.count, unit_registry(item.unit)
|
|
key_dim = (k, unit.dimensionality)
|
|
|
|
total_size = item_size * count * unit
|
|
|
|
try:
|
|
summed[key_dim] += total_size
|
|
except KeyError:
|
|
summed[key_dim] = total_size
|
|
|
|
summed = {k: compact_unit(v) for k, v in summed.items()}
|
|
return summed
|
|
|
|
|
|
def get_inventory_by_keys(keys: Tuple, inventory_id: int):
|
|
items = get_items(inventory_id)
|
|
|
|
summed = sum_by_key(items, lambda i: tuple(getattr(i, k) or '' for k in keys))
|
|
|
|
inventory = []
|
|
for key, quantity in sorted(summed.items(), key=lambda i: i[0][0]):
|
|
quantity = cleanup_quantity(quantity)
|
|
total_size, units = quantity.to_tuple()
|
|
unit = units[0][0]
|
|
|
|
d = dict(total_size=total_size, unit=unit)
|
|
d.update(dict(zip(keys, key[0])))
|
|
inventory.append(d)
|
|
|
|
return inventory
|
|
|
|
|
|
get_inventory_by_category = partial(get_inventory_by_keys, ('category',))
|
|
get_inventory_by_subcategory = partial(get_inventory_by_keys, ('category', 'subcategory'))
|
|
get_inventory_by_name = partial(get_inventory_by_keys, ('brand', 'name'))
|
|
INVENTORY_UNITS: dict = None
|
|
UNIT_PRECISION = 5
|
|
|
|
|
|
def initialize_inventory_units() -> dict:
|
|
global INVENTORY_UNITS
|
|
if INVENTORY_UNITS:
|
|
return INVENTORY_UNITS
|
|
|
|
unit_registry = get_unit_registry()
|
|
INVENTORY_UNITS = {
|
|
('ounce', 1): (16, unit_registry.pound),
|
|
('pound', 1): (2000, unit_registry.ton),
|
|
}
|
|
|
|
|
|
def compact_unit(quantity: Quantity) -> Quantity:
|
|
"""
|
|
Convert a Quantity to it's more readable format. Such as 2000 pounds to 1 ton.
|
|
"""
|
|
initialize_inventory_units()
|
|
|
|
number, (units,) = quantity.to_tuple()
|
|
next_unit = INVENTORY_UNITS.get(units)
|
|
if not next_unit:
|
|
# No units after this one, return as is.
|
|
return round(quantity, UNIT_PRECISION)
|
|
|
|
max_quantity, next_unit = next_unit
|
|
if number >= max_quantity:
|
|
# Number is too high for this unit, move to the next unit.
|
|
return compact_unit(quantity.to(next_unit))
|
|
|
|
# No units were necessary, return as is.
|
|
quantity = round(quantity, UNIT_PRECISION)
|
|
return quantity
|
|
|
|
|
|
def quantity_to_tuple(quantity: Quantity) -> Tuple[Decimal, Quantity]:
|
|
unit_registry = get_unit_registry()
|
|
decimal, (units,) = quantity.to_tuple()
|
|
unit, _ = units
|
|
quantity, unit = round(decimal, UNIT_PRECISION), unit_registry(unit)
|
|
return quantity, unit
|
|
|
|
|
|
def cleanup_quantity(quantity: Quantity) -> Quantity:
|
|
"""Remove trailing zeros from a Quantity."""
|
|
num, unit = quantity_to_tuple(quantity)
|
|
num = round(num, UNIT_PRECISION)
|
|
num = str(num).rstrip('0').rstrip('.')
|
|
return Decimal(num) * unit
|
|
|
|
|
|
class InventoriesConfig(ConfigFile):
|
|
file_name = 'inventories.yaml'
|
|
default_config = dict(
|
|
inventories=[],
|
|
version=1,
|
|
)
|
|
|
|
@property
|
|
def inventories(self) -> dict:
|
|
return self._config['inventories']
|
|
|
|
@inventories.setter
|
|
def inventories(self, value: dict):
|
|
self.update({'inventories': value})
|
|
|
|
@property
|
|
def version(self) -> int:
|
|
return self._config['version']
|
|
|
|
@version.setter
|
|
def version(self, value: int):
|
|
self.update({'version': value})
|
|
|
|
|
|
INVENTORIES_CONFIG: InventoriesConfig = InventoriesConfig()
|
|
TEST_INVENTORIES_CONFIG: InventoriesConfig = None
|
|
|
|
|
|
def get_inventories_config():
|
|
global TEST_INVENTORIES_CONFIG
|
|
if isinstance(TEST_INVENTORIES_CONFIG, ConfigFile):
|
|
return TEST_INVENTORIES_CONFIG
|
|
|
|
global INVENTORIES_CONFIG
|
|
return INVENTORIES_CONFIG
|
|
|
|
|
|
def set_test_inventories_config(enabled: bool):
|
|
global TEST_INVENTORIES_CONFIG
|
|
if enabled:
|
|
TEST_INVENTORIES_CONFIG = InventoriesConfig()
|
|
else:
|
|
TEST_INVENTORIES_CONFIG = None
|
|
|
|
|
|
def save_inventories_file():
|
|
"""Write all inventories and their respective items to a WROLPi Config file."""
|
|
config = get_inventories_config()
|
|
|
|
inventories = []
|
|
for inventory in get_inventories():
|
|
inventories.append(inventory.dict())
|
|
|
|
if not inventories:
|
|
raise NoInventories('No Inventories are in the database!')
|
|
|
|
with increment_inventories_version() as version:
|
|
if config.version and config.version > version:
|
|
raise InventoriesVersionMismatch(
|
|
f'Inventories config version is {config.version} but DB version is {get_inventories_version()}')
|
|
|
|
config.inventories = inventories
|
|
config.version = version
|
|
config.save()
|
|
|
|
|
|
@before_startup
|
|
def import_inventories_file():
|
|
config = get_inventories_config()
|
|
|
|
inventories = get_inventories()
|
|
inventories_names = {i.name for i in inventories}
|
|
new_inventories = [i for i in config.inventories if i['name'] not in inventories_names]
|
|
with get_db_session(commit=True) as session:
|
|
for inventory in new_inventories:
|
|
items = inventory['items']
|
|
inventory = Inventory(
|
|
name=inventory['name'],
|
|
created_at=inventory['created_at'],
|
|
deleted_at=inventory['deleted_at'],
|
|
)
|
|
session.add(inventory)
|
|
# Get the Inventory from the DB so we can use it's ID.
|
|
session.flush()
|
|
session.refresh(inventory)
|
|
|
|
for item in items:
|
|
del item['inventory_id']
|
|
item = Item(inventory_id=inventory.id, **item)
|
|
item.inventory_id = inventory.id
|
|
session.add(item)
|