DockerfileLeanFoundation

#
#   LEAN Foundation Docker Container
#   Cross platform deployment for multiple brokerages
#   Intended to be used in conjunction with Dockerfile. This is just the foundation common OS+Dependencies required.
#

# Use base system for cleaning up wayward processes
FROM phusion/baseimage:jammy-1.0.1

MAINTAINER QuantConnect <contact@quantconnect.com>

# Use baseimage-docker's init system.
CMD ["/sbin/my_init"]

# Install OS Packages:
# Misc tools for running Python.NET and IB inside a headless container.
RUN apt-get update && apt-get -y install wget curl unzip \
   && apt-get install -y git bzip2 zlib1g-dev \
   xvfb libxrender1 libxtst6 libxi6 libglib2.0-dev libopenmpi-dev libstdc++6 openmpi-bin \
   pandoc libcurl4-openssl-dev libgtk2.0.0 build-essential \
   && apt-get clean && apt-get autoclean && apt-get autoremove --purge -y \
   && rm -rf /var/lib/apt/lists/*

# Install dotnet sdk & runtime
RUN add-apt-repository ppa:dotnet/backports && apt-get update && apt-get install -y dotnet-sdk-9.0 && \
    apt-get clean && apt-get autoclean && apt-get autoremove --purge -y && rm -rf /var/lib/apt/lists/*

# Set PythonDLL variable for PythonNet
ENV PYTHONNET_PYDLL="/opt/miniconda3/lib/libpython3.11.so"

# Install miniconda
ENV CONDA="Miniconda3-py311_24.9.2-0-Linux-x86_64.sh"
ENV PATH="/opt/miniconda3/bin:${PATH}"
RUN wget -q https://cdn.quantconnect.com/miniconda/${CONDA} && \
    bash ${CONDA} -b -p /opt/miniconda3 && rm -rf ${CONDA} && \
    conda config --set solver classic && \
    conda config --set auto_update_conda false

# Install java runtime for h2o lib
RUN wget https://download.oracle.com/java/17/archive/jdk-17.0.12_linux-x64_bin.deb \
    && dpkg -i jdk-17.0.12_linux-x64_bin.deb \
    && update-alternatives --install /usr/bin/java java /usr/lib/jvm/jdk-17.0.12-oracle-x64/bin/java 1 \
    && rm jdk-17.0.12_linux-x64_bin.deb

# Avoid pip install read timeouts
ENV PIP_DEFAULT_TIMEOUT=120

# Install all packages
RUN pip install --no-cache-dir      \
    cython==3.0.9                   \
    pandas==2.1.4                   \
    scipy==1.11.4                   \
    numpy==1.26.4                   \
    wrapt==1.16.0                   \
    astropy==7.0.0                  \
    beautifulsoup4==4.12.3          \
    dill==0.3.8                     \
    jsonschema==4.23.0              \
    lxml==5.3.0                     \
    msgpack==1.1.0                  \
    numba==0.59.1                   \
    xarray==2024.11.0               \
    plotly==5.24.1                  \
    jupyterlab==4.3.2               \
    ipywidgets==8.1.5               \
    jupyterlab-widgets==3.0.13      \
    tensorflow==2.18.0              \
    docutils==0.21.2                \
    cvxopt==1.3.2                   \
    gensim==4.3.3                   \
    keras==3.7.0                    \
    lightgbm==4.5.0                 \
    nltk==3.9.1                     \
    graphviz==0.20.3                \
    cmdstanpy==1.2.4                \
    copulae==0.7.9                  \
    featuretools==1.31.0            \
    PuLP==2.9.0                     \
    pymc==5.19.0                    \
    rauth==0.7.3                    \
    scikit-learn==1.4.2             \
    scikit-optimize==0.10.2         \
    aesara==2.9.4                   \
    tsfresh==0.20.2                 \
    tslearn==0.6.3                  \
    tweepy==4.14.0                  \
    PyWavelets==1.7.0               \
    umap-learn==0.5.7               \
    fastai==2.7.18                  \
    arch==7.2.0                     \
    copulas==0.12.0                 \
    creme==0.6.1                    \
    cufflinks==0.17.3               \
    gym==0.26.2                     \
    deap==1.4.1                     \
    pykalman==0.9.7                 \
    cvxpy==1.6.0                    \
    pyportfolioopt==1.5.6           \
    pmdarima==2.0.4                 \
    pyro-ppl==1.9.1                 \
    riskparityportfolio==0.6.0      \
    sklearn-json==0.1.0             \
    statsmodels==0.14.4             \
    QuantLib==1.36                  \
    xgboost==2.1.3                  \
    dtw-python==1.5.3               \
    gluonts==0.16.0                 \
    gplearn==0.4.2                  \
    jax==0.4.35                     \
    jaxlib==0.4.35                  \
    keras-rl==0.4.2                 \
    pennylane==0.39.0               \
    PennyLane-Lightning==0.39.0     \
    pennylane-qiskit==0.36.0        \
    qiskit==1.2.4                   \
    neural-tangents==0.6.5          \
    mplfinance==0.12.10b0           \
    hmmlearn==0.3.3                 \
    catboost==1.2.7                 \
    fastai2==0.0.30                 \
    scikit-tda==1.1.1               \
    ta==0.11.0                      \
    seaborn==0.13.2                 \
    optuna==4.1.0                   \
    findiff==0.10.2                 \
    sktime==0.26.0                  \
    hyperopt==0.2.7                 \
    bayesian-optimization==2.0.0    \
    pingouin==0.5.5                 \
    quantecon==0.7.2                \
    matplotlib==3.7.5               \
    sdeint==0.3.0                   \
    pandas_market_calendars==4.4.2  \
    dgl==2.1.0                      \
    ruptures==1.1.9                 \
    simpy==4.1.1                    \
    scikit-learn-extra==0.3.0       \
    ray==2.40.0                     \
    "ray[tune]"==2.40.0             \
    "ray[rllib]"==2.40.0            \
    "ray[data]"==2.40.0             \
    "ray[train]"==2.40.0            \
    fastText==0.9.3                 \
    h2o==3.46.0.6                   \
    prophet==1.1.6                  \
    torch==2.5.1                    \
    torchvision==0.20.1             \
    ax-platform==0.4.3              \
    alphalens-reloaded==0.4.5       \
    pyfolio-reloaded==0.9.8         \
    altair==5.5.0                   \
    modin==0.26.1                   \
    persim==0.3.7                   \
    ripser==0.6.10                  \
    pydmd==2024.12.1                \
    spacy==3.7.5                    \
    pandas-ta==0.3.14b              \
    pytorch-ignite==0.5.1           \
    tensorly==0.9.0                 \
    mlxtend==0.23.3                 \
    shap==0.46.0                    \
    lime==0.2.0.1                   \
    tensorflow-probability==0.25.0  \
    mpmath==1.3.0                   \
    tensortrade==1.0.3              \
    polars==1.16.0                  \
    stockstats==0.6.2               \
    autokeras==2.0.0                \
    QuantStats==0.0.64              \
    hurst==0.0.5                    \
    numerapi==2.19.1                \
    pymdptoolbox==4.0-b3            \
    panel==1.5.4                    \
    hvplot==0.11.1                  \
    line-profiler==4.2.0            \
    py-heat==0.0.6                  \
    py-heat-magic==0.0.2            \
    bokeh==3.6.2                    \
    tensorflow-decision-forests==1.11.0 \
    river==0.21.0                   \
    stumpy==1.13.0                  \
    pyvinecopulib==0.6.5            \
    ijson==3.3.0                    \
    jupyter-resource-usage==1.1.0   \
    injector==0.22.0                \
    openpyxl==3.1.5                 \
    xlrd==2.0.1                     \
    mljar-supervised==1.1.9         \
    dm-tree==0.1.8                  \
    lz4==4.3.3                      \
    ortools==9.9.3963               \
    py_vollib==1.0.1                \
    thundergbm==0.3.17              \
    yellowbrick==1.5                \
    livelossplot==0.5.5             \
    gymnasium==1.0.0                \
    interpret==0.6.7                \
    DoubleML==0.9.0                 \
    jupyter-bokeh==4.0.5            \
    imbalanced-learn==0.12.4        \
    openai==1.57.0                  \
    lazypredict==0.2.14a1           \
    darts==0.31.0                   \
    fastparquet==2024.11.0          \
    tables==3.10.1                  \
    dimod==0.12.17                  \
    dwave-samplers==1.3.0           \
    python-statemachine==2.5.0      \
    pymannkendall==1.4.3            \
    Pyomo==6.8.2                    \
    gpflow==2.9.2                   \
    pyarrow==15.0.1                 \
    dwave-ocean-sdk==8.0.1          \
    chardet==5.2.0                  \
    stable-baselines3==2.4.0        \
    Shimmy==2.0.0                   \
    pystan==3.10.0                  \
    FixedEffectModel==0.0.5         \
    transformers==4.46.3            \
    Rbeast==0.1.23                  \
    langchain==0.2.17               \
    pomegranate==1.1.1              \
    MAPIE==0.9.1                    \
    mlforecast==0.15.1              \
    tensorrt==10.7.0                \
    x-transformers==1.42.24         \
    Werkzeug==3.1.3                 \
    TPOT==0.12.2                    \
    llama-index==0.12.2             \
    mlflow==2.18.0                  \
    ngboost==0.5.1                  \
    pycaret==3.3.2                  \
    control==0.10.1                 \
    pgmpy==0.1.26                   \
    mgarch==0.3.0                   \
    jupyter-ai==2.28.2              \
    keras-tcn==3.5.0                \
    neuralprophet[live]==0.9.0      \
    Riskfolio-Lib==6.1.1            \
    fuzzy-c-means==1.7.2            \
    EMD-signal==1.6.4               \
    dask[complete]==2024.9.0        \
    nolds==0.6.1                    \
    feature-engine==1.6.2           \
    pytorch-tabnet==4.1.0           \
    opencv-contrib-python-headless==4.10.0.84 \
    POT==0.9.5                      \
    alibi-detect==0.12.0            \
    datasets==2.21.0                \
    scikeras==0.13.0                \
    accelerate==0.34.2              \
    peft==0.13.2                    \
    FlagEmbedding==1.2.11           \
    contourpy==1.3.1                \
    tensorboardX==2.6.2.2           \
    scikit-image==0.22.0            \
    scs==3.2.7                      \
    thinc==8.2.5                    \
    cesium==0.12.1                  \
    cvxportfolio==1.4.0             \
    tsfel==0.1.9                    \
    ipympl==0.9.4                   \
    PyQt6==6.7.1                    \
    nixtla==0.6.4                   \
    tigramite==5.2.6.7              \
    pytorch-forecasting==1.2.0      \
    chronos-forecasting[training]==1.4.1 \
    setuptools==73.0.1              \
    tinygrad==0.10.0

RUN conda install -c nvidia -y cuda-compiler=12.3.2 && conda clean -y --all

ENV CUDA_MODULE_LOADING=LAZY
ENV XLA_FLAGS=--xla_gpu_cuda_data_dir=/opt/miniconda3/
ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cublas/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cuda_cupti/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cuda_nvrtc/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cuda_runtime/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cudnn/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cufft/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/curand/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cusolver/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/cusparse/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/nccl/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/nvjitlink/lib/:/opt/miniconda3/lib/python3.11/site-packages/nvidia/nvtx/lib/
# reduces GPU memory usage
ENV PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True

# required for numba to work correctly
RUN ln -s /opt/miniconda3/lib/python3.11/site-packages/nvidia/cuda_runtime/lib/libcudart.so.12 /opt/miniconda3/lib/python3.11/site-packages/nvidia/cuda_runtime/lib/libcudart.so

# iisignature requires numpy to be already installed. cupy requires cuda installed
RUN pip install --no-cache-dir iisignature==0.24 cupy-cuda12x==13.3.0 mamba-ssm[causal-conv1d]==2.2.4

# TF uses this version, pytorch a slightly older (9.1) but cuDNN is backwards compatible
RUN pip install --no-cache-dir nvidia-cudnn-cu12==9.3.0.75

# Install dwave tool
RUN dwave install --all -y

# Install 'ipopt' solver for 'Pyomo'
RUN conda install -c conda-forge -y ipopt==3.14.16 coincbc==2.10.12 openmpi=5.0.6    \
    && conda clean -y --all

# Install spacy models
RUN python -m spacy download en_core_web_md && python -m spacy download en_core_web_sm

# Install PyTorch Geometric
RUN TORCH=$(python -c "import torch; print(torch.__version__)") && \
    CUDA=$(python -c "import torch; print('cu' + torch.version.cuda.replace('.', ''))") && \
    pip install --no-cache-dir -f https://pytorch-geometric.com/whl/torch-${TORCH}+${CUDA}.html \
    torch-scatter==2.1.2 torch-sparse==0.6.18 torch-cluster==1.6.3 torch-spline-conv==1.2.2 torch-geometric==2.6.1

# Install nltk data
RUN python -m nltk.downloader -d /usr/share/nltk_data punkt && \
    python -m nltk.downloader -d /usr/share/nltk_data punkt_tab && \
    python -m nltk.downloader -d /usr/share/nltk_data vader_lexicon && \
    python -m nltk.downloader -d /usr/share/nltk_data stopwords && \
    python -m nltk.downloader -d /usr/share/nltk_data wordnet

# Install Pyrb
RUN wget -q https://cdn.quantconnect.com/pyrb/pyrb-master-250054e.zip && \
    unzip -q pyrb-master-250054e.zip && cd pyrb-master && \
    pip install . && cd .. && rm -rf pyrb-master && rm pyrb-master-250054e.zip

# Install SSM
RUN wget -q https://cdn.quantconnect.com/ssm/ssm-master-646e188.zip && \
    unzip -q ssm-master-646e188.zip && cd ssm-master && \
    pip install . && cd .. && rm -rf ssm-master && rm ssm-master-646e188.zip

# Install TA-lib for python
RUN wget -q https://cdn.quantconnect.com/ta-lib/ta-lib-0.4.0-src.tar.gz && \
    tar -zxvf ta-lib-0.4.0-src.tar.gz && cd ta-lib && \
    ./configure --prefix=/usr && make && make install && \
    cd .. && rm -rf ta-lib && rm ta-lib-0.4.0-src.tar.gz && \
    pip install --no-cache-dir TA-Lib==0.5.1

# chronos-forecasting we manually copy the 'scripts' folder which holds the fine tuning tools
RUN wget -q https://cdn.quantconnect.com/chronos-forecasting/chronos-forecasting-main-133761a.zip && \
    unzip -q chronos-forecasting-main-133761a.zip && cd chronos-forecasting-main && \
    cp -r scripts /opt/miniconda3/lib/python3.11/site-packages/chronos/ && \
    cd .. && rm -rf chronos-forecasting-main && rm chronos-forecasting-main-133761a.zip

RUN echo "{\"argv\":[\"python\",\"-m\",\"ipykernel_launcher\",\"-f\",\"{connection_file}\"],\"display_name\":\"Foundation-Py-Default\",\"language\":\"python\",\"metadata\":{\"debugger\":true}}" > /opt/miniconda3/share/jupyter/kernels/python3/kernel.json

# Install wkhtmltopdf and xvfb to support HTML to PDF conversion of reports
RUN apt-get update && apt install -y xvfb wkhtmltopdf && \
    apt-get clean && apt-get autoclean && apt-get autoremove --purge -y && rm -rf /var/lib/apt/lists/*

# Install fonts for matplotlib
RUN wget -q https://cdn.quantconnect.com/fonts/foundation.zip && unzip -q foundation.zip && rm foundation.zip \
    && mv "lean fonts/"* /usr/share/fonts/truetype/ && rm -rf "lean fonts/" "__MACOSX/"

# Install IB Gateway: Installs to /root/ibgateway
RUN mkdir -p /root/ibgateway && \
    wget -q https://cdn.quantconnect.com/interactive/ibgateway-stable-standalone-linux-x64.v10.19.2a.sh && \
    chmod 777 ibgateway-stable-standalone-linux-x64.v10.19.2a.sh && \
    ./ibgateway-stable-standalone-linux-x64.v10.19.2a.sh -q -dir /root/ibgateway && \
    rm ibgateway-stable-standalone-linux-x64.v10.19.2a.sh

RUN apt-get update && apt-get install -y dotnet-sdk-6.0 && \
    apt-get clean && apt-get autoclean && apt-get autoremove --purge -y && rm -rf /var/lib/apt/lists/*

# label definitions
LABEL strict_python_version=3.11.11
LABEL python_version=3.11
LABEL target_framework=net6.0