docs(qec): misc docs updates

[bmhowe23]

Description

Documentation-only change that closes several gaps for the 0.7.0 QEC release,
where new or existing decoder features were not reflected in the docs. No code
or behavior changes.

Decoders newly documented

• PyMatching (MWPM) decoder — new API fragment pymatching_api.rst
  (error_rate_vec, merge_strategy and its five values, optional O
  observables matrix, matchable-graph requirement), wired into the Python and
  C++ "Built-in Decoders" lists, plus a tutorial section in decoders.rst.
• Chromobius color-code decoder — new API fragment chromobius_api.rst
  (DEM-native construction from Stim DEM text, the "no parity-check matrix"
  restriction, ≤64-observable limit, and the five boolean options), wired into
  both API pages and decoders.rst.

Existing 0.7.0 features/options now documented

• Surface-code orientation construction option (XV/XH/ZV/ZH, aliases
  O1–O4, default ZH) with Python and C++ usage examples.
• trt_decoder composite/global decoder chaining (global_decoder /
  global_decoder_params, including the "supply params — even empty — when
  global_decoder is set" caveat).
• cuda_device_id placement knob added to the trt_decoder and
  nv-qldpc-decoder parameter tables.
• Note that Python decode results are NumPy arrays as of 0.7.0
  (DecoderResult / BatchDecoderResult / AsyncDecoderResult).

All option names, values, and defaults were verified against the plugin
sources (pymatching.cpp, chromobius.cpp, surface_code.cpp, decoder.cpp,
trt_decoder.cpp, py_decoder.cpp).

This branch is based on main and is intended to be cherry-picked onto
releases/v0.7.0.

Runtime / performance impact

N/A — documentation only.

Self-review checklist

Before requesting review

• I reviewed my own full diff in GitHub or my editor.
• PR is in Draft if it is not yet ready for review.
• Temporary / debugging changes have been removed.
• Local test logs reviewed; no unexplained warnings or errors. (Docs build with sphinx -n -W --keep-going succeeds with zero warnings; the script's invalid-HTML checks pass.)
• CI logs reviewed; no unexplained warnings or errors.
• Full CI has been run.

Scope and size

• PR is under ~1000 lines, or an exception is justified in the description.
• Refactoring-only changes are isolated in their own PR(s).
• No existing tests were disabled or modified just to make this PR pass.

Tests

• New functionality has new tests. Documentation-only change; no code paths added.
• Tests fail if the new functionality is broken. N/A (docs only).
• Negative tests added where exceptions are expected. N/A (docs only).
• Truth data added. N/A (docs only).
• CI runtime impact considered; team notified if significant. (Docs-only; negligible.)

Documentation

• Public-facing APIs have Doxygen docs.
• User-visible behavior changes have public docs, or a follow-up is tracked.

Code style

• Naming follows the existing convention for the area being modified. (reStructuredText; mirrors the existing *_api.rst fragment pattern.)

Dependencies

• No new third-party dependencies.

🤖 Generated with Claude Code

[melody-ren]

Using this PR as a catch all place: I think decoder_context is not in the docs yet

[bmhowe23]

Using this PR as a catch all place: I think decoder_context is not in the docs yet

I think it's actually already documented - it's under the Python casing DecoderContext (PascalCase vs snake_case as in many [but not all] of the Python vs C++ APIs). Coverage today:

• Python: cudaq_qec.DecoderContext (with :members:) in api/qec/python_api.rst.
• C++: cudaq::qec::decoder_context in api/qec/cpp_api.rst.