docs: add Docker-based workflow for building documentation

[GaneshPatil7517]

This PR adds a Docker-based workflow for building DataFusion documentation, allowing contributors to build docs without installing additional host dependencies.

The container includes all required tools (Rust 1.92.0, cargo-depgraph, Sphinx, etc.) and provides a simple, reproducible build process.

This addresses #19777.

What's included:

• Dockerfile with all doc build dependencies
• docker-compose.yml for easy usage
• Updated docs/README.md with clear instructions
• Cross-platform compatibility (handles Windows/Unix line endings)

Usage:

docker-compose run --rm docs bash build.sh
# or
docker build -t datafusion-docs -f docs/Dockerfile .
docker run --rm -v $(pwd):/work datafusion-docs bash build.sh

[Copilot]

Pull request overview

This PR adds Docker-based infrastructure to enable building DataFusion documentation in a containerized environment without requiring manual installation of dependencies (Rust, cargo-depgraph, Sphinx, etc.).

Changes:

• Adds Dockerfile with all necessary documentation build dependencies (Rust 1.92.0, Python packages, cargo-depgraph, graphviz)
• Updates docker-compose.yml to include a docs service for easy documentation building
• Updates docs/README.md with Docker-based workflow instructions as the recommended approach

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 6 comments.

File	Description
docs/Dockerfile	Defines container image with all doc build dependencies, handles line endings for cross-platform compatibility
docker-compose.yml	Adds docs service configuration for simplified Docker workflow
docs/README.md	Documents Docker-based build option as recommended approach, maintains existing local installation instructions

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[GaneshPatil7517]

Thanks for the feedback @Jefffrey! I've simplified the approach significantly:

Changes made:

✅ Switched to rust:1.92.0-bookworm base image (install Python on top instead of vice versa)
✅ Removed docker-compose.yml - overkill for this use case
✅ Removed entrypoint.sh and CRLF conversion logic - users should configure git properly
✅ Removed COPY . /work - use volume mount at runtime instead
✅ Removed unnecessary comments
Usage is now simply:
docker build -t datafusion-docs ./docs
docker run --rm -v $(pwd):/datafusion datafusion-docs

Tested locally and the docs build successfully. Please take another look!> docker build -t datafusion-docs ./docs
docker run --rm -v $(pwd):/datafusion datafusion-docs

Tested locally and the docs build successfully. Please take another look!

[Jefffrey]

We should be using the build.sh script because that generates the dependency chart for us (and is what is motivating this issue)

[GaneshPatil7517]

Thanks for the feedback @Jefffrey! All addressed:
✅ Using rust:bookworm (latest) instead of pinned version
✅ Removed python3-venv
✅ Using build.sh instead of make html - dependency graph now generates correctly
✅ Added dos2unix for Windows line ending compatibility
Tested locally - docs build succeeds with the dependency graph generated. Please take another look!

[Jefffrey]

When I try the docker run command it seems to download & install some components each time, slowing down the doc build process:

datafusion (docs/dockerized-docs-build)$ docker run --rm -v $(pwd):/datafusion datafusion-docs
info: syncing channel updates for '1.92.0-aarch64-unknown-linux-gnu'
info: latest update on 2025-12-11, rust version 1.92.0 (ded5c06cf 2025-12-08)
info: downloading component 'clippy'
info: downloading component 'rustfmt'
info: installing component 'clippy'
info: installing component 'rustfmt'

Could you look into this?