gmat-run v0.4 — CI on setup-gmat, R2025a/R2026a matrix, Mission.summary()

[djankov]

gmat-run v0.4 is on PyPI. This one is about CI debt rather than new features — every PR now runs against R2025a and R2026a on Ubuntu / Windows / macOS × Python 3.10 / 3.11 / 3.12 (18 cells), the install step is one line, and a canonical-image canary catches drift between the action and ghcr.io/astro-tools/gmat. The one user-visible API addition is Mission.summary() plus notebook reprs.

from gmat_run import Mission

mission = Mission.load("flyby.script")

# New in v0.4 — structured snapshot of the parsed script
summary = mission.summary()
summary.install.version              # 'R2026a'
[r.kind for r in summary.resources]  # ['Spacecraft', 'ForceModel', 'Propagator', ...]
[c.name for c in summary.commands]   # ['Propagate', 'Maneuver', 'Propagate']

# Also new — drop `mission` (or `result`) on the last line of a notebook cell
# and Jupyter renders a structured table instead of `<gmat_run.Mission ...>`.
mission

What's new in v0.4

• Mission.summary() and notebook reprs — Mission.summary() returns a MissionSummary with the resolved GmatInstall, the parsed ResourceGroups and CommandOutline for the mission sequence. Mission and Results both gain _repr_html_ so a notebook cell that ends in either renders a structured table instead of a raw <…> repr (#99).
• CI on astro-tools/setup-gmat@v0 — the inline GMAT install/cache step is replaced by the setup-gmat action. The action handles download, cache, BuildApiStartupFile.py, and GMAT_ROOT export — the workflow shrinks to one step (#95). A new docs cookbook page, Run gmat-run in your CI, walks through wiring the same setup into a downstream project's workflow, with caching, optional extras, and a multi-version matrix (#97).
• Multi-version CI — the matrix doubles to exercise both R2025a and R2026a on every PR (Ubuntu / Windows / macOS × Py 3.10 / 3.11 / 3.12 × R2025a / R2026a — 18 cells). Both releases are now first-class supported (#96).
• Coverage gate raised to 90% overall (was 80%), enforced on the Ubuntu / Python 3.12 / R2026a cell. Per-package gate on gmat_run/parsers/ stays at 95% (#98).
• Canonical-image canary — a separate CI cell exercises pip install gmat-run and a Mission.load → run → assert non-empty ReportFile round-trip inside ghcr.io/astro-tools/gmat:R2026a on every PR. The action and the canonical image can drift apart silently; this is the canary that catches it before downstream container-mode CI users do (#100).

Full notes: https://github.com/astro-tools/gmat-run/blob/main/CHANGELOG.md#040--2026-05-03

Install

pip install -U gmat-run                 # core
pip install -U "gmat-run[ccsds-ndm]"    # adds CCSDS-OEM export (from v0.3)
pip install -U "gmat-run[astropy]"      # adds time-scale conversion (from v0.3)
pip install -U "gmat-run[spiceypy]"     # adds SPK ephemeris support (from v0.2)

Same baseline: Python 3.10–3.12 and a local GMAT install. R2025a and R2026a are now both exercised on every PR.

Links

• PyPI: https://pypi.org/project/gmat-run/
• Docs: https://astro-tools.github.io/gmat-run/
• Source: https://github.com/astro-tools/gmat-run
• Release notes: https://github.com/astro-tools/gmat-run/releases/tag/v0.4.0

Feedback wanted

• Wired the CI cookbook into your own project's GitHub Actions — does the four-step workflow actually drop in clean, or did you have to massage something? Issue or comment with the diff.
• The notebook reprs are new surface area — anything you'd want surfaced on Mission/Results that isn't there yet? MissionSummary is the place to grow it.
• Running container-mode CI inside ghcr.io/astro-tools/gmat? Heads up that the canary now exercises that path on every PR — flag any drift you spot.