roboquant 2.1.3__tar.gz → 2.2.0__tar.gz

{roboquant-2.1.3 → roboquant-2.2.0}/.github/workflows/verify.yml

@@ -8,6 +8,9 @@ on:
   pull_request:
     branches: [ "main" ]
 
+env:
+  PYRIGHT_PYTHON_FORCE_VERSION: latest
+
 jobs:
   verify:
 

{roboquant-2.1.3 → roboquant-2.2.0}/LICENSE

@@ -186,7 +186,7 @@
       same "printed page" as the copyright notice for easier
       identification within third-party archives.
 
-   Copyright 2024, 2025 Neural Layer
+   Copyright 2024, 2025, 2026 Neural Layer
 
    Licensed under the Apache License, Version 2.0 (the "License");
    you may not use this file except in compliance with the License.

{roboquant-2.1.3 → roboquant-2.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: roboquant
-Version: 2.1.3
+Version: 2.2.0
 Summary: A fast algorithmic trading platform with support for AI based strategies
 Project-URL: Homepage, https://roboquant.org
 Project-URL: Repository, https://github.com/neurallayer/roboquant.py.git
@@ -43,10 +43,10 @@ Description-Content-Type: text/markdown
 
 # ![roboquant logo](https://github.com/neurallayer/roboquant.py/raw/main/docs/roboquant_header.png)
 
-![PyPI - Version](https://img.shields.io/pypi/v/roboquant)
-![PyPI - License](https://img.shields.io/pypi/l/roboquant)
-![PyPI - Python Version](https://img.shields.io/pypi/pyversions/roboquant)
-![PyPI - Status](https://img.shields.io/pypi/status/roboquant)
+![PyPI - Version](https://img.shields.io/pypi/v/roboquant?)
+![PyPI - License](https://img.shields.io/pypi/l/roboquant?)
+![PyPI - Python Version](https://img.shields.io/pypi/pyversions/roboquant?)
+![PyPI - Status](https://img.shields.io/pypi/status/roboquant?)
 ![GitHub Actions Workflow Status](https://img.shields.io/github/actions/workflow/status/neurallayer/roboquant.py/verify.yml)
 [![discord](https://img.shields.io/discord/954650958300856340?label=discord)](https://discord.com/channels/954650958300856340/954650958300856343)
 

{roboquant-2.1.3 → roboquant-2.2.0}/README.md

@@ -1,10 +1,10 @@
 
 # ![roboquant logo](https://github.com/neurallayer/roboquant.py/raw/main/docs/roboquant_header.png)
 
-![PyPI - Version](https://img.shields.io/pypi/v/roboquant)
-![PyPI - License](https://img.shields.io/pypi/l/roboquant)
-![PyPI - Python Version](https://img.shields.io/pypi/pyversions/roboquant)
-![PyPI - Status](https://img.shields.io/pypi/status/roboquant)
+![PyPI - Version](https://img.shields.io/pypi/v/roboquant?)
+![PyPI - License](https://img.shields.io/pypi/l/roboquant?)
+![PyPI - Python Version](https://img.shields.io/pypi/pyversions/roboquant?)
+![PyPI - Status](https://img.shields.io/pypi/status/roboquant?)
 ![GitHub Actions Workflow Status](https://img.shields.io/github/actions/workflow/status/neurallayer/roboquant.py/verify.yml)
 [![discord](https://img.shields.io/discord/954650958300856340?label=discord)](https://discord.com/channels/954650958300856340/954650958300856343)
 

roboquant-2.2.0/docs/DESIGN.md

@@ -0,0 +1,85 @@
+# Design Principles
+
+## Modular Pipeline Architecture
+
+The library is built around a clean separation of five orthogonal concerns:
+
+| Component | Responsibility | Has access to Account? |
+|-----------|---------------|----------------------|
+| **Feed** | Provides market data events | No |
+| **Strategy** | Generates trading signals from events | No |
+| **Trader** | Converts signals into orders (risk/sizing) | Yes |
+| **Broker** | Executes orders, maintains account state | Yes (owns Account) |
+| **Journal** | Logs/tracks every step (read-only) | Read-only snapshot |
+
+Each component is an abstract base class with pluggable implementations, making every part of the pipeline independently swappable.
+
+## The Run Loop
+
+The core of the system is the `roboquant.run()` function, which connects all components in a streaming event loop:
+
+```
+for each event in feed.play(timeframe):
+    1. broker.sync(event)                   — update account, execute fills
+    2. strategy.create_signals(event)       — generate signals from market data
+    3. trader.create_orders(signals, ...)   — apply risk rules, produce orders
+    4. broker.place_orders(orders)          — submit orders to the broker
+    5. journal.track(event, account, ...)   — record metrics (optional)
+```
+
+### Step-by-step
+
+1. **`broker.sync(event)`** — Updates the account with the latest market data. Open orders from previous steps are tested against prices and executed if conditions are met. No look-ahead bias: orders placed at time `t` only execute at time `t+1`.
+
+2. **`strategy.create_signals(event)`** — The strategy examines the event's price data and returns a list of `Signal` objects. Each signal has an asset, a rating (typically -1.0 to 1.0), and a type (`ENTRY`, `EXIT`, or `ENTRY_EXIT`). Strategies are **pure decision-makers** — they know nothing about cash, positions, or risk.
+
+3. **`trader.create_orders(signals, event, account)`** — The trader applies risk management rules (position sizing, shorting constraints, order limits) and converts signals into concrete `Order` objects. Unlike strategies, traders **have full access to the Account** (cash, positions, buying power).
+
+4. **`broker.place_orders(orders)`** — New orders are submitted to the broker. In `SimBroker`, they are stored and evaluated for execution when the next event arrives.
+
+5. **`journal.track(...)`** — Optional logging and metrics collection. Journals are passive observers that never modify state.
+
+## Minimal Backtest
+
+```python
+import roboquant as rq
+
+feed = rq.feeds.YahooFeed("JPM", "IBM", start_date="2015-01-01")
+account = rq.run(feed, rq.strategies.EMACrossover())
+print(account)
+```
+
+This works because `run()` provides sensible defaults: `SimBroker` (USD 1M deposit, 0% slippage) and `FlexTrader` (conservative position sizing).
+
+## Custom Backtest
+
+```python
+feed = rq.feeds.YahooFeed("AAPL", "MSFT", start_date="2020-01-01")
+strategy = rq.strategies.EMACrossover()
+trader = rq.traders.FlexTrader(max_order_perc=0.1, shorting=True)
+broker = rq.brokers.SimBroker(deposit=500_000)
+journal = rq.journals.MetricsJournal()
+
+account = rq.run(feed, strategy, trader=trader, broker=broker, journal=journal)
+print(account)
+print(journal.metrics())
+```
+
+## Strategy/Trader Separation
+
+This is the most important design choice in the library:
+
+- **Strategies** produce signals from market data only. They implement `create_signals(event) -> list[Signal]` and have no access to account state. This keeps them pure, testable, and reusable across any trader configuration.
+- **Traders** implement `create_orders(signals, event, account) -> list[Order]` and are responsible for risk management, position sizing, and order construction. They know nothing about indicators or market data beyond what is in the signal.
+
+Example: The same `EMACrossover` strategy can be used with a conservative trader (2% max order) in backtesting and a different trader (20% max order, shorting enabled) in live trading — without changing a line of strategy code.
+
+## Key Principles
+
+- **Event-driven streaming** — Everything is built around `Event` objects produced lazily by feeds, supporting both backtesting and live trading with the same pipeline.
+- **Broker owns the Account** — The Account is never modified directly by user code; it is always the broker's canonical view returned by `sync()`.
+- **Default-everywhere** — `roboquant.run(feed, strategy)` works out of the box with sensible defaults, making the simplest case a one-liner.
+- **Immutable core types** — `Asset`, `Signal`, and `Event` are immutable; `Order` uses `cancel()`/`modify()` returning new objects.
+- **Strategy composition** — `MultiStrategy` combines multiple strategies with configurable conflict resolution (`first`, `last`, `mean`, `none`).
+- **Pluggable pricing** — Price type strings (`"OPEN"`, `"CLOSE"`, `"HIGH"`, `"LOW"`, `"DEFAULT"`) allow strategies and traders to choose which price to use for evaluation.
+- **Multi-currency** — Built-in support via `Amount`, `Wallet`, and pluggable `CurrencyConverter` (ECB, static, one-to-one).

{roboquant-2.1.3 → roboquant-2.2.0}/roboquant/__init__.py

@@ -3,7 +3,7 @@ The `roboquant` package contains the `run` method and a number of shared classes
 like `Account`, `Asset` and `Event`.
 """
 
-__version__ = "2.1.3"
+__version__ = "2.2.0"
 
 import logging
 
@@ -12,7 +12,7 @@ from roboquant import feeds
 from roboquant import journals
 from roboquant import strategies
 from roboquant import traders
-from roboquant import ml
+
 from .account import Account, Position
 from .event import Event, PriceItem, Bar, TradePrice, Quote
 from .signal import Signal, SignalType
@@ -31,7 +31,6 @@ __all__ = [
     "journals",
     "strategies",
     "traders",
-    "ml",
     "Account",
     "Position",
     "Event",

roboquant-2.2.0/roboquant/ai/__init__.py

@@ -0,0 +1,6 @@
+"""roboquant.ai package.
+Provides a number of AI related classes and methods to support the development of AI based trading strategies.
+It relies on Stable Baselines3 for reinforcement learning and PyTorch for deep learning.
+
+The package provides also several features to support the development of AI based trading strategies.
+"""

{roboquant-2.1.3/roboquant/ml → roboquant-2.2.0/roboquant/ai}/features.py

@@ -18,7 +18,7 @@ FloatArray = NDArray[np.float32]
 class Feature(Generic[T]):
     """Base class for different types of features.
     The ones included by default are either based either an `Event` or an `Account`.
-    Typically Event features are used for input and Account features are used for label/output."""
+    Typically Event features are used for input and Account features are used for reward/label/output."""
 
     @abstractmethod
     def calc(self, value: T) -> FloatArray:

{roboquant-2.1.3/roboquant/ml → roboquant-2.2.0/roboquant/ai}/rl.py

@@ -15,7 +15,7 @@ from roboquant.brokers.simbroker import SimBroker
 from roboquant.event import Event
 from roboquant.feeds.feed import Feed
 from roboquant.journals.journal import Journal
-from roboquant.ml.features import Feature
+from roboquant.ai.features import Feature
 from roboquant.signal import Signal
 from roboquant.strategies.strategy import Strategy
 from roboquant.timeframe import Timeframe
@@ -23,7 +23,7 @@ from roboquant.traders.flextrader import FlexTrader
 from roboquant.traders.trader import Trader
 
 
-register(id="roboquant/StrategyEnv-v0", entry_point="roboquant.ml.envs:StrategyEnv")
+register(id="roboquant/StrategyEnv-v0", entry_point="roboquant.ai.envs:StrategyEnv")
 logger = logging.getLogger(__name__)
 
 

{roboquant-2.1.3/roboquant/ml → roboquant-2.2.0/roboquant/ai}/strategies.py

@@ -11,7 +11,7 @@ from torch.utils.data import DataLoader, Dataset
 
 from roboquant.asset import Asset
 from roboquant.event import Event
-from roboquant.ml.features import Feature, NormalizeFeature
+from roboquant.ai.features import Feature, NormalizeFeature
 from roboquant.signal import Signal
 from roboquant.strategies.strategy import Strategy
 

{roboquant-2.1.3 → roboquant-2.2.0}/roboquant/brokers/broker.py

@@ -148,11 +148,11 @@ class LiveBroker(Broker):
         """
         order = Order(
             asset=asset,
-            size=-Decimal(size),  # Negative size for sell orders
+            size=-Decimal(size), # Negative size for sell orders
             limit=float(limit),
             tif=tif,
         )
-        order.fill = - Decimal(fill)
+        order.fill = - Decimal(fill) # Negative fill for sell orders
         order.id = str(id)
         return order
 

roboquant-2.2.0/roboquant/crypto/__init__.py

@@ -0,0 +1,3 @@
+from roboquant.crypto import CryptoFeed, CryptoBroker
+
+__all__ = ["CryptoFeed", "CryptoBroker"]

{roboquant-2.1.3 → roboquant-2.2.0}/tests/integration/test_ccxt.py

@@ -3,7 +3,7 @@ import ccxt
 from datetime import datetime
 
 from roboquant.asset import Crypto
-from roboquant.feeds.cryptofeed import CryptoFeed
+from roboquant.crypto.cryptofeed import CryptoFeed
 from tests.common import run_price_item_feed
 
 

{roboquant-2.1.3 → roboquant-2.2.0}/tests/samples/crypto_backtest.py

@@ -4,7 +4,7 @@
 # %%
 import ccxt
 import roboquant as rq
-from roboquant.feeds.cryptofeed import CryptoFeed
+from roboquant.crypto.cryptofeed import CryptoFeed
 
 # %%
 exchange = ccxt.binance()

{roboquant-2.1.3 → roboquant-2.2.0}/tests/samples/sb3_strategy.py

@@ -3,8 +3,8 @@ from sb3_contrib import RecurrentPPO
 from sb3_contrib.common.recurrent.policies import RecurrentActorCriticPolicy
 from roboquant import run
 from roboquant.feeds.yahoo import YahooFeed
-from roboquant.ml.features import BarFeature, EquityFeature, CombinedFeature, SMAFeature, PriceFeature
-from roboquant.ml.rl import TradingEnv, SB3PolicyStrategy
+from roboquant.ai.features import BarFeature, EquityFeature, CombinedFeature, SMAFeature, PriceFeature
+from roboquant.ai.rl import TradingEnv, SB3PolicyStrategy
 
 # %%
 symbols = ["IBM", "JPM", "MSFT", "BA"]

{roboquant-2.1.3 → roboquant-2.2.0}/tests/samples/sb3_strategy_quotes.py

@@ -4,8 +4,8 @@ from sb3_contrib import RecurrentPPO
 from roboquant import run
 from roboquant.alpaca.feed import AlpacaHistoricStockFeed
 from roboquant.asset import Stock
-from roboquant.ml.features import EquityFeature, QuoteFeature
-from roboquant.ml.rl import TradingEnv, SB3PolicyStrategy
+from roboquant.ai.features import EquityFeature, QuoteFeature
+from roboquant.ai.rl import TradingEnv, SB3PolicyStrategy
 from roboquant.timeframe import Timeframe
 from dotenv import load_dotenv
 

{roboquant-2.1.3 → roboquant-2.2.0}/tests/samples/talib_feature.py

@@ -3,7 +3,7 @@
 import roboquant.ta as ta
 import roboquant as rq
 from roboquant.asset import Asset
-from roboquant.ml.features import TaFeature
+from roboquant.ai.features import TaFeature
 from roboquant.strategies.buffer import OHLCVBuffer
 
 # %%

{roboquant-2.1.3 → roboquant-2.2.0}/tests/samples/talib_strategy.py

@@ -3,7 +3,6 @@
 # from the TaLib library. The strategy combines the Relative Strength Index (RSI) and Bollinger Bands
 
 # %%
-# Make sure ta-lib 0.6.4 or higher is installed before running this sample
 import roboquant.ta as ta
 import roboquant as rq
 from roboquant.strategies import OHLCVBuffer, TaStrategy

{roboquant-2.1.3 → roboquant-2.2.0}/tests/samples/torch_lstm.py

@@ -6,8 +6,8 @@ import torch.nn.functional as F
 import roboquant as rq
 from roboquant.asset import Stock
 from roboquant.journals.basicjournal import BasicJournal
-from roboquant.ml.features import BarFeature, CombinedFeature, MaxReturnFeature, PriceFeature, SMAFeature
-from roboquant.ml.strategies import TimeSeriesStrategy, logger
+from roboquant.ai.features import BarFeature, CombinedFeature, MaxReturnFeature, PriceFeature, SMAFeature
+from roboquant.ai.strategies import TimeSeriesStrategy, logger
 
 
 # %%

{roboquant-2.1.3 → roboquant-2.2.0}/tests/samples/torch_transformer.py

@@ -6,8 +6,8 @@ from torch import nn
 import roboquant as rq
 from roboquant.asset import Stock
 from roboquant.journals.basicjournal import BasicJournal
-from roboquant.ml.features import BarFeature, CombinedFeature, MaxReturnFeature, PriceFeature, SMAFeature, DayOfMonthFeature
-from roboquant.ml.strategies import TimeSeriesStrategy, logger
+from roboquant.ai.features import BarFeature, CombinedFeature, MaxReturnFeature, PriceFeature, SMAFeature, DayOfMonthFeature
+from roboquant.ai.strategies import TimeSeriesStrategy, logger
 
 
 # %%

{roboquant-2.1.3 → roboquant-2.2.0}/tests/unit/test_features.py

@@ -4,7 +4,7 @@ import unittest
 import numpy as np
 from roboquant.event import Event
 
-from roboquant.ml.features import (
+from roboquant.ai.features import (
     CacheFeature,
     CombinedFeature,
     NormalizeFeature,

{roboquant-2.1.3 → roboquant-2.2.0}/tests/unit/test_rnn.py

@@ -6,8 +6,8 @@ import numpy as np
 
 import roboquant as rq
 from roboquant.asset import Stock
-from roboquant.ml.features import BarFeature, CombinedFeature, PriceFeature, SMAFeature
-from roboquant.ml.strategies import TimeSeriesStrategy, SequenceDataset
+from roboquant.ai.features import BarFeature, CombinedFeature, PriceFeature, SMAFeature
+from roboquant.ai.strategies import TimeSeriesStrategy, SequenceDataset
 from tests.common import get_feed