mixdyn 0.1.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (218) hide show
  1. mixdyn-0.1.0/.github/workflows/ci.yml +116 -0
  2. mixdyn-0.1.0/.github/workflows/release.yml +117 -0
  3. mixdyn-0.1.0/.gitignore +11 -0
  4. mixdyn-0.1.0/.python-version +1 -0
  5. mixdyn-0.1.0/CHANGELOG.md +95 -0
  6. mixdyn-0.1.0/CONTRIBUTING.md +64 -0
  7. mixdyn-0.1.0/LICENSE +202 -0
  8. mixdyn-0.1.0/NOTICE +2 -0
  9. mixdyn-0.1.0/PKG-INFO +298 -0
  10. mixdyn-0.1.0/README.md +270 -0
  11. mixdyn-0.1.0/SKILL.md +165 -0
  12. mixdyn-0.1.0/allowed-runtime-packages.txt +30 -0
  13. mixdyn-0.1.0/docs/adr/0001-dependency-policy.md +58 -0
  14. mixdyn-0.1.0/docs/adr/0002-data-model.md +85 -0
  15. mixdyn-0.1.0/docs/adr/0004-engines-share-a-decomposition-not-a-base-class.md +57 -0
  16. mixdyn-0.1.0/docs/adr/0005-geo-plans.md +53 -0
  17. mixdyn-0.1.0/docs/adr/0006-population-national-media-and-aggregation.md +83 -0
  18. mixdyn-0.1.0/docs/adr/0007-the-model-over-geos.md +92 -0
  19. mixdyn-0.1.0/docs/adr/0008-the-ridge-engine.md +127 -0
  20. mixdyn-0.1.0/docs/adr/0009-the-sampler.md +107 -0
  21. mixdyn-0.1.0/docs/adr/0010-the-bayesian-model.md +84 -0
  22. mixdyn-0.1.0/docs/adr/0011-curve-families-and-their-weights.md +72 -0
  23. mixdyn-0.1.0/docs/adr/0012-chains-started-by-pathfinder.md +89 -0
  24. mixdyn-0.1.0/docs/adr/0013-the-bayesian-model-over-geos.md +136 -0
  25. mixdyn-0.1.0/docs/adr/0014-the-nested-model.md +204 -0
  26. mixdyn-0.1.0/docs/adr/0015-lift-estimates-in-the-likelihood.md +111 -0
  27. mixdyn-0.1.0/docs/adr/0016-one-definition-of-each-return.md +77 -0
  28. mixdyn-0.1.0/docs/adr/0017-log-normal-noise-about-the-mean.md +72 -0
  29. mixdyn-0.1.0/docs/adr/0018-lift-estimates-over-geos-and-in-the-nested-model.md +127 -0
  30. mixdyn-0.1.0/docs/adr/0019-a-plan-for-a-posteriors-expected-return.md +118 -0
  31. mixdyn-0.1.0/docs/adr/0020-the-dlm-over-media.md +177 -0
  32. mixdyn-0.1.0/docs/adr/0021-the-scorecards-harness.md +237 -0
  33. mixdyn-0.1.0/docs/adr/0022-kept-specs-drawn-until-they-mix.md +56 -0
  34. mixdyn-0.1.0/docs/adr/0023-a-stratified-sample-of-kept-specs.md +77 -0
  35. mixdyn-0.1.0/docs/adr/0024-r-hat-held-where-stan-holds-it.md +62 -0
  36. mixdyn-0.1.0/docs/adr/0025-divergences-read-against-stans-own-runs.md +52 -0
  37. mixdyn-0.1.0/docs/adr/0026-each-effects-discount-and-the-ends-of-a-grid.md +94 -0
  38. mixdyn-0.1.0/docs/adr/0027-a-forecast-drawn-with-its-signs-held.md +101 -0
  39. mixdyn-0.1.0/docs/adr/0028-the-static-models-posterior-predictive.md +87 -0
  40. mixdyn-0.1.0/docs/adr/0029-a-plan-for-a-posteriors-worst-share.md +77 -0
  41. mixdyn-0.1.0/docs/adr/0030-a-lift-tests-value-to-the-plan.md +221 -0
  42. mixdyn-0.1.0/docs/adr/0031-a-set-of-lift-tests-chosen-by-its-value.md +124 -0
  43. mixdyn-0.1.0/docs/adr/0032-a-refit-corrects-a-tests-value.md +92 -0
  44. mixdyn-0.1.0/docs/adr/0033-a-lift-on-the-kpi-through-the-mediator.md +120 -0
  45. mixdyn-0.1.0/docs/adr/0034-the-posteriors-plan-searched-on-its-grid.md +98 -0
  46. mixdyn-0.1.0/docs/adr/0035-a-records-compute-and-a-fit-past-its-chains-bounds.md +123 -0
  47. mixdyn-0.1.0/docs/adr/0036-the-intercept-sampled-as-the-level.md +102 -0
  48. mixdyn-0.1.0/docs/adr/0037-a-lift-read-over-the-panels-own-cells.md +175 -0
  49. mixdyn-0.1.0/docs/adr/0038-a-lift-test-read-by-the-dlm.md +143 -0
  50. mixdyn-0.1.0/docs/adr/0039-one-runner-for-every-arm.md +108 -0
  51. mixdyn-0.1.0/docs/adr/0040-the-claims-gate-and-forecasts-from-a-bootstrap.md +189 -0
  52. mixdyn-0.1.0/docs/adr/0041-the-ridge-arm-and-the-chc-arm.md +248 -0
  53. mixdyn-0.1.0/docs/adr/0042-a-channel-bought-nationally-planned-as-one-spend.md +51 -0
  54. mixdyn-0.1.0/docs/adr/0043-a-geo-world-planned-by-the-engines-arms.md +146 -0
  55. mixdyn-0.1.0/docs/adr/0044-a-no-test-plan-read-against-the-equal-split.md +204 -0
  56. mixdyn-0.1.0/docs/adr/0045-a-candidates-standard-error-from-placebo-windows.md +72 -0
  57. mixdyn-0.1.0/docs/adr/0046-a-lift-test-read-period-by-period.md +134 -0
  58. mixdyn-0.1.0/docs/adr/0047-a-geo-test-read-into-a-lift-estimate.md +147 -0
  59. mixdyn-0.1.0/docs/adr/0048-synthetic-difference-in-differences-as-the-paper-states-it.md +176 -0
  60. mixdyn-0.1.0/docs/adr/0049-a-geo-tests-design-randomised-within-matched-pairs.md +157 -0
  61. mixdyn-0.1.0/docs/adr/0050-a-plans-table-read-by-one-definition-of-each-return.md +159 -0
  62. mixdyn-0.1.0/docs/adr/0051-scenarios-read-as-differences-with-their-carry-out.md +99 -0
  63. mixdyn-0.1.0/docs/adr/0052-a-flighted-plan-scored-against-the-same-carry-in.md +111 -0
  64. mixdyn-0.1.0/docs/adr/0053-a-return-prior-that-counts-the-carry-out.md +139 -0
  65. mixdyn-0.1.0/docs/adr/0054-the-licence-and-the-road-to-publication.md +98 -0
  66. mixdyn-0.1.0/docs/adr/0055-a-geo-tests-power-read-by-the-reader-that-reads-it.md +169 -0
  67. mixdyn-0.1.0/docs/adr/0056-a-switchback-test-read-by-its-own-randomisation.md +350 -0
  68. mixdyn-0.1.0/docs/adr/0058-readers-audited-on-null-panels.md +152 -0
  69. mixdyn-0.1.0/docs/adr/0059-simulation-based-calibration-recorded.md +127 -0
  70. mixdyn-0.1.0/docs/adr/0060-the-health-card.md +149 -0
  71. mixdyn-0.1.0/docs/adr/0061-sensitivity-to-unobserved-confounding.md +241 -0
  72. mixdyn-0.1.0/docs/adr/0062-the-file-format-and-the-public-surface.md +112 -0
  73. mixdyn-0.1.0/docs/adr/0063-versioned-documents-read-by-their-declared-type.md +151 -0
  74. mixdyn-0.1.0/docs/adr/0064-the-transforms-two-families-each-in-its-engines-convention.md +117 -0
  75. mixdyn-0.1.0/docs/adr/0065-convergence-diagnostics-and-where-their-thresholds-come-from.md +107 -0
  76. mixdyn-0.1.0/docs/adr/0066-the-dlm-core-chc-dlm-blocks-on-the-panels-calendar.md +121 -0
  77. mixdyn-0.1.0/docs/adr/0068-levers-measured-against-their-counterfactual-level.md +219 -0
  78. mixdyn-0.1.0/docs/adr/0069-calendar-events-beside-the-level.md +79 -0
  79. mixdyn-0.1.0/docs/adr/0070-a-conversions-kpis-cost-per-incremental-kpi.md +62 -0
  80. mixdyn-0.1.0/docs/adr/0071-a-posteriors-plan-for-a-target-return.md +97 -0
  81. mixdyn-0.1.0/docs/adr/0072-a-models-factsheet-and-the-decision-on-one-page.md +153 -0
  82. mixdyn-0.1.0/docs/adr/0073-twin-worlds-a-history-cannot-tell-apart.md +120 -0
  83. mixdyn-0.1.0/docs/adr/0074-spend-set-by-demand.md +180 -0
  84. mixdyn-0.1.0/docs/adr/0076-a-return-prior-that-may-be-0.md +109 -0
  85. mixdyn-0.1.0/docs/adr/0077-the-calibration-loop-as-a-scorecard-arm.md +87 -0
  86. mixdyn-0.1.0/docs/adr/0078-the-decision-reader-and-geo-arms-of-the-scorecard.md +167 -0
  87. mixdyn-0.1.0/docs/adr/0079-a-blocked-holdout-a-tempered-refresh-and-the-refresh-cell.md +205 -0
  88. mixdyn-0.1.0/docs/adr/0080-the-claim-over-every-root-of-a-run.md +125 -0
  89. mixdyn-0.1.0/docs/adr/0081-the-claim-by-family-and-rung-and-the-next-run-s-development-set.md +162 -0
  90. mixdyn-0.1.0/docs/adr/0082-the-tested-rung-s-curve-family-read-from-the-lift-rows.md +125 -0
  91. mixdyn-0.1.0/docs/adr/0083-a-geo-panel-planned-by-the-model-over-geos.md +170 -0
  92. mixdyn-0.1.0/docs/adr/0084-tuned-rival-arms-and-the-rivals-in-the-refresh-cell.md +116 -0
  93. mixdyn-0.1.0/docs/adr/0085-recurring-events-read-off-the-records-and-a-no-test-arm.md +160 -0
  94. mixdyn-0.1.0/docs/adr/0086-the-margin-arm-linear-margins-shrunk-and-planned-at-a-corner.md +344 -0
  95. mixdyn-0.1.0/docs/adr/0087-the-curved-margin-variant-saturated-margins-water-filled-in-the-adstock-range.md +228 -0
  96. mixdyn-0.1.0/docs/adr/0088-the-tested-rung-rule-anchored-margins-and-the-gs-composite.md +170 -0
  97. mixdyn-0.1.0/docs/adr/0089-the-curved-tested-rung-rule-a-curve-posterior-off-the-tests-and-the-gs-curved-composite.md +237 -0
  98. mixdyn-0.1.0/docs/adr/0090-a-closed-form-model-over-geos-fitted-by-empirical-bayes.md +222 -0
  99. mixdyn-0.1.0/docs/adr/0091-the-noise-mixture-a-flat-nuisance-prior-and-error-models-weighed-by-their-evidence.md +235 -0
  100. mixdyn-0.1.0/docs/adr/0092-the-time-guard-a-channel-whose-spend-follows-the-week-is-held-at-its-status-quo.md +187 -0
  101. mixdyn-0.1.0/docs/adr/0093-the-time-guard-at-the-tested-rung-and-the-gs-held-composite.md +150 -0
  102. mixdyn-0.1.0/docs/adr/0094-the-default-read-against-a-declared-variant-beside-the-claim.md +46 -0
  103. mixdyn-0.1.0/docs/adr/0095-public-on-pypi-from-a-fresh-history.md +72 -0
  104. mixdyn-0.1.0/docs/api/index.md +627 -0
  105. mixdyn-0.1.0/docs/release-check.md +82 -0
  106. mixdyn-0.1.0/docs/scope.md +15 -0
  107. mixdyn-0.1.0/justfile +34 -0
  108. mixdyn-0.1.0/pyproject.toml +83 -0
  109. mixdyn-0.1.0/scripts/api_index.py +148 -0
  110. mixdyn-0.1.0/scripts/check_runtime_deps.py +85 -0
  111. mixdyn-0.1.0/scripts/family12_curves.py +76 -0
  112. mixdyn-0.1.0/scripts/real_data_protocol.py +352 -0
  113. mixdyn-0.1.0/scripts/study_dlm_prior.py +187 -0
  114. mixdyn-0.1.0/scripts/study_endogenous_spend.py +525 -0
  115. mixdyn-0.1.0/scripts/study_events.py +126 -0
  116. mixdyn-0.1.0/scripts/study_geo_design.py +209 -0
  117. mixdyn-0.1.0/scripts/study_levers.py +475 -0
  118. mixdyn-0.1.0/scripts/study_lift_overlap.py +353 -0
  119. mixdyn-0.1.0/scripts/study_lift_series.py +206 -0
  120. mixdyn-0.1.0/scripts/study_recurring_events.py +150 -0
  121. mixdyn-0.1.0/src/mixdyn/__init__.py +47 -0
  122. mixdyn-0.1.0/src/mixdyn/api/__init__.py +75 -0
  123. mixdyn-0.1.0/src/mixdyn/api/documents.py +2070 -0
  124. mixdyn-0.1.0/src/mixdyn/api/serialize.py +481 -0
  125. mixdyn-0.1.0/src/mixdyn/calibration/__init__.py +21 -0
  126. mixdyn-0.1.0/src/mixdyn/calibration/anchors.py +441 -0
  127. mixdyn-0.1.0/src/mixdyn/calibration/lift.py +178 -0
  128. mixdyn-0.1.0/src/mixdyn/calibration/tables.py +383 -0
  129. mixdyn-0.1.0/src/mixdyn/data/__init__.py +47 -0
  130. mixdyn-0.1.0/src/mixdyn/data/feedback.py +208 -0
  131. mixdyn-0.1.0/src/mixdyn/data/geos.py +191 -0
  132. mixdyn-0.1.0/src/mixdyn/data/issues.py +83 -0
  133. mixdyn-0.1.0/src/mixdyn/data/panel.py +475 -0
  134. mixdyn-0.1.0/src/mixdyn/data/read.py +1001 -0
  135. mixdyn-0.1.0/src/mixdyn/data/recurring.py +230 -0
  136. mixdyn-0.1.0/src/mixdyn/decisions/__init__.py +84 -0
  137. mixdyn-0.1.0/src/mixdyn/decisions/_quantile.py +50 -0
  138. mixdyn-0.1.0/src/mixdyn/decisions/flighting.py +277 -0
  139. mixdyn-0.1.0/src/mixdyn/decisions/geo_plan.py +474 -0
  140. mixdyn-0.1.0/src/mixdyn/decisions/margin_plan.py +404 -0
  141. mixdyn-0.1.0/src/mixdyn/decisions/plan_table.py +418 -0
  142. mixdyn-0.1.0/src/mixdyn/decisions/posterior.py +1267 -0
  143. mixdyn-0.1.0/src/mixdyn/decisions/returns.py +177 -0
  144. mixdyn-0.1.0/src/mixdyn/decisions/scenarios.py +475 -0
  145. mixdyn-0.1.0/src/mixdyn/experiments/__init__.py +123 -0
  146. mixdyn-0.1.0/src/mixdyn/experiments/candidates.py +488 -0
  147. mixdyn-0.1.0/src/mixdyn/experiments/geo_design.py +777 -0
  148. mixdyn-0.1.0/src/mixdyn/experiments/reading.py +660 -0
  149. mixdyn-0.1.0/src/mixdyn/experiments/sdid.py +517 -0
  150. mixdyn-0.1.0/src/mixdyn/experiments/switchback.py +609 -0
  151. mixdyn-0.1.0/src/mixdyn/experiments/value.py +1273 -0
  152. mixdyn-0.1.0/src/mixdyn/inference/__init__.py +6 -0
  153. mixdyn-0.1.0/src/mixdyn/inference/compiled.py +287 -0
  154. mixdyn-0.1.0/src/mixdyn/inference/diagnostics.py +587 -0
  155. mixdyn-0.1.0/src/mixdyn/inference/loo.py +253 -0
  156. mixdyn-0.1.0/src/mixdyn/inference/nuts.py +490 -0
  157. mixdyn-0.1.0/src/mixdyn/inference/pathfinder.py +248 -0
  158. mixdyn-0.1.0/src/mixdyn/inference/sensitivity.py +463 -0
  159. mixdyn-0.1.0/src/mixdyn/models/__init__.py +79 -0
  160. mixdyn-0.1.0/src/mixdyn/models/bayes_geo.py +800 -0
  161. mixdyn-0.1.0/src/mixdyn/models/bayes_mmm.py +1864 -0
  162. mixdyn-0.1.0/src/mixdyn/models/demand.py +87 -0
  163. mixdyn-0.1.0/src/mixdyn/models/dlm.py +785 -0
  164. mixdyn-0.1.0/src/mixdyn/models/dlm_choice.py +718 -0
  165. mixdyn-0.1.0/src/mixdyn/models/dlm_decomposition.py +149 -0
  166. mixdyn-0.1.0/src/mixdyn/models/dlm_diagnostics.py +176 -0
  167. mixdyn-0.1.0/src/mixdyn/models/dlm_lift.py +527 -0
  168. mixdyn-0.1.0/src/mixdyn/models/dlm_media.py +526 -0
  169. mixdyn-0.1.0/src/mixdyn/models/geo_dlm.py +382 -0
  170. mixdyn-0.1.0/src/mixdyn/models/geo_eb.py +632 -0
  171. mixdyn-0.1.0/src/mixdyn/models/margin_mmm.py +996 -0
  172. mixdyn-0.1.0/src/mixdyn/models/nested.py +505 -0
  173. mixdyn-0.1.0/src/mixdyn/models/ridge.py +181 -0
  174. mixdyn-0.1.0/src/mixdyn/models/ridge_mmm.py +514 -0
  175. mixdyn-0.1.0/src/mixdyn/models/ridge_search.py +493 -0
  176. mixdyn-0.1.0/src/mixdyn/py.typed +0 -0
  177. mixdyn-0.1.0/src/mixdyn/transforms/__init__.py +30 -0
  178. mixdyn-0.1.0/src/mixdyn/transforms/media.py +144 -0
  179. mixdyn-0.1.0/src/mixdyn/transforms/saturation.py +149 -0
  180. mixdyn-0.1.0/uv.lock +748 -0
  181. mixdyn-0.1.0/validation/aa_audit.md +139 -0
  182. mixdyn-0.1.0/validation/aa_audit.py +299 -0
  183. mixdyn-0.1.0/validation/geo_aa_flag.mac +73 -0
  184. mixdyn-0.1.0/validation/geo_eb.mac +153 -0
  185. mixdyn-0.1.0/validation/half_normal_roi.mac +96 -0
  186. mixdyn-0.1.0/validation/hall_reader.mac +141 -0
  187. mixdyn-0.1.0/validation/lift_logconcave.mac +120 -0
  188. mixdyn-0.1.0/validation/long_term_null.md +316 -0
  189. mixdyn-0.1.0/validation/long_term_null.py +358 -0
  190. mixdyn-0.1.0/validation/margin_noise.mac +110 -0
  191. mixdyn-0.1.0/validation/meridian_calibration_score.mac +68 -0
  192. mixdyn-0.1.0/validation/refresh_prior.mac +75 -0
  193. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-base_prior-normal.json +30676 -0
  194. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-base_prior-normal.md +81 -0
  195. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-centred-search.json +30676 -0
  196. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-centred-search.md +81 -0
  197. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-centred-tv.json +30676 -0
  198. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-centred-tv.md +81 -0
  199. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-noise_prior-log-normal.json +30676 -0
  200. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-noise_prior-log-normal.md +81 -0
  201. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-slope-fixed.json +28253 -0
  202. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo-slope-fixed.md +76 -0
  203. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo.json +30674 -0
  204. mixdyn-0.1.0/validation/sbc/bayes_geo-centred-geo.md +81 -0
  205. mixdyn-0.1.0/validation/sbc/bayes_geo-small.json +26406 -0
  206. mixdyn-0.1.0/validation/sbc/bayes_geo-small.md +81 -0
  207. mixdyn-0.1.0/validation/sbc/bayes_mmm-small.json +15187 -0
  208. mixdyn-0.1.0/validation/sbc/bayes_mmm-small.md +57 -0
  209. mixdyn-0.1.0/validation/sbc/dlm-small.json +9065 -0
  210. mixdyn-0.1.0/validation/sbc/dlm-small.md +31 -0
  211. mixdyn-0.1.0/validation/sbc/nested-small.json +18409 -0
  212. mixdyn-0.1.0/validation/sbc/nested-small.md +70 -0
  213. mixdyn-0.1.0/validation/sbc.py +838 -0
  214. mixdyn-0.1.0/validation/sdid_simplex_weights.mac +100 -0
  215. mixdyn-0.1.0/validation/sensitivity.mac +140 -0
  216. mixdyn-0.1.0/validation/sign_flip_interval.mac +78 -0
  217. mixdyn-0.1.0/validation/switchback_carryover.mac +168 -0
  218. mixdyn-0.1.0/validation/time_guard.mac +97 -0
@@ -0,0 +1,116 @@
1
+ name: ci
2
+
3
+ on:
4
+ push:
5
+ branches: [main]
6
+ pull_request:
7
+
8
+ permissions:
9
+ contents: read
10
+
11
+ jobs:
12
+ check:
13
+ runs-on: ubuntu-latest
14
+ steps:
15
+ - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
16
+ - name: Install uv
17
+ uses: astral-sh/setup-uv@1c37ad07a6a961277cf70c0d37d6f313000f5884 # v10.3.0
18
+ with:
19
+ python-version: "3.14"
20
+ - name: Sync dependencies
21
+ # --locked fails when uv.lock is stale against pyproject.toml rather than quietly
22
+ # re-resolving, so the dependency gate below reads the lock that was reviewed.
23
+ run: uv sync --locked
24
+ - name: Format check
25
+ run: uv run ruff format --check .
26
+ - name: Lint
27
+ run: uv run ruff check .
28
+ - name: Type check
29
+ run: uv run ty check
30
+ - name: Dependency gate
31
+ run: uv run python scripts/check_runtime_deps.py
32
+
33
+ # A posterior test samples four chains of 2000 iterations one after another, minutes each: a geo
34
+ # world's about three times a one-geo world's, a Student-t world's about three times a normal
35
+ # one's (`--durations` prints them). Run in turn they took most of the suite's half hour. The
36
+ # groups below split them by those times, and `rest` runs the suite without exactly the node ids
37
+ # the groups name, so a test no group names, a new world's among them, runs there: removing a
38
+ # group means removing its ids from `rest`'s too. `rest` runs again on the oldest CPythons the
39
+ # engine supports (ADR 0054) and on the newest, its node ids the same list.
40
+ test:
41
+ name: test (${{ matrix.group }})
42
+ runs-on: ubuntu-latest
43
+ strategy:
44
+ fail-fast: false
45
+ matrix:
46
+ include:
47
+ - group: posterior-1
48
+ tests: >-
49
+ tests/test_bayes_geo.py::test_the_posterior_is_stan_s[student]
50
+ - group: posterior-2
51
+ tests: >-
52
+ tests/test_bayes_geo.py::test_the_posterior_is_stan_s[normal]
53
+ tests/test_bayes_geo.py::test_the_posterior_is_stan_s[tanh]
54
+ tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[lognormal-uniform]
55
+ tests/test_bayes_mmm.py::test_on_qr_coordinates_the_posterior_is_stan_s
56
+ - group: posterior-3
57
+ tests: >-
58
+ tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[normal-uniform]
59
+ tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[normal-pathfinder]
60
+ tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[student-uniform]
61
+ tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[tanh-uniform]
62
+ tests/test_nested.py::test_the_joint_posterior_is_stan_s[both]
63
+ tests/test_nested.py::test_the_joint_posterior_is_stan_s[mediated]
64
+ - group: rest
65
+ tests: &rest >-
66
+ tests
67
+ --deselect tests/test_bayes_geo.py::test_the_posterior_is_stan_s[student]
68
+ --deselect tests/test_bayes_geo.py::test_the_posterior_is_stan_s[normal]
69
+ --deselect tests/test_bayes_geo.py::test_the_posterior_is_stan_s[tanh]
70
+ --deselect tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[lognormal-uniform]
71
+ --deselect tests/test_bayes_mmm.py::test_on_qr_coordinates_the_posterior_is_stan_s
72
+ --deselect tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[normal-uniform]
73
+ --deselect tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[normal-pathfinder]
74
+ --deselect tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[student-uniform]
75
+ --deselect tests/test_bayes_mmm.py::test_the_posterior_is_stan_s[tanh-uniform]
76
+ --deselect tests/test_nested.py::test_the_joint_posterior_is_stan_s[both]
77
+ --deselect tests/test_nested.py::test_the_joint_posterior_is_stan_s[mediated]
78
+ - group: rest-3.12
79
+ python: "3.12"
80
+ tests: *rest
81
+ - group: rest-3.13
82
+ python: "3.13"
83
+ tests: *rest
84
+ - group: rest-3.15
85
+ python: "3.15"
86
+ tests: *rest
87
+ steps:
88
+ - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
89
+ - name: Install uv
90
+ uses: astral-sh/setup-uv@1c37ad07a6a961277cf70c0d37d6f313000f5884 # v10.3.0
91
+ with:
92
+ python-version: ${{ matrix.python || '3.14' }}
93
+ - name: Sync dependencies
94
+ run: uv sync --locked
95
+ - name: Cache JAX's compiled programs
96
+ # JAX's own key for a program hashes its HLO, jaxlib's version, the compile options and
97
+ # the CPU's model and features, so a program compiled on another runner model is never
98
+ # loaded here. Programs that compile in under a second are not kept.
99
+ uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
100
+ with:
101
+ path: ${{ runner.temp }}/jax-cache
102
+ key: jax-${{ matrix.group }}-${{ hashFiles('uv.lock') }}-${{ github.sha }}
103
+ restore-keys: jax-${{ matrix.group }}-${{ hashFiles('uv.lock') }}-
104
+ - name: Runner CPU
105
+ # A digest pinned on one CPU class moves on another: OpenBLAS and numpy pick their kernels
106
+ # by the CPU, so the log names the class each leg ran on.
107
+ run: |
108
+ lscpu | grep -E '^Model name:'
109
+ uv run python -c "import numpy; numpy.show_runtime()"
110
+ - name: Tests
111
+ env:
112
+ JAX_COMPILATION_CACHE_DIR: ${{ runner.temp }}/jax-cache
113
+ # set -f: a node id's brackets are not a glob.
114
+ run: |
115
+ set -f
116
+ uv run pytest ${{ matrix.tests }} --durations=10
@@ -0,0 +1,117 @@
1
+ name: release
2
+
3
+ # Cut a release by pushing a signed tag, `git tag -s v0.1.0 -m ... && git push origin v0.1.0`.
4
+ # Publishing uses PyPI Trusted Publishing (OIDC), no stored token. One-time PyPI-side setup: add the
5
+ # repository, this `release.yml` workflow and the `pypi` environment as a trusted publisher of the
6
+ # `mixdyn` project. The build job below stops on a wheel that carries the `Private :: Do Not
7
+ # Upload` classifier, which PyPI refuses, before anything is uploaded.
8
+
9
+ on:
10
+ push:
11
+ tags: ["v*"]
12
+
13
+ # Nothing by default: each job below names exactly the scopes it needs.
14
+ permissions: {}
15
+
16
+ jobs:
17
+ build:
18
+ runs-on: ubuntu-latest
19
+ permissions:
20
+ contents: read
21
+ steps:
22
+ - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
23
+ - name: Install uv
24
+ uses: astral-sh/setup-uv@1c37ad07a6a961277cf70c0d37d6f313000f5884 # v10.3.0
25
+ with:
26
+ # the floor, not the newest: the wheel is `py3-none-any`, so the build interpreter does not
27
+ # reach the artefact, and building on the oldest supported one keeps that true
28
+ python-version: "3.12"
29
+ - name: Build sdist and wheel
30
+ run: uv build
31
+ # The tag names the version the wheel must carry, and the wheel must be uploadable: PyPI
32
+ # refuses the private classifier, and a refusal after the sdist went up would leave half a
33
+ # release on the index.
34
+ - name: Check the wheel's version, licence and classifiers
35
+ run: |
36
+ version="${GITHUB_REF_NAME#v}"
37
+ wheel=$(ls dist/*.whl)
38
+ case "$wheel" in
39
+ "dist/mixdyn-$version-py3-none-any.whl") ;;
40
+ *) echo "the wheel $wheel does not carry the tag's version $version"; exit 1 ;;
41
+ esac
42
+ unzip -p "$wheel" '*.dist-info/METADATA' > METADATA
43
+ grep -qx 'License-Expression: Apache-2.0' METADATA
44
+ if grep -q 'Private :: Do Not Upload' METADATA; then
45
+ echo "the wheel still carries the private classifier: remove it from pyproject.toml"
46
+ exit 1
47
+ fi
48
+ # The notes are the changelog's section for this version, not a pointer to it, extracted so
49
+ # the two cannot drift. They are built here, before the upload, so a version without a
50
+ # section stops the release before PyPI has it.
51
+ - name: Build the release notes from CHANGELOG.md
52
+ run: |
53
+ version="${GITHUB_REF_NAME#v}"
54
+ awk -v v="## [$version]" '
55
+ index($0, v) == 1 { found = 1 }
56
+ found && /^## \[/ && index($0, v) != 1 { exit }
57
+ found { print }
58
+ ' CHANGELOG.md > notes.md
59
+ if [ ! -s notes.md ]; then
60
+ echo "CHANGELOG.md has no '## [$version]' section: refusing to publish empty notes"
61
+ exit 1
62
+ fi
63
+ # GitHub caps a release body at 125 000 characters and rejects the call above it.
64
+ if [ "$(wc -c < notes.md)" -gt 120000 ]; then
65
+ head -c 119000 notes.md > notes.trimmed
66
+ printf '\n\n_Truncated; the full entry is in [CHANGELOG.md](CHANGELOG.md)._\n' \
67
+ >> notes.trimmed
68
+ mv notes.trimmed notes.md
69
+ fi
70
+ wc -c notes.md
71
+ - uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
72
+ with:
73
+ name: dist
74
+ path: dist/
75
+ - uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
76
+ with:
77
+ name: notes
78
+ path: notes.md
79
+
80
+ publish:
81
+ needs: build
82
+ runs-on: ubuntu-latest
83
+ environment: pypi
84
+ permissions:
85
+ id-token: write # OIDC token for PyPI Trusted Publishing
86
+ steps:
87
+ - uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131 # v7.0.0
88
+ with:
89
+ name: dist
90
+ path: dist/
91
+ - name: Publish to PyPI
92
+ uses: pypa/gh-action-pypi-publish@dc37677b2e1c63e2034f94d8a5b11f265b73ba33 # v1.14.2
93
+
94
+ # Last on purpose: a Release is the claim that this version is out, and PyPI is the step that can
95
+ # still fail.
96
+ github-release:
97
+ needs: publish
98
+ runs-on: ubuntu-latest
99
+ permissions:
100
+ contents: write # create the Release object
101
+ steps:
102
+ - uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131 # v7.0.0
103
+ with:
104
+ name: dist
105
+ path: dist/
106
+ - uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131 # v7.0.0
107
+ with:
108
+ name: notes
109
+ path: .
110
+ - name: Create the GitHub Release
111
+ env:
112
+ GH_TOKEN: ${{ github.token }}
113
+ run: |
114
+ gh release create "$GITHUB_REF_NAME" dist/* \
115
+ --repo "$GITHUB_REPOSITORY" \
116
+ --title "$GITHUB_REF_NAME" \
117
+ --notes-file notes.md
@@ -0,0 +1,11 @@
1
+ __pycache__/
2
+ *.py[cod]
3
+ .venv/
4
+ .venv-*/
5
+ build/
6
+ dist/
7
+ *.egg-info/
8
+ .pytest_cache/
9
+ .ruff_cache/
10
+ .hypothesis/
11
+ .coverage
@@ -0,0 +1 @@
1
+ 3.14
@@ -0,0 +1,95 @@
1
+ # Changelog
2
+
3
+ Each release's changes. A release's notes on GitHub are its section here, extracted by
4
+ `.github/workflows/release.yml`. The project follows the versioning rules of `docs/api/index.md`:
5
+ before 1.0 a minor release may change the public surface, and a patch release never removes or
6
+ changes a public name.
7
+
8
+ ## [0.1.0] — 2026-11-20
9
+
10
+ The first public release. The engine's history before it is not public; this section lists what
11
+ the release holds.
12
+
13
+ ### Measured
14
+
15
+ - **A pre-registered non-inferiority claim on the causaldyn bench (run M3).** On simulated worlds
16
+ whose truth is known, mixdyn's plans are within δ = 0.05 of regret per euro of budget of those
17
+ of Meridian 2.1.0, PyMC-Marketing 1.3.0, Orbit 1.1.5.1, the status quo and the equal split, on
18
+ 17 pre-registered family-rungs. Each verdict was tested at one-sided 0.025 in a fixed sequence,
19
+ on fresh confirmatory seeds: 17 of 17 held, against 16.94 expected.
20
+ - At k = 0 (history only) `margin held` has a lower regret than every rival on every
21
+ confirmatory national family. At the tested rung (four go-dark lift tests) `gs held`, and on
22
+ the geo families `geo eb`, are level with PyMC-Marketing and have a lower regret than the
23
+ other rivals.
24
+ - **Losses, stated with the claim:** S-curves with no lift test (family 4, k = 0) lose to
25
+ PyMC-Marketing by +0.184 [+0.136, +0.232] and to Orbit by +0.171 [+0.118, +0.223] a euro, not
26
+ within δ. At the tested rung, drifting effects (family 1) and truth outside the history's range
27
+ (family 6) lose to PyMC-Marketing by +0.029 [+0.002, +0.055] and +0.039 [+0.022, +0.055], not
28
+ within δ. Holm's step-down finds two losses within δ, families 3 and 5 at the tested rung
29
+ against PyMC-Marketing (+0.008 and +0.020).
30
+ - **No headline:** families 12 and 13 at k = 0, and family 12 at the tested rung, needed more
31
+ worlds than the run's cap and are exploratory, so "within δ of every maintained MMM" is not
32
+ said. Bench families 11, 16 and 21 were not in the run. PyMC-Marketing ran its static model.
33
+ - The pre-registration is `scorecard/preregistration/m3.md` (SHA-256
34
+ `d684b25ccae0e3574fe2f962097e399c9f400f0641e52431cf794e131763d924`), frozen on 2026-11-17 in
35
+ the then-private repository: there is no public time-stamp before the run. The README states
36
+ the full table, every loss and what the run did not cover.
37
+
38
+ ### Added
39
+
40
+ - **Data** (`mixdyn.data`, `read_panel`): a weekly or daily panel of a KPI, media spend, controls
41
+ and levers, national or over a geo hierarchy, read from dicts, pandas or polars frames and
42
+ checked, every finding an issue with a code; national media allotted to geos by population;
43
+ recurring events read off the panel's own records; the flag for spend that reads the KPI
44
+ (ADR 0074).
45
+ - **Transforms** (`mixdyn.transforms`): adstock kernels and saturation curves in two families,
46
+ each in its engine's convention (ADR 0064).
47
+ - **Engines** (`mixdyn.models`):
48
+ - the ridge engine, its search over transforms and its calibration by geo tests (ADR 0008);
49
+ - the discount dynamic linear model over media, with its decomposition, diagnostics, discount
50
+ choice and lift tests read by the filter (ADRs 0020, 0038, 0066);
51
+ - the Bayesian engine, one geo and over geos, with returns as parameters and lift and geo tests
52
+ in the likelihood, sampled by the engine's own NUTS started by Pathfinder (ADRs 0009 to 0015,
53
+ 0018, 0036);
54
+ - the nested model, media through a mediator to the KPI (ADRs 0014, 0033);
55
+ - the margin model: each channel's margin by regression under a mixture of error models,
56
+ linear or curved, with empirical-Bayes shrinkage (ADRs 0086, 0087, 0091);
57
+ - the closed-form model over geos fitted by empirical Bayes (ADR 0090);
58
+ - a demand proxy, query volume for paid search, and the one role a model gives it (ADR 0074).
59
+ - **Inference** (`mixdyn.inference`): NUTS, Pathfinder, convergence diagnostics with thresholds
60
+ read where Stan reads them (ADR 0065), PSIS-LOO and the stacking of models' predictive
61
+ distributions, the health card (ADR 0060), and sensitivity to an
62
+ unobserved confounder: Cinelli and Hazlett's robustness values and the marginal sensitivity
63
+ model's `Gamma*` (ADR 0061).
64
+ - **Experiments** (`mixdyn.experiments`): geo-test readers (difference in differences, time-based
65
+ regression, synthetic control, synthetic difference in differences), a switchback reader, geo
66
+ designs randomised within matched pairs with their power, and the value of a lift test to the
67
+ plan and the set of tests chosen by it (ADRs 0030 to 0032, 0047 to 0049, 0055, 0056).
68
+ - **Calibration** (`mixdyn.calibration`): lift estimates and their tables, and margins anchored on
69
+ lift tests (ADRs 0015, 0088, 0089).
70
+ - **Decisions** (`mixdyn.decisions`): one definition of each return, plans for a posterior's
71
+ expected return, its worst share and a target return, the no-test plan against the plan in
72
+ place, flighting, scenarios, the plan table, geo plans, and the margin rule's plans with its
73
+ observed-range box and time guard (ADRs 0016, 0019, 0029, 0034, 0044, 0050 to 0052, 0071, 0092).
74
+ - **API** (`mixdyn.api`): a factsheet of a fitted model, the decision on one page, the estimands of
75
+ the tests read, a file format that pickles nothing (`save`, `load`, `FORMAT_VERSION` 1) and
76
+ versioned JSON documents (ADRs 0062, 0063, 0072).
77
+ - **The scorecard's harness** (`scorecard/`, in the repository, not the wheel): declares a run,
78
+ exports the bench's worlds, runs every arm, this engine's and the rivals', scores every record
79
+ and reads the pre-registered claim (ADRs 0021, 0039, 0040, 0080, 0081). It needs the bench
80
+ (`causaldyn-bench`) on `PYTHONPATH`.
81
+ - `SKILL.md`, an Agent Skill that walks one panel through read, fit, diagnostics, tests, plan and
82
+ report, each step gated by the engine's own checks; `scripts/real_data_protocol.py`, the
83
+ protocol a partner would run to score predictions of withheld lift tests on real data.
84
+
85
+ ### Packaging
86
+
87
+ - On PyPI as `mixdyn`, Apache-2.0 (`License-Expression: Apache-2.0`, with `LICENSE` and `NOTICE`
88
+ in the wheel and the sdist), contributions under the Developer Certificate of Origin
89
+ (`CONTRIBUTING.md`; ADRs 0054 and 0095).
90
+ - Python 3.12 or later; CI runs the suite on 3.12, 3.13, 3.14 and 3.15.
91
+ - Runtime dependencies: `causal-hybrid-control>=1.2.0`, `jax>=0.4.30`, `numpy>=2.3`,
92
+ `scipy>=1.16`; `uv.lock` locks causal-hybrid-control 1.3.1. The packages the engine may install
93
+ at run time are an allow-list that CI checks against the lock (ADR 0001).
94
+ - The wheel holds `mixdyn` alone; the sdist leaves out `evidence/` and `scorecard/`.
95
+ - Releases are published from a signed tag through PyPI's Trusted Publishing, with no stored token.
@@ -0,0 +1,64 @@
1
+ # Contributing
2
+
3
+ Issues and pull requests are welcome. Please open an issue before a large change, so that two people
4
+ do not build the same thing differently.
5
+
6
+ ## Setup
7
+
8
+ ```bash
9
+ just sync # .venv exactly as uv.lock has it
10
+ just check # format, lint, types, the dependency gate, tests: what CI runs
11
+ ```
12
+
13
+ Python 3.12 and up; development runs on 3.14 (`.python-version`). `uv.lock` is committed and
14
+ authoritative: run everything through `uv run`, never a bare `python`, `pytest` or `ruff`, or you
15
+ are testing a different dependency set from CI's.
16
+
17
+ ## The gates
18
+
19
+ CI runs these, and a pull request is expected to pass them locally first:
20
+
21
+ ```bash
22
+ uv run ruff format --check .
23
+ uv run ruff check .
24
+ uv run ty check # at Python 3.12, the lowest supported version
25
+ uv run python scripts/check_runtime_deps.py
26
+ uv run pytest
27
+ ```
28
+
29
+ The full suite samples several posteriors and takes a while; CI splits it into groups
30
+ (`.github/workflows/ci.yml`) and runs the rest of it on Python 3.12, 3.13, 3.14 and 3.15.
31
+
32
+ ## What a change should look like
33
+
34
+ - **Runtime packages are an allow-list** (`allowed-runtime-packages.txt`, ADR 0001). A new one needs
35
+ an issue first. Special-purpose packages (a probabilistic-programming language, an MMM library, a
36
+ statistics framework) are used as oracles in a scratch environment, never in the lock.
37
+ - **Every numerical routine has an independent oracle**: a closed form, another implementation,
38
+ coverage over seeds, or simulation-based calibration. Say which in the test's docstring.
39
+ - **Tests are two-sided.** A test that only shows the good case passing does not show that the bad
40
+ case fails; prefer one that fails when the implementation is reverted.
41
+ - **No number in a docstring, the README or the changelog that a test or a script does not
42
+ produce.**
43
+ - **Scope claims precisely.** If a result holds only for one geo, a known variance or a linear
44
+ response, the docstring says so.
45
+ - **A claim that mixdyn beats another tool** names its axis, its family of worlds and its paired
46
+ interval from a pre-registered scorecard run, and states the losses beside it.
47
+ - A decision the code will rely on gets an ADR in `docs/adr/`: context, decision, consequences,
48
+ alternatives.
49
+ - Everything written, code, comments, docstrings, commit messages and issues, is in English.
50
+ - Commit messages use a conventional-commit prefix (`feat:`, `fix:`, `refactor:`, `test:`,
51
+ `docs:`, `chore:`) and explain the *why* in the body.
52
+
53
+ ## Sign-off
54
+
55
+ Contributions are accepted under the [Developer Certificate of Origin](https://developercertificate.org/)
56
+ (DCO) 1.1: by signing off a commit you certify that you wrote it, or otherwise have the right to
57
+ submit it under the project's licence. Sign off each commit with `git commit -s`, which adds
58
+
59
+ ```
60
+ Signed-off-by: Your Name <you@example.org>
61
+ ```
62
+
63
+ A commit without a sign-off is not merged. There is no contributor licence agreement.
64
+ Contributions are licensed under the project's licence, Apache-2.0 (ADRs 0054 and 0095).
mixdyn-0.1.0/LICENSE ADDED
@@ -0,0 +1,202 @@
1
+
2
+ Apache License
3
+ Version 2.0, January 2004
4
+ http://www.apache.org/licenses/
5
+
6
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
7
+
8
+ 1. Definitions.
9
+
10
+ "License" shall mean the terms and conditions for use, reproduction,
11
+ and distribution as defined by Sections 1 through 9 of this document.
12
+
13
+ "Licensor" shall mean the copyright owner or entity authorized by
14
+ the copyright owner that is granting the License.
15
+
16
+ "Legal Entity" shall mean the union of the acting entity and all
17
+ other entities that control, are controlled by, or are under common
18
+ control with that entity. For the purposes of this definition,
19
+ "control" means (i) the power, direct or indirect, to cause the
20
+ direction or management of such entity, whether by contract or
21
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
22
+ outstanding shares, or (iii) beneficial ownership of such entity.
23
+
24
+ "You" (or "Your") shall mean an individual or Legal Entity
25
+ exercising permissions granted by this License.
26
+
27
+ "Source" form shall mean the preferred form for making modifications,
28
+ including but not limited to software source code, documentation
29
+ source, and configuration files.
30
+
31
+ "Object" form shall mean any form resulting from mechanical
32
+ transformation or translation of a Source form, including but
33
+ not limited to compiled object code, generated documentation,
34
+ and conversions to other media types.
35
+
36
+ "Work" shall mean the work of authorship, whether in Source or
37
+ Object form, made available under the License, as indicated by a
38
+ copyright notice that is included in or attached to the work
39
+ (an example is provided in the Appendix below).
40
+
41
+ "Derivative Works" shall mean any work, whether in Source or Object
42
+ form, that is based on (or derived from) the Work and for which the
43
+ editorial revisions, annotations, elaborations, or other modifications
44
+ represent, as a whole, an original work of authorship. For the purposes
45
+ of this License, Derivative Works shall not include works that remain
46
+ separable from, or merely link (or bind by name) to the interfaces of,
47
+ the Work and Derivative Works thereof.
48
+
49
+ "Contribution" shall mean any work of authorship, including
50
+ the original version of the Work and any modifications or additions
51
+ to that Work or Derivative Works thereof, that is intentionally
52
+ submitted to Licensor for inclusion in the Work by the copyright owner
53
+ or by an individual or Legal Entity authorized to submit on behalf of
54
+ the copyright owner. For the purposes of this definition, "submitted"
55
+ means any form of electronic, verbal, or written communication sent
56
+ to the Licensor or its representatives, including but not limited to
57
+ communication on electronic mailing lists, source code control systems,
58
+ and issue tracking systems that are managed by, or on behalf of, the
59
+ Licensor for the purpose of discussing and improving the Work, but
60
+ excluding communication that is conspicuously marked or otherwise
61
+ designated in writing by the copyright owner as "Not a Contribution."
62
+
63
+ "Contributor" shall mean Licensor and any individual or Legal Entity
64
+ on behalf of whom a Contribution has been received by Licensor and
65
+ subsequently incorporated within the Work.
66
+
67
+ 2. Grant of Copyright License. Subject to the terms and conditions of
68
+ this License, each Contributor hereby grants to You a perpetual,
69
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
70
+ copyright license to reproduce, prepare Derivative Works of,
71
+ publicly display, publicly perform, sublicense, and distribute the
72
+ Work and such Derivative Works in Source or Object form.
73
+
74
+ 3. Grant of Patent License. Subject to the terms and conditions of
75
+ this License, each Contributor hereby grants to You a perpetual,
76
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
77
+ (except as stated in this section) patent license to make, have made,
78
+ use, offer to sell, sell, import, and otherwise transfer the Work,
79
+ where such license applies only to those patent claims licensable
80
+ by such Contributor that are necessarily infringed by their
81
+ Contribution(s) alone or by combination of their Contribution(s)
82
+ with the Work to which such Contribution(s) was submitted. If You
83
+ institute patent litigation against any entity (including a
84
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
85
+ or a Contribution incorporated within the Work constitutes direct
86
+ or contributory patent infringement, then any patent licenses
87
+ granted to You under this License for that Work shall terminate
88
+ as of the date such litigation is filed.
89
+
90
+ 4. Redistribution. You may reproduce and distribute copies of the
91
+ Work or Derivative Works thereof in any medium, with or without
92
+ modifications, and in Source or Object form, provided that You
93
+ meet the following conditions:
94
+
95
+ (a) You must give any other recipients of the Work or
96
+ Derivative Works a copy of this License; and
97
+
98
+ (b) You must cause any modified files to carry prominent notices
99
+ stating that You changed the files; and
100
+
101
+ (c) You must retain, in the Source form of any Derivative Works
102
+ that You distribute, all copyright, patent, trademark, and
103
+ attribution notices from the Source form of the Work,
104
+ excluding those notices that do not pertain to any part of
105
+ the Derivative Works; and
106
+
107
+ (d) If the Work includes a "NOTICE" text file as part of its
108
+ distribution, then any Derivative Works that You distribute must
109
+ include a readable copy of the attribution notices contained
110
+ within such NOTICE file, excluding those notices that do not
111
+ pertain to any part of the Derivative Works, in at least one
112
+ of the following places: within a NOTICE text file distributed
113
+ as part of the Derivative Works; within the Source form or
114
+ documentation, if provided along with the Derivative Works; or,
115
+ within a display generated by the Derivative Works, if and
116
+ wherever such third-party notices normally appear. The contents
117
+ of the NOTICE file are for informational purposes only and
118
+ do not modify the License. You may add Your own attribution
119
+ notices within Derivative Works that You distribute, alongside
120
+ or as an addendum to the NOTICE text from the Work, provided
121
+ that such additional attribution notices cannot be construed
122
+ as modifying the License.
123
+
124
+ You may add Your own copyright statement to Your modifications and
125
+ may provide additional or different license terms and conditions
126
+ for use, reproduction, or distribution of Your modifications, or
127
+ for any such Derivative Works as a whole, provided Your use,
128
+ reproduction, and distribution of the Work otherwise complies with
129
+ the conditions stated in this License.
130
+
131
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
132
+ any Contribution intentionally submitted for inclusion in the Work
133
+ by You to the Licensor shall be under the terms and conditions of
134
+ this License, without any additional terms or conditions.
135
+ Notwithstanding the above, nothing herein shall supersede or modify
136
+ the terms of any separate license agreement you may have executed
137
+ with Licensor regarding such Contributions.
138
+
139
+ 6. Trademarks. This License does not grant permission to use the trade
140
+ names, trademarks, service marks, or product names of the Licensor,
141
+ except as required for reasonable and customary use in describing the
142
+ origin of the Work and reproducing the content of the NOTICE file.
143
+
144
+ 7. Disclaimer of Warranty. Unless required by applicable law or
145
+ agreed to in writing, Licensor provides the Work (and each
146
+ Contributor provides its Contributions) on an "AS IS" BASIS,
147
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
148
+ implied, including, without limitation, any warranties or conditions
149
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
150
+ PARTICULAR PURPOSE. You are solely responsible for determining the
151
+ appropriateness of using or redistributing the Work and assume any
152
+ risks associated with Your exercise of permissions under this License.
153
+
154
+ 8. Limitation of Liability. In no event and under no legal theory,
155
+ whether in tort (including negligence), contract, or otherwise,
156
+ unless required by applicable law (such as deliberate and grossly
157
+ negligent acts) or agreed to in writing, shall any Contributor be
158
+ liable to You for damages, including any direct, indirect, special,
159
+ incidental, or consequential damages of any character arising as a
160
+ result of this License or out of the use or inability to use the
161
+ Work (including but not limited to damages for loss of goodwill,
162
+ work stoppage, computer failure or malfunction, or any and all
163
+ other commercial damages or losses), even if such Contributor
164
+ has been advised of the possibility of such damages.
165
+
166
+ 9. Accepting Warranty or Additional Liability. While redistributing
167
+ the Work or Derivative Works thereof, You may choose to offer,
168
+ and charge a fee for, acceptance of support, warranty, indemnity,
169
+ or other liability obligations and/or rights consistent with this
170
+ License. However, in accepting such obligations, You may act only
171
+ on Your own behalf and on Your sole responsibility, not on behalf
172
+ of any other Contributor, and only if You agree to indemnify,
173
+ defend, and hold each Contributor harmless for any liability
174
+ incurred by, or claims asserted against, such Contributor by reason
175
+ of your accepting any such warranty or additional liability.
176
+
177
+ END OF TERMS AND CONDITIONS
178
+
179
+ APPENDIX: How to apply the Apache License to your work.
180
+
181
+ To apply the Apache License to your work, attach the following
182
+ boilerplate notice, with the fields enclosed by brackets "[]"
183
+ replaced with your own identifying information. (Don't include
184
+ the brackets!) The text should be enclosed in the appropriate
185
+ comment syntax for the file format. We also recommend that a
186
+ file or class name and description of purpose be included on the
187
+ same "printed page" as the copyright notice for easier
188
+ identification within third-party archives.
189
+
190
+ Copyright [yyyy] [name of copyright owner]
191
+
192
+ Licensed under the Apache License, Version 2.0 (the "License");
193
+ you may not use this file except in compliance with the License.
194
+ You may obtain a copy of the License at
195
+
196
+ http://www.apache.org/licenses/LICENSE-2.0
197
+
198
+ Unless required by applicable law or agreed to in writing, software
199
+ distributed under the License is distributed on an "AS IS" BASIS,
200
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
201
+ See the License for the specific language governing permissions and
202
+ limitations under the License.
mixdyn-0.1.0/NOTICE ADDED
@@ -0,0 +1,2 @@
1
+ mixdyn
2
+ Copyright 2026 Ilia Gradina