Consolidate normalization pipeline

[amc-corey-cox]

Summary

• Compilers now receive derived_specification (with schema-inferred defaults) instead of the raw specification, eliminating a duplicate induce_missing_values call in PythonCompiler
• Documents the two-phase normalization pipeline (load-time structural normalization vs schema-bind-time semantic induction) in the Transformer class docstring and derived_specification property
• Fixes compiler output to include inferred populated_from and range values — previously these showed as None in compiled markdown/python output

Test plan

• All 586 tests pass (531 unit + 55 compliance)
• Regenerated golden files for markdown and python compiler output
• Ruff clean

Closes #124

[Copilot]

Pull request overview

This PR consolidates the transformation-spec normalization pipeline by clearly separating load-time structural normalization from schema-bind-time semantic induction, and updates compilation pathways to use the schema-derived specification so compiler outputs include inferred fields (e.g., populated_from, range).

Changes:

• Route compiler invocations (CLI + tests) through ObjectTransformer.derived_specification instead of the raw spec.
• Remove redundant induce_missing_values invocation from PythonCompiler.
• Document the two-phase normalization pipeline and regenerate golden compiled markdown output.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/linkml_map/transformer/transformer.py	Documents the two-phase normalization pipeline and expands derived_specification documentation.
src/linkml_map/compiler/python_compiler.py	Removes internal deep-copy + induction from the python compiler so it compiles the already-derived spec.
src/linkml_map/cli/cli.py	CLI compile now compiles the derived specification.
tests/test_compliance/test_compliance_suite.py	Compliance helper now compiles the derived spec when available.
tests/test_compiler/test_python_compiler.py	Python compiler unit test now compiles the derived spec via ObjectTransformer.
tests/input/examples/personinfo_basic/output/personinfo_compiled.md	Golden output updated to reflect inferred populated_from / range values.

Comments suppressed due to low confidence (1)

src/linkml_map/transformer/transformer.py:250

• derived_specification can raise and/or cache a partially-induced spec when source_schemaview is unset or if induce_missing_values errors. Because _derived_specification is assigned before induction, an exception leaves the cache populated and subsequent calls will skip induction. Consider (1) returning None early when source_schemaview is not set (consistent with the | None return type), and (2) computing into a local variable and only assigning _derived_specification after induce_missing_values succeeds.

        if self._derived_specification is None:
            if self.specification is None:
                return None
            self._apply_source_schema_patches()
            self._derived_specification = deepcopy(self.specification)
            induce_missing_values(self._derived_specification, self.source_schemaview)