feat(cal): add structured measurement context

[jeliason]

Summary

• add closed enums for specimen acquisition, compartment, material, preservation, sampling, detection, analyte pool, and normalization
• attach the complete factual context to calibration and cross-scenario readouts
• require measurement context in literature-target completeness checks while leaving mechanistic assertions without experimental readouts
• expose the schema through the public API and MCP schema output
• update the extraction guide and README with the complete authoring contract

Maple records measurement facts only. It does not decide which fields define a statistical discrepancy correction; downstream analyses must predeclare active fields, reference levels, and prior weights.

Compatibility

Existing bare Readout objects still parse. Literature targets without measurement_context are reported incomplete and should be migrated; mechanistic targets remain exempt.

Validation

• ruff check --no-cache src tests
• pytest -p no:cacheprovider -m "not integration"
• 540 passed, 6 skipped