BDSCM provenance

Imported from https://github.com/CausalDataScience/TransPFN
commit c5a2af9cef017913240ca5687885e0cd9248cbdc.
The original imported snapshot hashes are recorded below. Four Python sources and the catalog JSON remain unchanged. scm.py adapts the
preset to the shared StructuralCausalModel evaluator; outcome.py adds an optional
score offset for CSCM, retaining the original default calculation.
Original TransPFN files remain available for reproduction; CausalEngine is the
maintenance location for this imported implementation going forward.
No historical upstream license declaration was found for these files. On
October 26, 2026, Yonghan Jung explicitly confirmed his authority to license this
imported TransPFN/BDSCM material and authorized its public distribution under the
CausalEngine Noncommercial License 1.0 in the root LICENSE. This present author
authorization does not assert a historical upstream grant. Original provenance
and snapshot hashes are preserved below. See LICENSING.md for the scoped policy.

Original SHA-256 values:
a7a052b3bad7db9fb42e9136e72848d1256f5a7a923088adf9bff758af2d67a3  __init__.py
88a53d74bc87b521abab04db77444d41142696be4b65e914c5f23010308b98d2  covariates.py
ab60f52a91341dfe5c5549fef441caba87e5d3124211a3dfe6e2cb82d6b49c22  data/tabdpt_openml.json
296dba66a4e73c183b54115fc19ac5d5c7a0bb9284177ba4eea4a183d378dee7  noise.py
ea98ac31f80c1e76e42dafa635ecb0e6005123923ee5908f459423308be7fa5b  outcome.py
87c2716d899606bfc69d990899e84eabdb7a5bf0740029eaf2cea18ff43f5fac  scm.py
fd76b1c2fd38c55df63c4b50b646ab31fcb932a84bc3750fe1f4871a77e0abc6  treatment.py
564c45e450fc0a3d21b9199cf2b11a1e38f1671099b8ef31669de6d5b9f54267  utils.py

CausalEngine 0.1.0a6 adds optional pre-transform score offsets to covariates.py;
the default BDSCM transform and original source hashes remain unchanged.
