data2dsl

Ticket 097: Reduce CC in adapter normalizers

• ID: ticket-097
• Owner: agent
• Status: IN_PROGRESS
• Workflow state: EDIT
• Created: 2026-09-18

Goal and scope

Reduce cyclomatic complexity to <=15 (regix cc_max) in the adapter normalizers flagged by the health findings:

• oql.py: normalize_spec + normalize_telemetry (CC 38/38)
• intent.py: normalize (CC 29)
• deta.py: normalize (CC 28)
• planfile.py: normalize (CC 19)
• sumd.py: extract_table_metric (CC 22)

Approach: shared observation/evidence/value helpers extracted into data2dsl_adapter_parts/common.py (observation_envelope, evidence_entry, error_observation, unsupported_observation, numeric_value, temperature_value, set_value) plus per-module metric dispatch and evidence builders. Public signatures and the observation/v0 envelope shape are unchanged.

src/data2dsl_batch.py (compare_batch CC 45, format_markdown_report CC 21) is outside this ticket’s allowedPaths and remains for a follow-up slice.

Acceptance criteria

• AC-01: lizard -C 15 src/data2dsl_adapter_parts/ reports no warnings.
• AC-02: python -m pytest -q passes (158 tests).
• AC-03: ./project/governance-check.sh GOV-PASS.

Changelog

• Extracted shared helpers into common.py; refactored the five flagged functions to table-driven dispatch + per-record evidence builders.
• Restored SUMDAdapter.normalize verbatim (not flagged; kept as-is).