# Minimal linux/arm64 runtime for the staged in-VM agent.
#
# Base = ubuntu:22.04 — binary-verified: the Cowork system prompt says the guest is
# "a lightweight Linux VM (Ubuntu 22)". Matching the distro matters (python 3.10, glibc,
# coreutils) so skills behave as they would in the real VM.
#
# IMPORTANT: this image contains NO Anthropic binary. The `claude` agent is the
# Linux/arm64 build that Claude Desktop stages on the user's own machine
# (~/Library/Application Support/Claude/claude-code-vm/<ver>/claude); the harness
# BIND-MOUNTS it into the container at run time (see src/runtime/container.ts).
#
#   docker build --platform linux/arm64 -t cowork-agent-base:2 -f docker/Dockerfile.agent .
#   # full-parity (OCR/Office/markitdown/opencv/table-extraction) image:
#   docker build --platform linux/arm64 --build-arg COWORK_FULL_PARITY=1 -t cowork-agent-full:2 -f docker/Dockerfile.agent .
#
# This image MIRRORS the real Cowork rootfs's preinstalled toolchain (binary-verified by
# mounting ~/.../vm_bundles/claudevm.bundle/rootfs.img): Node 22.22.3, the python
# document/data stack (openpyxl/pandas/numpy/pdfplumber/python-docx/matplotlib/…), and the
# node doc-gen globals. NOTE on openpyxl: pypi is off the runtime egress allowlist (so a skill
# can't `pip install` it at run time — faithful), BUT the real rootfs SHIPS it preinstalled —
# so omitting it here was an infidelity (a false negative). We preinstall the rootfs stack,
# pinned to the rootfs's versions.

# Base pinned by DIGEST, not by the floating `22.04` tag. This Dockerfile has no COPY/ADD — every byte
# comes from the base plus apt and pip — so with a floating base, rebuilding an unchanged recipe produced
# a different image, and "the recipe didn't change" said nothing about the contents. The digest below is
# the multi-arch INDEX digest (`docker buildx imagetools inspect ubuntu:22.04 --format
# '{{.Manifest.Digest}}'`); `--platform` selects the arm64 manifest from it. Refresh it DELIBERATELY (to
# pick up base security updates), never incidentally — a bump here changes every layer beneath.
FROM --platform=linux/arm64 ubuntu:22.04@sha256:3b06811b2afd352be909dd088a004166d665dc76d38b13eada33522a9d915c6f

ENV DEBIAN_FRONTEND=noninteractive

# Layer split: Layer A (this default) covers the document/data stack
# ~90% of skills use; Layer B (heavy: OCR/LibreOffice/opencv/onnxruntime) is opt-in.
ARG COWORK_FULL_PARITY=0

# --- Layer A: base toolchain + the lightweight apt deps the core python stack needs ---
RUN apt-get update && apt-get install -y --no-install-recommends \
      ca-certificates git ripgrep curl jq gnupg \
      python3 python3-pip python3-venv \
      poppler-utils ghostscript graphviz pandoc \
      libcairo2 libpango-1.0-0 libgl1 libglib2.0-0 libgdk-pixbuf-2.0-0 \
      libxml2 libxslt1.1 libmagic1 \
      fonts-dejavu-core fonts-liberation \
      ruby ffmpeg qpdf \
    && rm -rf /var/lib/apt/lists/*

# --- Node 22.22.3 via NodeSource (the real rootfs ships 22.22.3-1nodesource1; ubuntu's apt
#     `nodejs` is v12 — a fidelity gap). The `nodistro` repo serves the exact patch today; if
#     it rotates out, drop the `=22.22.3-*` pin to the nearest 22.x (functionally faithful). ---
RUN curl -fsSL https://deb.nodesource.com/setup_22.x | bash - \
    && (apt-get install -y "nodejs=22.22.3-1nodesource1" || apt-get install -y nodejs) \
    && rm -rf /var/lib/apt/lists/*

# --- Layer A python stack (pinned to the real rootfs). Upgrade pip first to match the rootfs's pip 25.x
#     (stock jammy pip 22.0 lacks --break-system-packages); PEP-668 → --break-system-packages is faithful
#     (the rootfs's pip is PEP-668-restricted) and a harmless no-op if no marker is present. All pins have
#     cp310 aarch64 wheels. ---
RUN python3 -m pip install --no-cache-dir --upgrade pip
RUN pip3 install --break-system-packages --no-cache-dir \
      numpy==2.2.6 pandas==2.3.3 \
      openpyxl==3.1.5 et_xmlfile==2.0.0 xlsxwriter==3.2.9 \
      python-docx==1.2.0 python-pptx==1.0.2 odfpy==1.4.1 \
      pdfplumber==0.11.9 pypdf==6.13.1 pdfminer.six==20251230 pikepdf==10.8.0 \
      pdf2image==1.17.0 img2pdf==0.6.3 pypdfium2==5.9.0 playa-pdf==1.1.0 \
      matplotlib==3.10.9 seaborn==0.13.2 pillow==12.2.0 reportlab==4.5.1 \
      lxml==6.1.1 beautifulsoup4==4.15.0 markdownify==1.2.2 marko==2.2.3 mistune==3.2.1 \
      tabulate==0.10.0 jsonschema==3.2.0 python-dotenv==1.2.2 pyyaml==6.0.3 requests==2.34.2 \
      graphviz==0.21 sympy==1.14.0 python-magic==0.4.24 Markdown==3.10.2 Jinja2==3.0.3

# --- Node doc-gen globals (the real rootfs installs these into a shared global prefix on NODE_PATH;
#     @anthropic-ai/sandbox-runtime is intentionally NOT mirrored — it's infra the harness emulates). ---
ENV NPM_CONFIG_PREFIX=/usr/local/lib/node_modules_global
ENV NODE_PATH=/usr/local/lib/node_modules_global/lib/node_modules
ENV PATH=/usr/local/lib/node_modules_global/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin
RUN npm install -g \
      docx@9.7.1 graphviz@0.0.9 markdown-toc@1.2.0 marked@18.0.5 \
      pdf-lib@1.17.1 pdfjs-dist@6.0.227 pptxgenjs@4.0.1 sharp@0.34.5 \
      ts-node@10.9.2 tsx@4.22.4 typescript@6.0.3 \
    && npm cache clean --force

# --- Locale: the real rootfs is C.UTF-8 (NOT en_US.UTF-8, which isn't generated). C.UTF-8 is
#     glibc-builtin on jammy (no locale-gen needed). Reproduce /etc/default/locale + the login
#     fix script, and set ENV so the effective preferred encoding is UTF-8 regardless of shell. ---
RUN printf 'LANG=C.UTF-8\n' > /etc/default/locale \
    && printf '#!/bin/sh\nexport LANG=C.UTF-8\n' > /etc/profile.d/01-locale-fix.sh \
    && chmod 0644 /etc/profile.d/01-locale-fix.sh
ENV LANG=C.UTF-8

# --- Sandbox env markers the real rootfs sets ---
ENV IS_SANDBOX=yes
ENV PYTHONUNBUFFERED=1
ENV VM_IMAGE_BUILD=2

# --- Layer B (opt-in): OCR / LibreOffice / opencv / onnxruntime / table-extraction.
#     tesseract ships ONLY eng+osd in the real rootfs — do NOT add language packs (that would
#     be an infidelity AND bloat). LibreOffice is the bulk of the image size. ---
RUN if [ "$COWORK_FULL_PARITY" = "1" ]; then \
      apt-get update && apt-get install -y --no-install-recommends \
        tesseract-ocr \
        imagemagick libmagickwand-6.q16-6 \
        default-jre-headless \
        libreoffice-calc libreoffice-writer libreoffice-impress libreoffice-core \
        fonts-noto-core fonts-noto-cjk fonts-ipafont-gothic fonts-ipafont-mincho \
        fonts-wqy-zenhei fonts-freefont-ttf \
      && rm -rf /var/lib/apt/lists/* \
      && pip3 install --break-system-packages --no-cache-dir \
        onnxruntime==1.23.2 magika==0.6.3 markitdown==0.1.6 \
        opencv-python==4.13.0.92 opencv-python-headless==4.13.0.92 \
        Wand==0.7.1 pytesseract==0.3.13 camelot-py==2.0.0 tabula-py==2.10.0 \
        unoserver==3.6 pyoo==1.4 ; \
    fi

# --- Run as uid 1000 `ubuntu` (the rootfs image's static /etc/passwd user; the ubuntu:22.04
#     *docker* base has no uid-1000 user, so we create it). NB this is a KNOWN divergence: at
#     runtime production's coworkd useradd's a dedicated per-session account (name = session slug,
#     uid=gid sequential >=1014, HOME=/sessions/<slug>) and spawns the agent as it — see
#     docs/fidelity-gaps.md "Guest runtime identity". Privileged steps above run as root; USER is
#     set LAST. ---
RUN useradd -u 1000 -m -s /bin/bash ubuntu \
    && mkdir -p /workspace /opt/cowork/cli-wrapper /smol/bin \
    && chown 1000:1000 /workspace \
    && printf 'prefix=/usr/local/lib/node_modules_global\ncache=/tmp/.npm\n' > /tmp/.npmrc \
    && chown 1000:1000 /tmp/.npmrc

# --- OCI image metadata (drives the GHCR package page: repo link, description, license, provenance).
#     Placed LAST so editing a label never invalidates the expensive toolchain layers above. The publish
#     workflow overrides title/description per variant (base vs full) with `docker build --label`. ---
LABEL org.opencontainers.image.title="cowork-agent-base" \
      org.opencontainers.image.description="Sandbox runtime image for cowork-harness: a fidelity-matched Ubuntu 22.04 (linux/arm64) mirror of Claude Cowork's guest rootfs toolchain (Node 22 + the python document/data stack). Contains NO Anthropic binary — the agent is bind-mounted from the user's own Claude Desktop install at run time. Full-parity variant (OCR/Office/opencv) builds from the same Dockerfile with --build-arg COWORK_FULL_PARITY=1." \
      org.opencontainers.image.source="https://github.com/yaniv-golan/cowork-harness" \
      org.opencontainers.image.documentation="https://github.com/yaniv-golan/cowork-harness#readme" \
      org.opencontainers.image.licenses="MIT"

# cwd is /sessions/<id>; the whole session root is bind-mounted writable at run time.
WORKDIR /sessions/local

USER ubuntu

# The agent binary is mounted at /usr/local/bin/claude at run time; no ENTRYPOINT —
# the Controller invokes it and drives stream-json over stdio.