Add snap2stamps-style pipeline definitions using sarpyx.snapflow operators

[sirbastiano]

Motivation

• Provide programmatic equivalents of the snap2stamps graph collection (v2.0.1) using the existing sarpyx.snapflow GPT operator methods to enable reproducible, composable SNAP-based Pipelines within this project.
• Expose common StaMPS/PSI and SBAS processing flows as reusable Python workflows to simplify automation and integration with existing GPT wrappers.

Description

• Added a new module sarpyx/snapflow/snap2stamps_pipelines.py which defines a PipelineStep dataclass and declarative graph-level pipeline mappings to GPT operator methods.
• Implemented composed workflows (psi_prep, psi_post_unwrap, psi_full, sbas_full) that chain the graph-level blocks into end-to-end sequences.
• Added execution helpers: _run_step (with per-step overrides and defaults for stamps_export and snaphu_export paths), run_graph_pipeline, run_workflow, and a convenience build_gpt constructor.

Testing

• Ran Python bytecode compilation with python -m compileall sarpyx/snapflow/snap2stamps_pipelines.py, which completed successfully.
• Performed an import/registry check with python -c "from sarpyx.snapflow.snap2stamps_pipelines import SNAP2STAMPS_GRAPH_PIPELINES, SNAP2STAMPS_WORKFLOWS; print(len(SNAP2STAMPS_GRAPH_PIPELINES), len(SNAP2STAMPS_WORKFLOWS))" which printed 13 4 confirming the module loads and registries are present.
• Attempted remote retrieval of the original snap2stamps graphs for direct diffing but network access to the upstream GitHub content was blocked in this environment (HTTP 403), so offline mapping was implemented against the available SNAP operator methods instead.

---

Codex Task