Dockerfile.custom_deployment

# syntax=docker/dockerfile:1.4
FROM python:3.12-slim AS base

WORKDIR /app
ENV PYTHONUNBUFFERED=1
ENV DEBIAN_FRONTEND=noninteractive

RUN --mount=type=cache,target=/var/cache/apt \
    apt-get update -qq && \
    apt-get install -yq --no-install-recommends \
    git \
    openssh-client \
    wget \
    gpg \
    software-properties-common \
    build-essential && \
    apt-get clean && rm -rf /var/lib/apt/lists/*

RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda.sh && \
    chmod +x ~/miniconda.sh && \
    bash ~/miniconda.sh -b -p /app/miniconda && \
    rm ~/miniconda.sh && \
    /app/miniconda/bin/conda init bash

# Set environment variables to point to conda environment
ENV VIRTUAL_ENV="/app/miniconda/bin"
ENV PATH="/app/miniconda/bin:$PATH"
ENV PYTHONPATH="/app/miniconda/lib/python3.12/site-packages:${PYTHONPATH:-}"

# Install uv & mamba
RUN pip3 install --no-cache-dir uv==0.5.21
RUN conda install -c conda-forge mamba -y

# Install R and kernels in the crow_env environment
RUN mamba install -c conda-forge -y \
    r-base=4.3.3 \
    r-recommended=4.3 \
    r-irkernel=1.3.2 \
    r-factominer=2.11 \
    r-rcolorbrewer=1.1_3 \
    r-devtools=2.4.5 \
    r-broom=1.0.7 \
    r-data.table=1.15.4 \
    r-enrichr=3.2 \
    r-factoextra=1.0.7 \
    r-ggnewscale=0.5.0 \
    r-ggrepel=0.9.6 \
    r-ggpubr=0.6.0 \
    r-ggvenn=0.1.10 \
    r-janitor=2.2.1 \
    r-multcomp=1.4_26 \
    r-matrix=1.6_5 \
    r-pheatmap=1.0.12 \
    r-tidyverse=2.0.0 \
    r-readxl=1.4.3 \
    r-reshape=0.8.9 \
    r-rstatix=0.7.2 \
    r-viridis=0.6.5 \
    udocker=1.3.17 \
    imbalanced-learn=0.13.0 \
    ipykernel=6.29.5 \
    sqlite=3.47.2 \
    anndata=0.11.1 \
    biopython=1.84 \
    datasets \
    ete3=3.1.3 \
    fcsparser=0.2.8 \
    keras=3.7.0 \
    jupyter=1.0.0 \
    matplotlib=3.10.0 \
    matplotlib-venn=1.1.1 \
    nbconvert=7.16.4 \
    numpy=1.26.4 \
    optuna=4.1.0 \
    openpyxl=3.1.5 \
    pandas=2.2.3 \
    plotly=5.24.1 \
    rpy2=3.5.11 \
    scipy=1.14.1 \
    scanpy=1.10.4 \
    seaborn=0.13.2 \
    scikit-learn=1.6.0 \
    statsmodels=0.14.4 \
    umap-learn=0.5.7

RUN python -m ipykernel install --user --name python3 --display-name "Python 3 (ipykernel)"
RUN R -e 'IRkernel::installspec(name = "R", displayname = "R (4.3.3)")'

RUN mamba install -c conda-forge -c bioconda -y \
    biokit=0.5.0 \
    gseapy=1.1.4 \
    blast=2.16.0 \
    clipkit=2.3.0 \
    clustalo=1.2.4 \
    fastqc=0.12.1 \
    iqtree=2.3.6 \
    mafft=7.526 \
    metaeuk=7.bba0d80 \
    mygene=3.2.2 \
    perl=5.32.1 \
    phykit=2.0.1 \
    pydeseq2=0.4.12 \
    spades=4.0.0 \
    trim-galore=0.6.10 \
    bioconductor-enhancedvolcano=1.20.0 \
    bioconductor-deseq2=1.42.0 \
    bioconductor-clusterprofiler=4.10.0 \
    bioconductor-org.hs.eg.db=3.18.0 \
    bioconductor-genomicranges=1.54.1 \
    bioconductor-summarizedexperiment=1.32.0 \
    bioconductor-apeglm=1.24.0 \
    bioconductor-flowcore=2.14.0

ENV UV_COMPILE_BYTECODE=1
ENV UV_LINK_MODE=copy

FROM base AS builder

ARG MODULE_NAME
ARG USE_INTERNAL_DEPS
ARG USE_GIT_FUTUREHOUSE_CLIENT


RUN mkdir -p ~/.ssh && \
    chmod 700 ~/.ssh && \
    ssh-keyscan github.com >> ~/.ssh/known_hosts && \
    printf "Host github.com\n    IdentityFile /root/.ssh/pqa_id_ed25519\n    IdentityFile /root/.ssh/aviary_id_ed25519\nHost gitlab.company.com\n    IdentityFile /root/.ssh/pqa_id_ed25519\n" > ~/.ssh/config

RUN --mount=type=cache,target=/var/cache/apt \
    apt-get update -qq && \
    apt-get install -yq --no-install-recommends \
    build-essential && \
    apt-get clean && rm -rf /var/lib/apt/lists/*

ENV VIRTUAL_ENV="/app/miniconda/bin"
ENV PATH="/app/miniconda/bin:$PATH"

COPY ./${MODULE_NAME} /app/${MODULE_NAME}

RUN mkdir -p /app/scripts
COPY ./scripts/run_crow_job.py /app/scripts/

RUN --mount=type=cache,target=/root/.cache/uv \
    --mount=type=ssh \
    --mount=type=secret,id=ssh_key,target=/root/.ssh/aviary_id_ed25519.tmp \
    if [ "$USE_INTERNAL_DEPS" = "true" ]; then \
        cp /root/.ssh/aviary_id_ed25519.tmp /root/.ssh/aviary_id_ed25519 && \
        chmod 400 /root/.ssh/aviary_id_ed25519 && \
        git clone git@github.com:Future-House/aviary-internal.git /app/aviary_internal && \
        cd /app/aviary_internal/aviary_internal && \
        uv pip install --system -e .; \
    else \
        echo 'Skipping aviary_internal install'; \
    fi && \
    if [ "$USE_GIT_FUTUREHOUSE_CLIENT" = "true" ]; then \
        git clone git@github.com:Future-House/crow-ecosystem.git /app/crow-ecosystem && \
        cd /app/crow-ecosystem/packages/futurehouse-client && \
        uv pip install --system -e .; \
    else \
        uv pip install --system futurehouse-client; \
    fi

WORKDIR /app/${MODULE_NAME}
RUN --mount=type=ssh \
    --mount=type=secret,id=pqa_ssh_key,target=/root/.ssh/pqa_id_ed25519.tmp \
    cp /root/.ssh/pqa_id_ed25519.tmp /root/.ssh/pqa_id_ed25519 && \
    chmod 400 /root/.ssh/pqa_id_ed25519 && \
    if [ -f "pyproject.toml" ]; then \
        uv pip install --system -e .; \
    elif [ -f "requirements.txt" ]; then \
        uv pip install --system -r requirements.txt; \
    else \
        echo "No pyproject.toml or requirements.txt found" && exit 1; \
    fi

RUN find /app -type l -delete && \
    rm -rf /app/.git

# Fix futurehouse_client imports by creating a symlink if we're using the git version
RUN if [ "$USE_GIT_FUTUREHOUSE_CLIENT" = "true" ]; then \
    ln -sf /app/crow-ecosystem/packages/futurehouse-client/futurehouse_client /app/miniconda/lib/python3.12/site-packages/futurehouse_client; \
    fi

FROM base AS runtime

COPY --from=builder /app/ /app/

ENV VIRTUAL_ENV="/app/miniconda/bin"
ENV PATH="/app/miniconda/bin:$PATH"
ENV PYTHONPATH="/app/miniconda/lib/python3.12/site-packages:${PYTHONPATH:-}"
CMD ["python", "scripts/run_crow_job.py"]