ENH: Support for Open-Meteo API in the Environment class

[Gui-FernandesBR]

Closes #520

Pull request type

• Code changes (bugfix, features)
• ReadMe, Docs and GitHub updates

Checklist

• Tests for the changes have been added (if needed)
• Docs have been reviewed and added / updated
• Lint (ruff check / ruff format) has passed locally
• All tests (pytest tests -m slow --runslow) have passed locally
• CHANGELOG.md — entries added manually as well

Current behavior

Reaching for reanalysis data means downloading files by hand from the Copernicus
Climate Data Store, and pulling a forecast means an OPeNDAP request through
netCDF4, which is heavy and slow. On top of that, the GEFS ensemble shortcut
is currently disabled because NOMADS OPeNDAP was deactivated, so there is no
working out-of-the-box ensemble source.

New behavior

Adds Open-Meteo as an atmospheric data source. It
serves pressure-level data as plain JSON over HTTPS, with no API key and no
netCDF4/OPeNDAP dependency (requests is already a RocketPy dependency, so
this adds none).

Two new model types:

env.set_atmospheric_model("open_meteo")                          # best_match
env.set_atmospheric_model("open_meteo", file="ecmwf_ifs025")     # specific model
env.set_atmospheric_model("open_meteo_ensemble", file="gfs05")   # 31 members
env.select_ensemble_member(10)

Past launches need no extra argument — a past date is transparently routed to
Open-Meteo's historical-forecast archive:

env.set_date([2024, 1, 10, 12])
env.set_atmospheric_model("open_meteo")   # historical-forecast API

The ensemble model stores every member, so select_ensemble_member() and
plots.ensemble_member_comparison() work exactly as they do for GEFS, with the
unperturbed control run exposed as member 0. Open-Meteo also reports the
elevation of the grid cell it answered for, so the launch site elevation is set
automatically.

Notable API findings

Each of the following returns HTTP 200, so none of them fail loudly. They were
found by probing the live API and they shaped the implementation:

• Open-Meteo's ERA5 archive endpoint serves no pressure-level data at all —
  nulls at every level and date tested. It is therefore not used; the
  historical-forecast API is the only archive that can feed a vertical profile.
• The historical archive starts around March 2021 (2021-03-15 comes back
  empty, 2021-03-23 is complete). Earlier dates now raise a warning naming the
  cutoff and pointing at the reanalysis and sounding models, instead of failing
  with a generic "not enough pressure levels" error.
• Only gfs05 (31 members) and ecmwf_ifs025 (51) publish complete
  pressure-level data. gem_global is the subtle one: it serves temperature
  and geopotential height but no pressure-level winds, so probing temperature
  alone makes it look usable. gfs025, icon_global and
  bom_access_global_ensemble return nulls everywhere. All of these are now
  rejected up front with an actionable message.
• Open-Meteo defaults to km/h, so wind_speed_unit=ms is requested explicitly;
  dropping it would silently inflate wind speeds by 3.6x.

Coverage of pressure levels varies per model (gfs_seamless reaches 30 hPa,
ecmwf_ifs025 stops at 50 hPa), so levels a model does not resolve are dropped
while parsing.

Drive-by bug fix

The model-type gates in _EnvironmentPrints and _EnvironmentPlots compared
against capitalised literals ("Ensemble", "Forecast"), but
set_atmospheric_model documents type as case-insensitive and stores it as the
user spelled it. So type="ensemble" silently printed no time period and no
member count, and skipped the ensemble comparison plot. They now compare
case-insensitively.

Breaking change

• No

Additional information

Implementation notes

• Open-Meteo reports wind as speed/direction rather than u/v, so
  convert_wind_speed_direction_to_components was added to
  environment/tools.py. It converts the meteorological blows-from convention
  into RocketPy's East/North components, and is round-trip tested against
  calculate_wind_heading — getting that convention wrong would flip the wind
  by 180°.
• Temperatures are converted from Celsius to Kelvin, pressure levels from hPa to
  Pa, and geopotential heights to geometric altitude.
• to_dict/from_dict round-trip both new types, including every ensemble
  member.

Verification

• 50 new unit tests, all offline: the fetchers are patched, and the suite was
  re-run with socket.connect blocked to confirm no network access.
• 6 new integration tests marked slow, exercising the live API (forecast across
  4 models, the historical archive, and the ensemble with member selection).
• The full unit suite passes: 1954 passed, 16 skipped.
• The full tests/integration/environment suite passes with --runslow: 29
  passed, 1 skipped.
• Every code block in the new docs page runs as part of the docs build; all five
  were executed against the live API.
• End-to-end sanity check: a full Calisto flight using open_meteo at Spaceport
  America gives 3421 m AGL apogee at Mach 0.82, with the 1405 m site elevation
  picked up automatically from the API.

Docs

New page at docs/user/environment/1-atm-models/open_meteo.rst, cross-referenced
from the forecast, reanalysis and ensemble pages, since Open-Meteo is the lighter
alternative in each of those cases — and currently the only working ensemble
source while the GEFS shortcut is down.

🤖 Generated with Claude Code

[github-actions]

Failed to generate code suggestions for PR

[codecov]

Codecov Report

❌ Patch coverage is 95.23810% with 10 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.80%. Comparing base (e0ff281) to head (c1bd678).
⚠️ Report is 41 commits behind head on develop.

Files with missing lines	Patch %	Lines
...ocketpy/environment/fetchers/open_meteo_fetcher.py	85.24%	9 Missing ⚠️
rocketpy/environment/environment.py	99.29%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1119      +/-   ##
===========================================
+ Coverage    82.18%   82.80%   +0.62%     
===========================================
  Files          122      129       +7     
  Lines        16355    16812     +457     
===========================================
+ Hits         13441    13922     +481     
+ Misses        2914     2890      -24     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.