Requires Python 3.12 or newer and an internet connection (market data is downloaded from Binance on demand — nothing large ships with the package).
Check what you have:
python --versionIf that prints 3.11 or lower, or "command not found", install Python from python.org/downloads first. On Windows, tick "Add python.exe to PATH" in the installer.
pip install martex-quantThat gives you the martex-quant command. Verify it:
martex-quant doctorUse the module form, which always works:
python -m pip install martex-quant
python -m martex_quant.cli doctorOn macOS and Linux you may need python3 instead of python.
Every release attaches a .whl (wheel) and a .tar.gz (source archive) to
its GitHub Releases page.
Download the .whl and install it directly — useful for an offline machine
or when you want to pin an exact build:
pip install ./martex_quant-1.0.0-py3-none-any.whlDo this if you want to change the code, add a strategy, or run the test suite.
git clone https://github.com/martex-dev/martex-quant.git
cd martex-quant
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -e .
pip install -r requirements-dev.txtConfirm the checks pass before you change anything. Run them one per line:
Windows PowerShell has no && operator, and chaining them there is a parser
error that silently runs nothing.
pytest
ruff check .
mypyWhichever option you pick, installing into a virtual environment rather than your system Python keeps this project's dependencies from colliding with anything else:
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install martex-quantYou must activate that environment (source .venv/bin/activate) in each new
terminal before martex-quant is on your PATH.
The core install deliberately stays small. Two extras exist:
pip install "martex-quant[research]" # TensorFlow — only for the TSLA CNN study
pip install "martex-quant[mt5]" # MetaTrader 5 broker adapter (Windows only)You do not need either to run the data pipeline, backtester, Monte Carlo, paper trading, or dashboard.
martex-quant init my-lab
cd my-lab
martex-quant quickstartinit creates the workspace — where your data lake, paper-trading records,
and a copy of the research corpus live. quickstart downloads real market
data, backtests it honestly, and explains the result.
Full command reference: USAGE.md.
martex-quant: command not found after installing.
Your virtual environment is not activated, or pip installed to a directory
that is not on your PATH. python -m martex_quant.cli always works as a
substitute.
Data pull fails with a connection or 451 error. Binance restricts access from some regions and blocks some cloud IP ranges. The data pipeline needs it; everything downstream needs data. A VPN or a different machine is the usual fix.
martex-quant doctor says the research corpus is missing.
You are running from a source tree that was never installed. Run
pip install -e . from the repository root.
Windows shows garbled characters in output.
Set PYTHONIOENCODING=utf-8 in your environment. Windows Terminal handles
this better than the legacy console.
A command fails and you cannot tell why.
Run martex-quant doctor first — it checks the install, the dependencies, the
corpus, and the workspace, and tells you the next command to run.