unify json atlases, CLI segreg changes and CLI segcentroids new

[m-reuter]

• Redesigned segreg into an LTA-focused centroid-registration CLI using --seg plus exactly one of --target-seg, --centroids, or --flipped as target.
• --centroids can either be path to a centroid JSON file or name of a provided atlas (fsaverage and min_icbm... see below).
• Switched centroids files to a single richer JSON format with required centroids and optional embedded geometry.
• Added new segcentroids CLI for writing centroid target JSON files from segmentations.
• Unified bundled targets onto that format and exposed both fsaverage and mni_icbm152_t1_tal_nlin_asym_09c.
• Added segcentroids --input ... --geometry ... mode to add or replace geometry in an existing target JSON without recomputing centroids.
• Updated tests and docs/README to cover the new CLI and target-file workflow.
• Validated with Ruff and the full pytest suite.

[Copilot]

Pull request overview

This PR redesigns segmentation-centroid registration workflows by standardizing centroid-target JSON files, updating the segreg CLI/API to focus on LTA output, and adding a dedicated segcentroids CLI for generating/editing centroid-target JSON files.

Changes:

• Unified bundled targets into a single “target JSON” format (required centroids + optional embedded geometry) and added an MNI centroid target alongside fsaverage.
• Refactored segreg API/CLI to use --seg plus exactly one of --target-seg, --centroids, or --flipped, and to always write an LTA.
• Added segcentroids CLI and updated tests/docs to cover the new target-file workflow.

Reviewed changes

Copilot reviewed 15 out of 15 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
tests/test_segreg.py	Updates/extends tests for new target JSON format, new bundled targets, new segreg CLI contract, and new segcentroids CLI.
pyproject.toml	Registers the new segcentroids console script.
neuroreg/segreg/register.py	Refactors registration entrypoints to target-seg vs centroid-target sources and adds export_segmentation_target.
neuroreg/segreg/labels.py	Renames the label preset literal from fsaverage_centroids to target_centroids.
neuroreg/segreg/io.py	Introduces rich target JSON read/write with optional geometry + legacy centroid-only compatibility.
neuroreg/segreg/data/mni_icbm152_t1_tal_nlin_asym_09c.json	Adds bundled MNI centroid target (centroids-only, no geometry).
neuroreg/segreg/data/fsaverage_data.json	Removes old standalone fsaverage geometry resource (superseded by unified target JSON).
neuroreg/segreg/data/fsaverage_centroids.json	Removes old standalone fsaverage centroids resource (superseded by unified target JSON).
neuroreg/segreg/data/fsaverage.json	Adds unified fsaverage target JSON with centroids + embedded geometry.
neuroreg/segreg/atlas.py	Reworks atlas helpers to load unified target JSON bundles and expose load_atlas_target.
neuroreg/cli/segreg.py	Updates CLI args/behavior to new target selection model and LTA-only output.
neuroreg/cli/segcentroids.py	New CLI to generate/edit centroid target JSON files and optionally embed/replace geometry.
neuroreg/cli/__init__.py	Documents the new CLI entrypoint.
doc/api/segreg.rst	Updates API docs to match refactors (removes old exports, adds new one).
README.md	Updates user-facing CLI documentation and adds segcentroids section.

Comments suppressed due to low confidence (1)

tests/test_segreg.py:537

• The indentation inside this @pytest.mark.parametrize payload is over-indented (e.g., the 7, line). This is likely to trip Ruff/pycodestyle continuation-indentation rules (E12x) and makes the parameter table harder to read. Reformat the tuples so the values align with the surrounding parentheses indentation.

    [
        (
                7,
                np.array(
                    [

---

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