v0.4.0

What's Changed

• feat(analysis): coverage gate — drop too-gappy series from correlation by @anym001 in #49
• revert(analysis): drop coverage gate; stamp raw correlation for validation by @anym001 in #50
• feat(analysis): stamp residual (deseasonalised) correlation coefficient by @anym001 in #51
• feat(analysis): correlate on the residual instead of the de-trended series by @anym001 in #52
• feat(analysis): require raw corroboration of residual correlations by @anym001 in #53
• feat(analysis): require year-over-year reproducibility for seasonality by @anym001 in #54
• feat(analysis): corroborate anomalies against full history and dedupe workout family by @anym001 in #55
• feat(analysis): require directional consistency for trend findings by @anym001 in #56
• feat(narrate): add --dry-run to render the model context without Ollama by @anym001 in #57
• Release v0.4.0 — corroboration-gated findings engine + narrate --dry-run by @anym001 in #58

Full Changelog: v0.3.0...v0.4.0

What's Changed

• feat(analysis): coverage gate — drop too-gappy series from correlation by @anym001 in #49
• revert(analysis): drop coverage gate; stamp raw correlation for validation by @anym001 in #50
• feat(analysis): stamp residual (deseasonalised) correlation coefficient by @anym001 in #51
• feat(analysis): correlate on the residual instead of the de-trended series by @anym001 in #52
• feat(analysis): require raw corroboration of residual correlations by @anym001 in #53
• feat(analysis): require year-over-year reproducibility for seasonality by @anym001 in #54
• feat(analysis): corroborate anomalies against full history and dedupe workout family by @anym001 in #55
• feat(analysis): require directional consistency for trend findings by @anym001 in #56
• feat(narrate): add --dry-run to render the model context without Ollama by @anym001 in #57
• Release v0.4.0 — corroboration-gated findings engine + narrate --dry-run by @anym001 in #58

Full Changelog: v0.3.0...v0.4.0