Files
aurum/backend/app/models/crypto.py
ZProger e8cd536584 feat: full CoinMarketCap-style crypto portfolio (buy/sell log, 1h/24h/7d%, P&L)
Replaces the plain quantity field with a CryptoTransaction log (same
"record events, derive the total" shape as Goal/GoalContribution): quantity
and average buy price are now computed via weighted-average cost — a buy
blends into the running average, a sell reduces quantity but leaves the
average cost of what's still held unchanged.

Switches price fetching from /simple/price to CoinGecko's /coins/markets
(still one batched call for every coin), which also returns 1h/24h/7d %
change in the same request. Old PATCH /crypto/holdings/{id} is replaced by
POST .../transactions (buy/sell, rejects selling more than held), GET
.../transactions (history), and DELETE /crypto/transactions/{id}.

Frontend: CryptoHoldingsTable replaces the simple list with a real table
(Name/Price/1h%/24h%/7d%/Holdings/Avg Buy Price/Profit-Loss/Actions),
horizontally scrollable on mobile rather than dropping columns. New
CryptoTransactionModal (buy/sell) and CryptoTransactionHistoryModal.

Backend: 79/79 tests passing (13 in test_crypto.py, up from 9). Also
verified live end-to-end against the real CoinGecko API with the user's own
key — create, buy, sell, oversell rejection, history, manual refresh, all
against real BTC/ETH prices in EUR — test data cleaned up after.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01855PiW5Wd8n4aRW4n9kByY
2026-08-30 04:25:03 +03:00

98 lines
5.0 KiB
Python

"""Crypto holdings: a thin layer on top of the existing Asset/AssetValuation
engine (see models/asset.py) rather than a parallel system. Net worth's
number for a holding still lives as ordinary AssetValuation rows, kept
current by services/crypto_service.py — Net Worth requires zero changes to
treat a crypto holding like any other manually tracked asset.
Quantity and average buy price are NOT stored — they're computed from the
CryptoTransaction log (same "you record it, we derive it" shape as
Goal/GoalContribution), because a sell doesn't just subtract quantity, it
also has to leave the average cost basis of what's still held unchanged
(see services/crypto_service.py's _compute_position).
"""
from datetime import date as date_
from datetime import datetime
from sqlalchemy import Date, DateTime, Enum, ForeignKey, Numeric, String, Text
from sqlalchemy.orm import Mapped, mapped_column, relationship
from app.db.base import Base
from app.models.asset import Asset
from app.models.enums import CryptoTransactionType
from app.models.mixins import TimestampMixin
class CryptoHolding(Base, TimestampMixin):
"""1:1 extension of an Asset (asset_class=CRYPTO) — asset_id doubles as
this table's own primary key rather than a separate id+unique
constraint, so the relationship can never drift into a 1:many by
accident. Deleting the Asset (the existing DELETE /assets/{id}) cascades
here too, and cascades again to every CryptoTransaction — there's no
separate "delete the whole holding" endpoint."""
__tablename__ = "crypto_holdings"
asset_id: Mapped[int] = mapped_column(ForeignKey("assets.id", ondelete="CASCADE"), primary_key=True)
# CoinGecko's stable coin id (e.g. "bitcoin") — NOT the ticker symbol,
# which collides across unrelated coins (CoinGecko lists dozens of
# differently-named coins that all use e.g. "SOL"-like tickers).
coingecko_id: Mapped[str] = mapped_column(String(100), nullable=False)
# Denormalized from CoinGecko at add-time so the holdings list can
# render a ticker/name/logo without an extra lookup per row.
symbol: Mapped[str] = mapped_column(String(20), nullable=False)
name: Mapped[str] = mapped_column(String(150), nullable=False)
thumb_url: Mapped[str | None] = mapped_column(String(500), nullable=True)
# Cached from the last successful sync (services/crypto_service.py) —
# in the app's display currency (see AppSettings.currency). Null until
# the first sync completes.
last_price: Mapped[Numeric | None] = mapped_column(Numeric(38, 18), nullable=True)
price_change_1h: Mapped[Numeric | None] = mapped_column(Numeric(10, 4), nullable=True)
price_change_24h: Mapped[Numeric | None] = mapped_column(Numeric(10, 4), nullable=True)
price_change_7d: Mapped[Numeric | None] = mapped_column(Numeric(10, 4), nullable=True)
asset: Mapped["Asset"] = relationship()
transactions: Mapped[list["CryptoTransaction"]] = relationship(
back_populates="holding", cascade="all, delete-orphan", order_by="CryptoTransaction.date"
)
class CryptoTransaction(Base, TimestampMixin):
"""One buy or sell against a holding. Quantity held and average buy
price are both derived by replaying this log in date order — see
services/crypto_service.py's _compute_position for why a sell can't
just be "negative quantity" the way it can for a Goal contribution: it
must leave the remaining coins' average cost basis alone."""
__tablename__ = "crypto_transactions"
id: Mapped[int] = mapped_column(primary_key=True)
asset_id: Mapped[int] = mapped_column(ForeignKey("crypto_holdings.asset_id", ondelete="CASCADE"), nullable=False)
type: Mapped[CryptoTransactionType] = mapped_column(
Enum(CryptoTransactionType, name="crypto_transaction_type", native_enum=False, length=10), nullable=False
)
quantity: Mapped[Numeric] = mapped_column(Numeric(38, 18), nullable=False)
# Price paid (buy) or received (sell) per unit, in the app's display
# currency at the time of the transaction — not re-derived later, since
# what you actually paid doesn't change with today's market price.
price_per_unit: Mapped[Numeric] = mapped_column(Numeric(38, 18), nullable=False)
date: Mapped[date_] = mapped_column(Date, nullable=False)
note: Mapped[str | None] = mapped_column(Text, nullable=True)
holding: Mapped["CryptoHolding"] = relationship(back_populates="transactions")
class CryptoSyncState(Base):
"""Singleton row (id=1, same get-or-create pattern as AppSettings) — the
single global timestamp of the last successful CoinGecko price refresh.
One timestamp for every holding, not one per holding: a refresh always
re-prices every tracked coin in a single batched request (see
services/crypto_service.py), so there's only ever one "last synced"
moment to track."""
__tablename__ = "crypto_sync_state"
id: Mapped[int] = mapped_column(primary_key=True)
last_synced_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)