Port JL advanced-physics compatibility to rift_O4d

[oshaughnessy-junior]

Summary

This is a conflict-resolved port of the core advanced-physics functionality from oshaughn/rift_O4d_JL (355c3f9f) onto oshaughnessy-junior/rift_O4d (4c1da52c). It preserves the newer rift_O4d calibration, interpolation, GPU, and pipeline behavior while adding:

• eccentricity and mean-anomaly propagation through ILE, CIP, pseudo-pipe, conversion, injection, and frame utilities;
• TEOBResumS a6c and hyperbolic E0/p_phi0 sampling, output, grid filtering, and waveform handling;
• generalized hyperpipeline column-layout detection for combinations of tides, EOS index, EOB, hyperbolic, eccentric, and distance fields;
• a TEOBResumS compatibility adapter with auto (default), dali, default, and legacy profiles, safe schema preflight, explicit overrides, and runtime fingerprinting;
• distance-invariant hyperbolic outcome classification using TEOBResumS radial dynamics rather than an absolute strain threshold;
• focused regression tests for the compatibility adapter and advanced-parameter data paths.

The old integrate_likelihood_extrinsic_batchmode_error full-script duplicate is intentionally not ported: it had diverged into a maintenance fork and does not add a distinct supported entry point.

Compatibility behavior

auto is the default. Known DALI interfaces receive their string-valued enum/boolean schema; other or future-looking interfaces conservatively receive the established default schema. An operator can pin RIFT_TEOBRESUMS_PROFILE=default|dali|legacy. The first schema is probed in a subprocess, preventing an incompatible extension from taking down the parent RIFT process.

Validation

• 31 passed across test_teobresums_compat.py, test_advanced_parameter_ports.py, and test_hyperpipeline_io.py.
• 154 passed for those suites plus the CI integrator-accounting, CIP-prior, and cosmology-source tests; the three new advanced EOB priors now have explicit support and normalization coverage.
• Full LISA gate: 266 passed, 1 skipped; the three ground-based external-TEOB options are explicitly classified as not applicable to the current LISA waveform path in the generated drift ledger.
• pyflakes clean for the new helper and tests; staged diff whitespace check clean.
• Local current DALI (9c4482d9) generated a finite hyperbolic waveform and classified the test point as zoom-whirl.
• CIT Jake environment (/home/jacob.lange/.conda/envs/20260204-Ecc-Tides-container-mylalsuite-env, DALI extension SHA-256 bade0374...) passed both hyperbolic generation and a combined eccentric + precessing + tidal + a6c waveform (3,353 finite samples).
• CIT legacy TEOBResumS (26f2343e) passed profile detection and finite waveform generation.
• CLI import/help gates passed for ILE, pseudo-pipe, CIP, converters, and frame utilities.

No files under RIFT/integrators/ change, so the posterior shape-recovery gate is not triggered.

Remaining follow-up

• Add CI environments containing representative real TEOBResumS binaries; dependency-free CI exercises the adapter logic, while real-backend tests currently require provisioned runtimes.
• Record the exposed TEOBResumS runtime fingerprint in durable run provenance, not only diagnostics.
• Run a full scheduler/DAG analysis with representative data before promoting from draft; waveform generation and parameter/file plumbing are covered here, but not a production-scale PE run.
• Decide and document the scientifically supported TEOBResumS commit set. Interface compatibility prevents crashes, but it cannot make distinct backend revisions scientifically identical.

[interhost-cross-review-canary-v2]

Independent automated review completed at the recorded exact commit. Detailed findings were withheld from public output by the private-context egress policy and require private human declassification.