ArtaModel IV — genderless, even (|Δθ|), every pair in both orders, every long-term relationship; edition III kept below

README.md

@@ -1,13 +1,98 @@
 ---
 license: cc0-1.0
 language: [en]
-tags: [astrology, sidereal, jyotisha, marriage, tabular, phase-model, artaquest]
-datasets: [artaquest-foundation/artamatch-sidereal]
+tags: [astrology, sidereal, jyotisha, relationships, tabular, phase-model, genderless, artaquest]
+datasets: [artaquest-foundation/artamatch-genderless, artaquest-foundation/artamatch-sidereal]
 metrics: [roc_auc]
 library_name: numpy
 ---
 
-# ArtaModel
+# ArtaModel IV — genderless
+
+**A genderless sidereal phase model of a long-term relationship**, the fourth edition of ArtaModel (Arash
+Ashrafnejad, ArtaQuest Foundation, 2026-08-19: "I want a genderless model from now on"). Two births and birthplaces
+— **in no order** — and the date the relationship began, one probability out: did it last thirty years?
+
+```
+y = | b + Σᵢ  aᵢ ·e^{i|θ1ᵢ − θ2ᵢ|}     the absolute synastry angle          (even under the swap)
+             + t1ᵢ·e^{i|θtᵢ − θ1ᵢ|}     the wedding sky to partner 1's chart
+             + t2ᵢ·e^{i|θtᵢ − θ2ᵢ|}     the wedding sky to partner 2's chart
+             + n1ᵢ·e^{i θ1ᵢ}            partner 1's own natal longitude
+             + n2ᵢ·e^{i θ2ᵢ}            partner 2's own natal longitude
+             + tnᵢ·e^{i θtᵢ} |²         the wedding sky itself
+```
+
+for each of fourteen bodies *i* (Sun, Moon, Mercury, Venus, Mars, Jupiter, Saturn, Uranus, Neptune, Pluto, Rāhu,
+Ketu, Chiron, Lilith). θ are **sidereal longitudes (Lahiri)** from Kerykeion (Swiss Ephemeris): births cast at
+**09:00 local time at the birthplace** (nobody's birth time is recorded — the dataset's convention), the start at
+12:00 UT. **Genderless, three ways:** no sex is read; every phase *difference* enters as its wrapped absolute value
+|Δθ| ∈ [0°, 180°], so each term is an even function of the swap; the training data carries every pair in both
+orders; and the scorer averages the two orders, so the answer is identical whichever way the partners are given.
+**Every long-term relationship** in Wikidata is in the data — marriages of every kind (same-sex included),
+unmarried partnerships, business and sporting partnerships, "significant person" pairs with family excluded.
+**A term exists only when both of its phases exist**: an unknown start day drops the wedding-sky terms, an unknown
+birth drops that partner's terms; a missing phase contributes exactly zero.
+
+## The deployed model, term by term
+
+Gradient boosting over **split single-sum fields**: each stage is one field `|bₖ + wₖ·e^{iφ}|²` on **one** phasor,
+chosen greedily as the phasor that best explains the current residual (all 84 phasors of all six terms compete at
+every stage), added to the logit as `stepₖ·(αₖ·u + cₖ)`. Fitted on **all the data — train and test rows with both
+natal charts, 35,894 rows (every pair in both orders)** — for 8 stages (the number the
+train-only fit chose on its inner temporal split). Of the 84 phasors offered it chose **3**:
+
+| phasor | body | term | stages | contribution to the logit swing | phase at which the field peaks |
+|---|---|---|---|---|---|
+| `a_uranus` | Uranus | a | 4 | 0.151 | 176° |
+| `t2_neptune` | Neptune | t2 | 2 | 0.099 | 195° |
+| `t1_neptune` | Neptune | t1 | 2 | 0.081 | 331° |
+
+Never chosen, at any stage:
+- **a** (a·e^{i|θ1−θ2|}): 13 bodies never chosen — sun, moon, mercury, venus, mars, jupiter, saturn, neptune, pluto, true_node, true_south_node, chiron, mean_lilith
+- **t1** (t1·e^{i|θt−θ1|}): 13 bodies never chosen — sun, moon, mercury, venus, mars, jupiter, saturn, uranus, pluto, true_node, true_south_node, chiron, mean_lilith
+- **t2** (t2·e^{i|θt−θ2|}): 13 bodies never chosen — sun, moon, mercury, venus, mars, jupiter, saturn, uranus, pluto, true_node, true_south_node, chiron, mean_lilith
+- **n1** (n1·e^{iθ1}): 14 bodies never chosen — sun, moon, mercury, venus, mars, jupiter, saturn, uranus, neptune, pluto, true_node, true_south_node, chiron, mean_lilith
+- **n2** (n2·e^{iθ2}): 14 bodies never chosen — sun, moon, mercury, venus, mars, jupiter, saturn, uranus, neptune, pluto, true_node, true_south_node, chiron, mean_lilith
+- **tn** (tn·e^{iθt}): 14 bodies never chosen — sun, moon, mercury, venus, mars, jupiter, saturn, uranus, neptune, pluto, true_node, true_south_node, chiron, mean_lilith
+
+**Read plainly:** `a_uranus` is the absolute gap between the two births measured by Uranus (4.3°/yr);
+`t1_neptune` and `t2_neptune` are each partner's age at the start measured by Neptune (2.2°/yr) — chosen as a pair,
+as a genderless model should. No natal phase, no wedding-sky phase, no fast body.
+
+## What it scores, honestly
+
+| on pairs born after 1900 (temporal hold-out; 7,631 pairs, both orders, symmetrised) | AUC |
+|---|---|
+| ArtaModel IV, fitted on train alone (inner split 0.6326, 8 stages) | **0.6252** (public board 0.6101) |
+| the plain columns — the two ages at the start, the absolute gap, the start year (LightGBM) | 0.6114 (public board 0.6000) |
+| equal-weight rank average of the two | 0.6289 (public board 0.6144) |
+| ArtaModel IV with the two ages held flat (AUC within 3-year age cells) | 0.5659 |
+
+What the model reads is the two partners' ages at the start and the absolute gap between their births, through the
+outer planets as clocks — the same finding as every edition before it, now without a sex anywhere in the model. It
+is exactly invariant to the ayanāṁśa, the birth hour and the birthplace; the age-cell-matched row is what is left
+once the ages are held flat. See `ARTAMODEL.md` for the study (editions III and IV).
+
+## Use
+
+```python
+from artamodel_score_iv import predict       # needs: numpy, kerykeion, timezonefinder
+r = predict("1936-08-04", 37.943, 23.647,     # partner 1: dob, lat, lon
+            "1924-05-14", 37.727, 26.909,     # partner 2 — the order does not matter
+            "1968-06-15")                     # start date (YYYY-01-01 = year only -> wedding-sky terms dropped)
+r["probability"], r["terms"], r["terms_swapped"]
+```
+
+`artamodel_iv_deployed.json` holds every stage's weights; `artamodel_iv.py` is the fit (with `artamodel.py`,
+`artamodel_deploy.py`, `kerykeion_phases.py`); `artamodel_iv.json` the leaderboard numbers. Data:
+[artaquest-foundation/artamatch-genderless](https://www.kaggle.com/datasets/artaquest-foundation/artamatch-genderless);
+competition: [artamatch-genderless](https://www.kaggle.com/competitions/artamatch-genderless). CC0.
+
+---
+
+## Edition III (superseded 2026-08-19): the gendered model, kept for the record
+
+### ArtaModel (third edition)
 
 **A sidereal phase model of a marriage**, named by Arash Ashrafnejad (ArtaQuest Foundation, 2026-08-18). Three dates
 and two places in — his birth and birthplace, hers, and the wedding date — one probability out: did the marriage

artamodel.py

@@ -36,6 +36,9 @@ from coherent_fit import Coherent, auc                                    # noqa
 
 TERMS = ("a", "m", "d", "mn", "dn", "tn")
 TERMS8 = TERMS + ("c", "tc")
+# Arash, 2026-08-19: "add the midpoint of natals as another term. also midpoint with their wedding. overall 3 more
+# terms" -> c (midpoint of the two natals), mw (midpoint of mom's natal with the wedding), dw (dad's with the wedding)
+TERMS9 = TERMS + ("c", "mw", "dw")
 BODIES14 = ["sun", "moon", "mercury", "venus", "mars", "jupiter", "saturn", "uranus", "neptune", "pluto",
             "true_node", "true_south_node", "chiron", "mean_lilith"]
 ANGLES = ["ascendant", "medium_coeli"]
@@ -51,27 +54,52 @@ def composite(D, M):
     return (D + diff / 2.0) % 360.0
 
 
-def phase_matrix(D, M, W, all_bodies, bodies, terms, angles_in_natal=False):
-    """(n, K) phases in degrees (NaN = the term does not exist for that row) and K labels."""
+# FOURTH EDITION — GENDERLESS (operator 2026-08-19): "I want a genderless model from now on ... (a, b, 1) should
+# also mean (b, a, 1) ... for each subtractive term add abs to ensure each term is an even function." The two
+# natal charts are slot 1 and slot 2 with no meaning attached; the files carry every pair in BOTH orders; and
+# every phase DIFFERENCE enters as its wrapped absolute value |Δθ| in [0°, 180°], so a term's value is unchanged
+# when the partners swap (and the wedding-sky terms are even in the same sense: |θt − θ|). Term names for this
+# edition: a (synastry |θ1 − θ2|), t1/t2 (the wedding sky to each partner, |θt − θ|), n1/n2 (each natal phase),
+# tn (the wedding sky's own phase). They map onto the earlier a/m/d/mn/dn/tn computations with `even=True`.
+TERMS_IV = ("a", "t1", "t2", "n1", "n2", "tn")
+_IV_TO_III = {"t1": "d", "t2": "m", "n1": "dn", "n2": "mn"}
+
+
+def absdiff(x, y):
+    """The wrapped absolute phase difference |x − y| in [0, 180] degrees — even in (x, y)."""
+    return np.abs((x - y + 180.0) % 360.0 - 180.0)
+
+
+def phase_matrix(D, M, W, all_bodies, bodies, terms, angles_in_natal=False, even=False):
+    """(n, K) phases in degrees (NaN = the term does not exist for that row) and K labels.
+    even=True: every subtractive term is the wrapped absolute difference (the genderless edition)."""
     col = {b: j for j, b in enumerate(all_bodies)}
     P, lab = [], []
     use = list(bodies) + (ANGLES if angles_in_natal else [])
-    for t in terms:
+    diff = absdiff if even else (lambda x, y: x - y)
+    for t0 in terms:
+        t = _IV_TO_III.get(t0, t0)
         for b in use:
             j = col[b]
             if t == "a":
-                P.append(M[:, j] - D[:, j])
+                P.append(diff(M[:, j], D[:, j]))
             elif t == "c":
                 P.append(composite(D[:, j], M[:, j]))
             elif t == "tc":
                 if b in ANGLES: continue
-                P.append(W[:, j] - composite(D[:, j], M[:, j]))
+                P.append(diff(W[:, j], composite(D[:, j], M[:, j])))
+            elif t == "mw":
+                if b in ANGLES: continue
+                P.append(composite(M[:, j], W[:, j]))          # midpoint of mom's natal and the wedding sky
+            elif t == "dw":
+                if b in ANGLES: continue
+                P.append(composite(D[:, j], W[:, j]))          # midpoint of dad's natal and the wedding sky
             elif t == "m":
                 if b in ANGLES: continue
-                P.append(W[:, j] - M[:, j])
+                P.append(diff(W[:, j], M[:, j]))
             elif t == "d":
                 if b in ANGLES: continue
-                P.append(W[:, j] - D[:, j])
+                P.append(diff(W[:, j], D[:, j]))
             elif t == "mn":
                 P.append(M[:, j])
             elif t == "dn":
@@ -81,7 +109,7 @@ def phase_matrix(D, M, W, all_bodies, bodies, terms, angles_in_natal=False):
                 P.append(W[:, j])
             else:
                 raise ValueError(t)
-            lab.append(f"{t}_{b}")
+            lab.append(f"{t0}_{b}")
     return (np.column_stack(P) if P else np.zeros((len(D), 0))), lab
 
 

artamodel_iv.json

@@ -0,0 +1,303 @@
+{
+ "edition": "IV",
+ "n_train_rows": 181596,
+ "n_test_rows": 15262,
+ "n_test_pairs": 7631,
+ "n_full_chart_train_rows": 20632,
+ "plain": {
+  "held": 0.6113819044977837,
+  "age_cell_matched": 0.523580650656717
+ },
+ "artamodel": {
+  "held": 0.6251903718901581,
+  "held_raw": 0.6252101256366446,
+  "age_cell_matched": 0.565921651972106,
+  "inner": 0.632624037191737,
+  "stages": 8,
+  "phasors_used": [
+   "a_uranus",
+   "t1_neptune",
+   "t2_pluto",
+   "t2_neptune"
+  ],
+  "pair_asymmetry": 0.006218895572088684
+ },
+ "ensemble": {
+  "held": 0.6288835478263907,
+  "age_cell_matched": 0.5594745705436661
+ },
+ "terms": {
+  "a": "|\u03b81 \u2212 \u03b82|: the absolute synastry angle between the two natal charts (even under swap)",
+  "t1": "|\u03b8t \u2212 \u03b81|: the wedding sky to partner 1's natal longitude",
+  "t2": "|\u03b8t \u2212 \u03b82|: the wedding sky to partner 2's natal longitude",
+  "n1": "partner 1's own natal longitude",
+  "n2": "partner 2's own natal longitude",
+  "tn": "the wedding-day longitude itself"
+ },
+ "labels": [
+  "a_sun",
+  "a_moon",
+  "a_mercury",
+  "a_venus",
+  "a_mars",
+  "a_jupiter",
+  "a_saturn",
+  "a_uranus",
+  "a_neptune",
+  "a_pluto",
+  "a_true_node",
+  "a_true_south_node",
+  "a_chiron",
+  "a_mean_lilith",
+  "t1_sun",
+  "t1_moon",
+  "t1_mercury",
+  "t1_venus",
+  "t1_mars",
+  "t1_jupiter",
+  "t1_saturn",
+  "t1_uranus",
+  "t1_neptune",
+  "t1_pluto",
+  "t1_true_node",
+  "t1_true_south_node",
+  "t1_chiron",
+  "t1_mean_lilith",
+  "t2_sun",
+  "t2_moon",
+  "t2_mercury",
+  "t2_venus",
+  "t2_mars",
+  "t2_jupiter",
+  "t2_saturn",
+  "t2_uranus",
+  "t2_neptune",
+  "t2_pluto",
+  "t2_true_node",
+  "t2_true_south_node",
+  "t2_chiron",
+  "t2_mean_lilith",
+  "n1_sun",
+  "n1_moon",
+  "n1_mercury",
+  "n1_venus",
+  "n1_mars",
+  "n1_jupiter",
+  "n1_saturn",
+  "n1_uranus",
+  "n1_neptune",
+  "n1_pluto",
+  "n1_true_node",
+  "n1_true_south_node",
+  "n1_chiron",
+  "n1_mean_lilith",
+  "n2_sun",
+  "n2_moon",
+  "n2_mercury",
+  "n2_venus",
+  "n2_mars",
+  "n2_jupiter",
+  "n2_saturn",
+  "n2_uranus",
+  "n2_neptune",
+  "n2_pluto",
+  "n2_true_node",
+  "n2_true_south_node",
+  "n2_chiron",
+  "n2_mean_lilith",
+  "tn_sun",
+  "tn_moon",
+  "tn_mercury",
+  "tn_venus",
+  "tn_mars",
+  "tn_jupiter",
+  "tn_saturn",
+  "tn_uranus",
+  "tn_neptune",
+  "tn_pluto",
+  "tn_true_node",
+  "tn_true_south_node",
+  "tn_chiron",
+  "tn_mean_lilith"
+ ],
+ "leaderboard": {
+  "F0": -0.32983794552870865,
+  "stages": [
+   {
+    "stage": 1,
+    "phasor": 7,
+    "step": 0.4,
+    "alpha": -0.3025812663117348,
+    "c": 0.16726821685620108,
+    "w_re": -0.21057191284516086,
+    "w_im": -0.425876200333399,
+    "b_re": 0.6107863105989085,
+    "b_im": 0.5605056821561912
+   },
+   {
+    "stage": 2,
+    "phasor": 36,
+    "step": 0.4,
+    "alpha": 0.519453130834745,
+    "c": -0.4099180846986107,
+    "w_re": -0.22061897085028806,
+    "w_im": -0.24300062023722413,
+    "b_re": -0.7282549974830534,
+    "b_im": -0.3886675397665921
+   },
+   {
+    "stage": 3,
+    "phasor": 22,
+    "step": 0.4,
+    "alpha": 0.48893778356724893,
+    "c": -0.30557951847828757,
+    "w_re": 0.2760151368679704,
+    "w_im": 0.19036050251246772,
+    "b_re": 0.710970393549423,
+    "b_im": 0.1646654555251957
+   },
+   {
+    "stage": 4,
+    "phasor": 7,
+    "step": 0.4,
+    "alpha": 0.42082875033094663,
+    "c": -0.2170703928424133,
+    "w_re": -0.4347436393823073,
+    "w_im": -0.13803958644238642,
+    "b_re": -0.4352382943096409,
+    "b_im": 0.17857939138405746
+   },
+   {
+    "stage": 5,
+    "phasor": 22,
+    "step": 0.4,
+    "alpha": 0.36385820296400506,
+    "c": -0.2863029007149194,
+    "w_re": 0.37205652641101605,
+    "w_im": 0.0036864218033502306,
+    "b_re": 0.7643472890323912,
+    "b_im": -0.292744659830657
+   },
+   {
+    "stage": 6,
+    "phasor": 7,
+    "step": 0.4,
+    "alpha": 0.22575105498589784,
+    "c": -0.23018974010738902,
+    "w_re": -0.45212517836925703,
+    "w_im": 0.11202375136041444,
+    "b_re": -0.5403941985609879,
+    "b_im": 0.4898483443114326
+   },
+   {
+    "stage": 7,
+    "phasor": 37,
+    "step": 0.4,
+    "alpha": 0.6364948915292409,
+    "c": -0.40972401951564086,
+    "w_re": -0.3412319535146097,
+    "w_im": 0.2234795041673507,
+    "b_re": 0.12619931698751752,
+    "b_im": 0.7645164673177844
+   },
+   {
+    "stage": 8,
+    "phasor": 22,
+    "step": 0.4,
+    "alpha": 0.39102210850368074,
+    "c": -0.10963081383635781,
+    "w_re": -0.373384819083809,
+    "w_im": 0.14228082101266815,
+    "b_re": -0.3073528966491033,
+    "b_im": 0.28233056845889576
+   }
+  ],
+  "inner_auc": 0.632624037191737,
+  "labels": [
+   "a_sun",
+   "a_moon",
+   "a_mercury",
+   "a_venus",
+   "a_mars",
+   "a_jupiter",
+   "a_saturn",
+   "a_uranus",
+   "a_neptune",
+   "a_pluto",
+   "a_true_node",
+   "a_true_south_node",
+   "a_chiron",
+   "a_mean_lilith",
+   "t1_sun",
+   "t1_moon",
+   "t1_mercury",
+   "t1_venus",
+   "t1_mars",
+   "t1_jupiter",
+   "t1_saturn",
+   "t1_uranus",
+   "t1_neptune",
+   "t1_pluto",
+   "t1_true_node",
+   "t1_true_south_node",
+   "t1_chiron",
+   "t1_mean_lilith",
+   "t2_sun",
+   "t2_moon",
+   "t2_mercury",
+   "t2_venus",
+   "t2_mars",
+   "t2_jupiter",
+   "t2_saturn",
+   "t2_uranus",
+   "t2_neptune",
+   "t2_pluto",
+   "t2_true_node",
+   "t2_true_south_node",
+   "t2_chiron",
+   "t2_mean_lilith",
+   "n1_sun",
+   "n1_moon",
+   "n1_mercury",
+   "n1_venus",
+   "n1_mars",
+   "n1_jupiter",
+   "n1_saturn",
+   "n1_uranus",
+   "n1_neptune",
+   "n1_pluto",
+   "n1_true_node",
+   "n1_true_south_node",
+   "n1_chiron",
+   "n1_mean_lilith",
+   "n2_sun",
+   "n2_moon",
+   "n2_mercury",
+   "n2_venus",
+   "n2_mars",
+   "n2_jupiter",
+   "n2_saturn",
+   "n2_uranus",
+   "n2_neptune",
+   "n2_pluto",
+   "n2_true_node",
+   "n2_true_south_node",
+   "n2_chiron",
+   "n2_mean_lilith",
+   "tn_sun",
+   "tn_moon",
+   "tn_mercury",
+   "tn_venus",
+   "tn_mars",
+   "tn_jupiter",
+   "tn_saturn",
+   "tn_uranus",
+   "tn_neptune",
+   "tn_pluto",
+   "tn_true_node",
+   "tn_true_south_node",
+   "tn_chiron",
+   "tn_mean_lilith"
+  ]
+ }
+}

artamodel_iv.py

@@ -0,0 +1,131 @@
+"""
+artamodel_iv.py — ArtaModel, FOURTH EDITION: genderless and even.
+
+Operator 2026-08-19: "I want a genderless model from now on. so duplicate all the train and test data. (a, b, 1)
+should also mean (b, a, 1) and add any longterm relationship to the dataset (including gay marriages and business
+partnerships). also for each subtractive terms add abs to ensure each term is an even function. then start over the
+competition."
+
+    y = | b + Σᵢ aᵢ·e^{i|θ1ᵢ − θ2ᵢ|} + t1ᵢ·e^{i|θtᵢ − θ1ᵢ|} + t2ᵢ·e^{i|θtᵢ − θ2ᵢ|} + n1ᵢ·e^{iθ1ᵢ} + n2ᵢ·e^{iθ2ᵢ} + tnᵢ·e^{iθtᵢ} |²
+
+θ1, θ2 = the two partners' sidereal (Lahiri) longitudes at 09:00 local at the birthplace, in the order the row
+gives them (the files carry every pair in both orders, so the model is symmetric by the data); θt = the wedding
+sky at 12:00 UT; every difference is the wrapped absolute value in [0°, 180°], so each term is an even function of
+the swap; the presence rule is unchanged (a term exists only when both its phases exist).
+
+Produces, from AQ_PHASES (kerykeion_phases.py on the edition-IV files):
+  submission_plain_iv.csv      LightGBM on the two ages at the start, |gap|, start year     -- the bar
+  submission_artamodel_iv.csv  the 6-term boosted-over-split construction (greedy, train-only fit)
+  submission_ensemble_iv.csv   equal-weight rank average of the two
+  artamodel_iv.json            every number the pages need (held out, age-cell-matched, pair-symmetry check)
+  artamodel_iv_deployed.json   the same construction fitted on train + test, term by term
+Every submission is SYMMETRISED: the two rows of a pair (p<n>a, p<n>b) receive the mean of their two scores.
+Usage: AQ_PHASES=/tmp/aq4feat/phases.npz AQ_SOL=/tmp/aq4comp/solution.csv AQ_OUT=/tmp/aq4sub python artamodel_iv.py
+"""
+import json
+import os
+import sys
+import time
+
+import numpy as np
+import pandas as pd
+from scipy.stats import rankdata
+
+HERE = os.path.dirname(os.path.abspath(__file__)); sys.path.insert(0, HERE); sys.path.insert(0, os.path.join(HERE, "..", "coherent"))
+from artamodel import BODIES14, TERMS_IV, auc, phase_matrix                    # noqa: E402
+from artamodel_deploy import boost_recorded, score, BODY_TEXT                   # noqa: E402
+from artamodel_full_stack import matched                                        # noqa: E402
+
+PH = os.environ.get("AQ_PHASES", "/tmp/aq4feat/phases.npz"); SOL = os.environ.get("AQ_SOL", "/tmp/aq4comp/solution.csv")
+OUT = os.environ.get("AQ_OUT", "/tmp/aq4sub"); T0 = time.time()
+log = lambda *a: print(f"[{time.time()-T0:6.0f}s]", *a, flush=True)
+r01 = lambda v: (rankdata(v) - 1) / max(1.0, len(v) - 1)
+TERM_TEXT_IV = {"a": "|θ1 − θ2|: the absolute synastry angle between the two natal charts (even under swap)",
+                "t1": "|θt − θ1|: the wedding sky to partner 1's natal longitude", "t2": "|θt − θ2|: the wedding sky to partner 2's natal longitude",
+                "n1": "partner 1's own natal longitude", "n2": "partner 2's own natal longitude", "tn": "the wedding-day longitude itself"}
+
+
+def pair_key(ids):
+    return np.array([i[:-1] if (i[-1] in "ab" and i[:-1].lstrip("p").isdigit()) else i for i in ids])
+
+
+def symmetrise(ids, s):
+    """Mean of the two orders of a pair — the genderless prediction; the two rows then carry the same score."""
+    k = pair_key(ids); df = pd.DataFrame({"k": k, "s": s}); m = df.groupby("k")["s"].transform("mean").to_numpy()
+    return m
+
+
+def explain_iv(model, labels):
+    rows = {}
+    for st in model["stages"]:
+        lab = labels[st["phasor"]]; term, body = lab.split("_", 1)
+        w = complex(st["w_re"], st["w_im"]); b = complex(st["b_re"], st["b_im"])
+        r = rows.setdefault(lab, {"phasor": lab, "term": term, "body": BODY_TEXT.get(body, body), "meaning": f"{TERM_TEXT_IV[term]}, for {BODY_TEXT.get(body, body)}",
+                                   "stages": 0, "contribution": 0.0, "phase_deg": None})
+        r["stages"] += 1; r["contribution"] += abs(st["step"] * st["alpha"]) * abs(w) ** 2
+        r["phase_deg"] = float(np.degrees(np.angle(b) - np.angle(w)) % 360.0)
+    used = sorted(rows.values(), key=lambda r: -r["contribution"]); unused = [l for l in labels if l not in rows]
+    return used, unused
+
+
+def main():
+    import lightgbm as lgb
+    os.makedirs(OUT, exist_ok=True)
+    Z = np.load(PH, allow_pickle=True); s1, s2 = list(Z["slots"]); bodies = list(Z["bodies"]); ids = Z["id_test"]; y = Z["y_train"].astype(np.int64)
+    Atr, Btr, Wtr = Z[f"theta_{s1}_train"], Z[f"theta_{s2}_train"], Z["theta_wed_train"]; Ate, Bte, Wte = Z[f"theta_{s1}_test"], Z[f"theta_{s2}_test"], Z["theta_wed_test"]
+    ptr, pte, pn = Z["plain_train"], Z["plain_test"], list(Z["plain_names"]); later = Z["yr_train"].astype(int).max(1)
+    sol = pd.read_csv(SOL).set_index("id"); lab = [c for c in sol.columns if c != "Usage"][0]; yte = sol.loc[ids, lab].to_numpy().astype(int)
+    j1 = ptr[:, pn.index("start_is_jan1")] == 1.0; j1e = pte[:, pn.index("start_is_jan1")] == 1.0; Wtr = Wtr.copy(); Wte = Wte.copy(); Wtr[j1] = np.nan; Wte[j1e] = np.nan
+    log(f"train {len(y):,} rows · test {len(ids):,} rows ({len(set(pair_key(ids))):,} pairs) · slots {s1}/{s2}")
+    # ---- plain: two ages, |gap|, start year (even by construction of |gap| and by the doubled rows) ----
+    ia, ib, ig, iy = pn.index(f"age_{s1}_at_start"), pn.index(f"age_{s2}_at_start"), pn.index("age_gap"), pn.index("start_year")
+    X = np.column_stack([ptr[:, ia], ptr[:, ib], np.abs(ptr[:, ig]), ptr[:, iy]]); Xe = np.column_stack([pte[:, ia], pte[:, ib], np.abs(pte[:, ig]), pte[:, iy]])
+    prm = dict(n_estimators=400, learning_rate=0.03, num_leaves=15, min_child_samples=100, colsample_bytree=0.8, subsample=0.8, subsample_freq=1, reg_lambda=10.0, verbose=-1)
+    p_plain = np.zeros(len(Xe))
+    for sd in range(3):
+        c = lgb.LGBMClassifier(random_state=sd, **prm); c.fit(X, y); p_plain += c.predict_proba(Xe)[:, 1] / 3
+    p_plain_s = symmetrise(ids, p_plain)
+    ages = pte[:, [ia, ib]]; cell = (np.floor(np.nan_to_num(np.maximum(ages[:, 0], ages[:, 1])) / 3) * 1000 + np.floor(np.nan_to_num(np.minimum(ages[:, 0], ages[:, 1])) / 3)).astype(int)
+    log(f"  plain: held {auc(yte, p_plain):.4f} (symmetrised {auc(yte, p_plain_s):.4f})  age-cell-matched {matched(yte, p_plain_s, cell):.4f}")
+    # ---- ArtaModel IV: even phase matrix, 6 terms x 14 bodies ----
+    B = [bodies.index(b) for b in BODIES14]
+    P, labels = phase_matrix(Atr, Btr, Wtr, bodies, BODIES14, TERMS_IV, even=True); Pe, _ = phase_matrix(Ate, Bte, Wte, bodies, BODIES14, TERMS_IV, even=True)
+    charts = np.isfinite(Atr[:, B]).all(1) & np.isfinite(Btr[:, B]).all(1)
+    lat = later[charts]; inner = lat > np.quantile(lat, 0.85)
+    lb = boost_recorded(P[charts], y[charts], inner, stages=4 * len(labels), nu=0.1)
+    s_lb = score(lb, Pe); s_lb_s = symmetrise(ids, s_lb)
+    log(f"  ArtaModel IV (leaderboard fit, {int(charts.sum()):,} full-chart rows, {len(lb['stages'])} stages, inner {lb['inner_auc']:.4f}): "
+        f"held {auc(yte, s_lb):.4f} (symmetrised {auc(yte, s_lb_s):.4f})  age-cell-matched {matched(yte, s_lb_s, cell):.4f}")
+    # symmetry check: how far apart are the two orders' raw scores? (0 = perfectly even model)
+    k = pair_key(ids); df = pd.DataFrame({"k": k, "s": r01(s_lb)}); spread = df.groupby("k")["s"].agg(lambda v: abs(v.max() - v.min())).mean()
+    log(f"  raw pair asymmetry of the ArtaModel ranks (mean |rank_a − rank_b|): {spread:.4f}")
+    p_ens = 0.5 * r01(p_plain_s) + 0.5 * r01(s_lb_s)
+    log(f"  ENSEMBLE plain + ArtaModel IV (equal-weight rank average): held {auc(yte, p_ens):.4f}  age-cell-matched {matched(yte, p_ens, cell):.4f}")
+    for nm, v in (("plain_iv", p_plain_s), ("artamodel_iv", s_lb_s), ("ensemble_iv", p_ens)):
+        pd.DataFrame({"id": ids, lab: r01(v)}).to_csv(os.path.join(OUT, f"submission_{nm}.csv"), index=False)
+    used, unused = explain_iv(lb, labels)
+    # ---- deployed: train + test full-chart rows, the number of stages the leaderboard fit chose ----
+    charts_te = np.isfinite(Ate[:, B]).all(1) & np.isfinite(Bte[:, B]).all(1)
+    Pall = np.vstack([P[charts], Pe[charts_te]]); yall = np.concatenate([y[charts], yte[charts_te]])
+    dep = boost_recorded(Pall, yall, np.zeros(len(yall), bool), stages=len(lb["stages"]), nu=0.1)
+    dused, dunused = explain_iv(dep, labels)
+    dep.update({"edition": "IV — genderless, even", "labels": labels, "terms": list(TERMS_IV), "bodies": BODIES14, "fitted_on": "train + test rows with both natal charts (every pair in both orders)",
+                "n_rows": int(len(yall)), "leaderboard_estimate": {"inner_auc": lb["inner_auc"], "held_out_auc": float(auc(yte, s_lb_s)), "note": "the same construction fitted on train alone, symmetrised over the pair"},
+                "phase_convention": {"zodiac": "sidereal, Lahiri", "engine": "Kerykeion 5.12.9 (Swiss Ephemeris)", "birth_time": "09:00 local at the birthplace", "wedding_time": "12:00 UT",
+                                     "even": "every phase difference is the wrapped absolute value |Δθ| in [0°, 180°]", "order": "slot 1 / slot 2 carry no meaning; score both orders and average",
+                                     "presence_rule": "a term exists only when both of its phases exist; a missing phase contributes zero"},
+                "explanation": {"used": dused, "unused": dunused}})
+    json.dump(dep, open(os.path.join(OUT, "artamodel_iv_deployed.json"), "w"), indent=1)
+    R = {"edition": "IV", "n_train_rows": int(len(y)), "n_test_rows": int(len(ids)), "n_test_pairs": int(len(set(k))), "n_full_chart_train_rows": int(charts.sum()),
+         "plain": {"held": float(auc(yte, p_plain_s)), "age_cell_matched": float(matched(yte, p_plain_s, cell))},
+         "artamodel": {"held": float(auc(yte, s_lb_s)), "held_raw": float(auc(yte, s_lb)), "age_cell_matched": float(matched(yte, s_lb_s, cell)), "inner": lb["inner_auc"], "stages": len(lb["stages"]),
+                       "phasors_used": [u["phasor"] for u in used], "pair_asymmetry": float(spread)},
+         "ensemble": {"held": float(auc(yte, p_ens)), "age_cell_matched": float(matched(yte, p_ens, cell))},
+         "terms": {t: TERM_TEXT_IV[t] for t in TERMS_IV}, "labels": labels, "leaderboard": {**lb, "labels": labels}}
+    json.dump(R, open(os.path.join(OUT, "artamodel_iv.json"), "w"), indent=1)
+    log(f"  deployed fit: {len(yall):,} rows · {len(dep['stages'])} stages · phasors used: {', '.join(u['phasor'] for u in dused)}")
+    log(f"wrote {OUT}/submission_{{plain,artamodel,ensemble}}_iv.csv, artamodel_iv.json, artamodel_iv_deployed.json")
+
+
+if __name__ == "__main__":
+    main()

artamodel_iv_deployed.json

@@ -0,0 +1,335 @@
+{
+ "F0": -0.32088433382887727,
+ "stages": [
+  {
+   "stage": 1,
+   "phasor": 7,
+   "step": 0.4,
+   "alpha": -0.3376948209619891,
+   "c": 0.1874793060543171,
+   "w_re": -0.24321451372901814,
+   "w_im": -0.4426184203580808,
+   "b_re": 0.6339725314962403,
+   "b_im": 0.6026368179437613
+  },
+  {
+   "stage": 2,
+   "phasor": 7,
+   "step": 0.4,
+   "alpha": -0.44634092077748555,
+   "c": 0.1116560749961018,
+   "w_re": 0.46754000603237206,
+   "w_im": 0.2560313440563457,
+   "b_re": -0.5437361440779536,
+   "b_im": -0.15134405303585938
+  },
+  {
+   "stage": 3,
+   "phasor": 22,
+   "step": 0.4,
+   "alpha": 0.4772890407464341,
+   "c": -0.24435517906829834,
+   "w_re": 0.49816339813880034,
+   "w_im": -0.06628727704385114,
+   "b_re": 0.4920847669878105,
+   "b_im": -0.4043385699859723
+  },
+  {
+   "stage": 4,
+   "phasor": 36,
+   "step": 0.4,
+   "alpha": 0.4121201205056047,
+   "c": -0.22911907658681685,
+   "w_re": -0.4990958449581369,
+   "w_im": 0.04409056714697404,
+   "b_re": -0.552262102086733,
+   "b_im": 0.38921839858489915
+  },
+  {
+   "stage": 5,
+   "phasor": 7,
+   "step": 0.4,
+   "alpha": 0.2891143341130896,
+   "c": -0.23242638188558676,
+   "w_re": 0.29176287067569107,
+   "w_im": 0.4353249731528996,
+   "b_re": 0.5499183154598283,
+   "b_im": 0.17856648674940595
+  },
+  {
+   "stage": 6,
+   "phasor": 36,
+   "step": 0.4,
+   "alpha": -0.30350135615690277,
+   "c": 0.1719921813087553,
+   "w_re": 0.2898425417803235,
+   "w_im": -0.6250770077916671,
+   "b_re": -0.5147192423447787,
+   "b_im": 0.6150611274518173
+  },
+  {
+   "stage": 7,
+   "phasor": 22,
+   "step": 0.4,
+   "alpha": 0.35284123130265743,
+   "c": -0.20570391043323596,
+   "w_re": 0.08529216192335105,
+   "w_im": 0.4735197468892639,
+   "b_re": 0.44026506888173583,
+   "b_im": 0.5399124753716616
+  },
+  {
+   "stage": 8,
+   "phasor": 7,
+   "step": 0.4,
+   "alpha": -0.41087990743918523,
+   "c": 0.07256860068983159,
+   "w_re": 0.3823720953371966,
+   "w_im": -0.2417660418214181,
+   "b_re": -0.44376920363750716,
+   "b_im": 0.32639335063421915
+  }
+ ],
+ "inner_auc": null,
+ "edition": "IV \u2014 genderless, even",
+ "labels": [
+  "a_sun",
+  "a_moon",
+  "a_mercury",
+  "a_venus",
+  "a_mars",
+  "a_jupiter",
+  "a_saturn",
+  "a_uranus",
+  "a_neptune",
+  "a_pluto",
+  "a_true_node",
+  "a_true_south_node",
+  "a_chiron",
+  "a_mean_lilith",
+  "t1_sun",
+  "t1_moon",
+  "t1_mercury",
+  "t1_venus",
+  "t1_mars",
+  "t1_jupiter",
+  "t1_saturn",
+  "t1_uranus",
+  "t1_neptune",
+  "t1_pluto",
+  "t1_true_node",
+  "t1_true_south_node",
+  "t1_chiron",
+  "t1_mean_lilith",
+  "t2_sun",
+  "t2_moon",
+  "t2_mercury",
+  "t2_venus",
+  "t2_mars",
+  "t2_jupiter",
+  "t2_saturn",
+  "t2_uranus",
+  "t2_neptune",
+  "t2_pluto",
+  "t2_true_node",
+  "t2_true_south_node",
+  "t2_chiron",
+  "t2_mean_lilith",
+  "n1_sun",
+  "n1_moon",
+  "n1_mercury",
+  "n1_venus",
+  "n1_mars",
+  "n1_jupiter",
+  "n1_saturn",
+  "n1_uranus",
+  "n1_neptune",
+  "n1_pluto",
+  "n1_true_node",
+  "n1_true_south_node",
+  "n1_chiron",
+  "n1_mean_lilith",
+  "n2_sun",
+  "n2_moon",
+  "n2_mercury",
+  "n2_venus",
+  "n2_mars",
+  "n2_jupiter",
+  "n2_saturn",
+  "n2_uranus",
+  "n2_neptune",
+  "n2_pluto",
+  "n2_true_node",
+  "n2_true_south_node",
+  "n2_chiron",
+  "n2_mean_lilith",
+  "tn_sun",
+  "tn_moon",
+  "tn_mercury",
+  "tn_venus",
+  "tn_mars",
+  "tn_jupiter",
+  "tn_saturn",
+  "tn_uranus",
+  "tn_neptune",
+  "tn_pluto",
+  "tn_true_node",
+  "tn_true_south_node",
+  "tn_chiron",
+  "tn_mean_lilith"
+ ],
+ "terms": [
+  "a",
+  "t1",
+  "t2",
+  "n1",
+  "n2",
+  "tn"
+ ],
+ "bodies": [
+  "sun",
+  "moon",
+  "mercury",
+  "venus",
+  "mars",
+  "jupiter",
+  "saturn",
+  "uranus",
+  "neptune",
+  "pluto",
+  "true_node",
+  "true_south_node",
+  "chiron",
+  "mean_lilith"
+ ],
+ "fitted_on": "train + test rows with both natal charts (every pair in both orders)",
+ "n_rows": 35894,
+ "leaderboard_estimate": {
+  "inner_auc": 0.632624037191737,
+  "held_out_auc": 0.6251903718901581,
+  "note": "the same construction fitted on train alone, symmetrised over the pair"
+ },
+ "phase_convention": {
+  "zodiac": "sidereal, Lahiri",
+  "engine": "Kerykeion 5.12.9 (Swiss Ephemeris)",
+  "birth_time": "09:00 local at the birthplace",
+  "wedding_time": "12:00 UT",
+  "even": "every phase difference is the wrapped absolute value |\u0394\u03b8| in [0\u00b0, 180\u00b0]",
+  "order": "slot 1 / slot 2 carry no meaning; score both orders and average",
+  "presence_rule": "a term exists only when both of its phases exist; a missing phase contributes zero"
+ },
+ "explanation": {
+  "used": [
+   {
+    "phasor": "a_uranus",
+    "term": "a",
+    "body": "Uranus",
+    "meaning": "|\u03b81 \u2212 \u03b82|: the absolute synastry angle between the two natal charts (even under swap), for Uranus",
+    "stages": 4,
+    "contribution": 0.15058023776611307,
+    "phase_deg": 175.96975487087076
+   },
+   {
+    "phasor": "t2_neptune",
+    "term": "t2",
+    "body": "Neptune",
+    "meaning": "|\u03b8t \u2212 \u03b82|: the wedding sky to partner 2's natal longitude, for Neptune",
+    "stages": 2,
+    "contribution": 0.09901603485407229,
+    "phase_deg": 195.04784294272702
+   },
+   {
+    "phasor": "t1_neptune",
+    "term": "t1",
+    "body": "Neptune",
+    "meaning": "|\u03b8t \u2212 \u03b81|: the wedding sky to partner 1's natal longitude, for Neptune",
+    "stages": 2,
+    "contribution": 0.08089028752347999,
+    "phase_deg": 331.01573734961244
+   }
+  ],
+  "unused": [
+   "a_sun",
+   "a_moon",
+   "a_mercury",
+   "a_venus",
+   "a_mars",
+   "a_jupiter",
+   "a_saturn",
+   "a_neptune",
+   "a_pluto",
+   "a_true_node",
+   "a_true_south_node",
+   "a_chiron",
+   "a_mean_lilith",
+   "t1_sun",
+   "t1_moon",
+   "t1_mercury",
+   "t1_venus",
+   "t1_mars",
+   "t1_jupiter",
+   "t1_saturn",
+   "t1_uranus",
+   "t1_pluto",
+   "t1_true_node",
+   "t1_true_south_node",
+   "t1_chiron",
+   "t1_mean_lilith",
+   "t2_sun",
+   "t2_moon",
+   "t2_mercury",
+   "t2_venus",
+   "t2_mars",
+   "t2_jupiter",
+   "t2_saturn",
+   "t2_uranus",
+   "t2_pluto",
+   "t2_true_node",
+   "t2_true_south_node",
+   "t2_chiron",
+   "t2_mean_lilith",
+   "n1_sun",
+   "n1_moon",
+   "n1_mercury",
+   "n1_venus",
+   "n1_mars",
+   "n1_jupiter",
+   "n1_saturn",
+   "n1_uranus",
+   "n1_neptune",
+   "n1_pluto",
+   "n1_true_node",
+   "n1_true_south_node",
+   "n1_chiron",
+   "n1_mean_lilith",
+   "n2_sun",
+   "n2_moon",
+   "n2_mercury",
+   "n2_venus",
+   "n2_mars",
+   "n2_jupiter",
+   "n2_saturn",
+   "n2_uranus",
+   "n2_neptune",
+   "n2_pluto",
+   "n2_true_node",
+   "n2_true_south_node",
+   "n2_chiron",
+   "n2_mean_lilith",
+   "tn_sun",
+   "tn_moon",
+   "tn_mercury",
+   "tn_venus",
+   "tn_mars",
+   "tn_jupiter",
+   "tn_saturn",
+   "tn_uranus",
+   "tn_neptune",
+   "tn_pluto",
+   "tn_true_node",
+   "tn_true_south_node",
+   "tn_chiron",
+   "tn_mean_lilith"
+  ]
+ }
+}

artamodel_score_iv.py

@@ -0,0 +1,118 @@
+"""
+artamodel_score_iv.py — score a pair with the deployed ArtaModel IV (genderless, even), from three dates and two
+places. numpy only for the model; Kerykeion for the phases (pip install kerykeion timezonefinder).
+
+    from artamodel_score_iv import predict
+    p = predict("1936-08-04", 37.943, 23.647, "1924-05-14", 37.727, 26.909, "1968-06-15")   # partner 1, partner 2, start
+
+The two partners carry no order: the model is scored in BOTH orders and the two logits are averaged, so the
+answer is exactly the same whichever way round you give them (every phase difference already enters as |Δθ|).
+Returns the probability that the relationship lasted thirty years and the term-by-term account of both orders.
+See README.md / ARTAMODEL.md for what the model is and, honestly, what it reads.
+"""
+import json
+import os
+import warnings
+
+import numpy as np
+
+warnings.filterwarnings("ignore")
+HERE = os.path.dirname(os.path.abspath(__file__))
+MODEL = json.load(open(os.path.join(HERE, "artamodel_iv_deployed.json")))
+BODIES = MODEL["bodies"]; TERMS = MODEL["terms"]; LABELS = MODEL["labels"]
+ANGLES = {"ascendant", "medium_coeli"}
+SLOW = {"jupiter", "saturn", "uranus", "neptune", "pluto", "true_node", "true_south_node", "chiron", "mean_lilith"}
+_TF = None
+
+
+def _tz(lat, lon):
+    global _TF
+    if _TF is None:
+        from timezonefinder import TimezoneFinder
+        _TF = TimezoneFinder()
+    return _TF.timezone_at(lng=lon, lat=lat) or "UTC"
+
+
+def _prec(d):
+    if not d or d == "0000-00-00":
+        return 0
+    return 1 if d.endswith("-00-00") else (2 if d.endswith("-00") else 3)
+
+
+def theta(date, lat=None, lon=None, hour=9, natal=True):
+    """Sidereal (Lahiri) longitudes of the model's bodies at 09:00 local (natal) or 12:00 UT (wedding); NaN where
+    the date's precision cannot place a body (year-only -> slow bodies only)."""
+    from kerykeion import AstrologicalSubject
+    out = np.full(len(BODIES), np.nan); p = _prec(date)
+    if p == 0 or (natal and (lat is None or lon is None)):
+        return out
+    y, m, d = int(date[:4]), max(1, int(date[5:7])), max(1, int(date[8:10]))
+    if natal:
+        s = AstrologicalSubject("x", y, m, d, hour, 0, lng=float(lon), lat=float(lat), tz_str=_tz(float(lat), float(lon)),
+                                city="x", nation="XX", zodiac_type="Sidereal", sidereal_mode="LAHIRI", online=False)
+    else:
+        s = AstrologicalSubject("w", y, m, d, 12, 0, lng=0.0, lat=51.48, tz_str="UTC", city="G", nation="GB",
+                                zodiac_type="Sidereal", sidereal_mode="LAHIRI", online=False)
+    for j, b in enumerate(BODIES):
+        if p == 1 and b not in SLOW:
+            continue
+        if p == 2 and b not in SLOW and b != "sun":
+            continue
+        try:
+            out[j] = float(getattr(s, b).abs_pos)
+        except Exception:
+            pass
+    return out
+
+
+def absdiff(x, y):
+    """The wrapped absolute phase difference in [0, 180] degrees — even in (x, y)."""
+    return np.abs((x - y + 180.0) % 360.0 - 180.0)
+
+
+def phases(t1, t2, tw):
+    """The 84 phases in the model's label order (NaN = the term does not exist). Every difference is |Δθ|."""
+    P = np.full(len(LABELS), np.nan)
+    col = {b: j for j, b in enumerate(BODIES)}
+    for k, lab in enumerate(LABELS):
+        t, b = lab.split("_", 1); j = col[b]
+        if t == "a": P[k] = absdiff(t1[j], t2[j])
+        elif t == "t1": P[k] = absdiff(tw[j], t1[j])
+        elif t == "t2": P[k] = absdiff(tw[j], t2[j])
+        elif t == "n1": P[k] = t1[j]
+        elif t == "n2": P[k] = t2[j]
+        elif t == "tn": P[k] = tw[j]
+    return P
+
+
+def logit(P):
+    """The deployed model's logit for one phase vector, with the per-stage account."""
+    rad = np.pi / 180.0
+    C, S = np.nan_to_num(np.cos(P * rad)), np.nan_to_num(np.sin(P * rad))
+    F = MODEL["F0"]; account = []
+    for st in MODEL["stages"]:
+        j = st["phasor"]
+        if not np.isfinite(P[j]):
+            account.append({"stage": st["stage"], "phasor": LABELS[j], "contribution": 0.0, "note": "term absent for this couple"}); continue
+        Zr = st["w_re"] * C[j] - st["w_im"] * S[j] + st["b_re"]; Zi = st["w_re"] * S[j] + st["w_im"] * C[j] + st["b_im"]
+        u = Zr * Zr + Zi * Zi; contrib = st["step"] * (st["alpha"] * u + st["c"])
+        F += contrib; account.append({"stage": st["stage"], "phasor": LABELS[j], "phase_deg": float(P[j] % 360), "contribution": float(contrib)})
+    return float(F), account
+
+
+def predict(dob_1, lat_1, lon_1, dob_2, lat_2, lon_2, start):
+    wed = start if start[5:] != "01-01" else start[:4] + "-00-00"          # a 1 January start is a year-only record
+    t1, t2, tw = theta(dob_1, lat_1, lon_1), theta(dob_2, lat_2, lon_2), theta(wed, natal=False)
+    F12, acc12 = logit(phases(t1, t2, tw)); F21, acc21 = logit(phases(t2, t1, tw))
+    F = 0.5 * (F12 + F21)
+    return {"probability": float(1 / (1 + np.exp(-F))), "logit": F, "logit_order_12": F12, "logit_order_21": F21,
+            "terms": acc12, "terms_swapped": acc21}
+
+
+if __name__ == "__main__":
+    r = predict("1936-08-04", 37.943, 23.647, "1924-05-14", 37.727, 26.909, "1968-06-15")
+    r2 = predict("1924-05-14", 37.727, 26.909, "1936-08-04", 37.943, 23.647, "1968-06-15")
+    assert abs(r["probability"] - r2["probability"]) < 1e-12, "the scorer is not symmetric"
+    print(f"  p(lasted 30 years) = {r['probability']:.3f}   logit {r['logit']:+.3f}   (swapped: {r2['probability']:.3f} — identical)")
+    for t in r["terms"][:8]:
+        print(f"    stage {t['stage']:>2}  {t['phasor']:<12} " + (f"φ={t['phase_deg']:6.1f}°  {t['contribution']:+.4f}" if "phase_deg" in t else t["note"]))

kerykeion_phases.py

@@ -100,8 +100,16 @@ def _work(args):
     return i, theta(dd, latd, lond, 9, True), theta(dm, latm, lonm, 9, True), theta(wed, None, None, 12, False)
 
 
+def slots(df):
+    """FOURTH EDITION (genderless, 2026-08-19): the files carry `dob_a/dob_b`; the earlier editions `dob_dad/dob_mom`.
+    The extractor reads whichever it is given and names its outputs by the same slot names."""
+    return ("a", "b") if "dob_a" in df.columns else ("dad", "mom")
+
+
 def build(df):
-    jobs = [(i, r.dob_dad, r.lat_dad, r.lon_dad, r.dob_mom, r.lat_mom, r.lon_mom, r.start)
+    s1, s2 = slots(df)
+    jobs = [(i, getattr(r, f"dob_{s1}"), getattr(r, f"lat_{s1}"), getattr(r, f"lon_{s1}"), getattr(r, f"dob_{s2}"),
+             getattr(r, f"lat_{s2}"), getattr(r, f"lon_{s2}"), r.start)
             for i, r in enumerate(df.itertuples(index=False))]
     with mp.Pool(max(1, mp.cpu_count() - 1)) as pool:
         res = pool.map(_work, jobs, chunksize=256)
@@ -112,27 +120,31 @@ def build(df):
 
 
 def main():
-    tr = pd.read_csv(f"{SRC}/train.csv", dtype={"dob_dad": str, "dob_mom": str, "start": str})
-    te = pd.read_csv(f"{SRC}/test.csv", dtype={"dob_dad": str, "dob_mom": str, "start": str})
-    LABEL = [c for c in tr.columns if c not in {"id", "dob_dad", "dob_mom", "lat_dad", "lon_dad", "lat_mom", "lon_mom", "start"}][0]
+    tr = pd.read_csv(f"{SRC}/train.csv", dtype=str); te = pd.read_csv(f"{SRC}/test.csv", dtype=str)
+    s1, s2 = slots(tr)
+    for df in (tr, te):
+        for c in (f"lat_{s1}", f"lon_{s1}", f"lat_{s2}", f"lon_{s2}"):
+            df[c] = pd.to_numeric(df[c], errors="coerce")
+    LABEL = [c for c in tr.columns if c not in {"id", f"dob_{s1}", f"dob_{s2}", f"lat_{s1}", f"lon_{s1}", f"lat_{s2}", f"lon_{s2}", "start"}][0]
+    d1, d2 = tr[f"dob_{s1}"], tr[f"dob_{s2}"]
     if LIMIT:
         tr, te = tr.head(LIMIT), te.head(max(200, LIMIT // 4)); log(f"AQ_LIMIT={LIMIT}: DRY RUN")
     log(f"train {len(tr):,} · test {len(te):,}")
     Dtr, Mtr, Wtr = build(tr); log("train phases")
     Dte, Mte, Wte = build(te); log("test phases")
     def plain(df):
-        yd = pd.to_numeric(df.dob_dad.str[:4], errors="coerce").where(df.dob_dad != "0000-00-00")
-        ym = pd.to_numeric(df.dob_mom.str[:4], errors="coerce").where(df.dob_mom != "0000-00-00")
+        yd = pd.to_numeric(df[f"dob_{s1}"].str[:4], errors="coerce").where(df[f"dob_{s1}"] != "0000-00-00")
+        ym = pd.to_numeric(df[f"dob_{s2}"].str[:4], errors="coerce").where(df[f"dob_{s2}"] != "0000-00-00")
         sy = df.start.str[:4].astype(float)
         return np.column_stack([sy - yd, sy - ym, ym - yd, sy, (df.start.str[5:] == "01-01").astype(float)])
     os.makedirs(OUT, exist_ok=True)
-    np.savez_compressed(f"{OUT}/phases.npz", theta_dad_train=Dtr, theta_mom_train=Mtr, theta_wed_train=Wtr,
-                        theta_dad_test=Dte, theta_mom_test=Mte, theta_wed_test=Wte, bodies=np.array(BODIES, dtype=object),
+    np.savez_compressed(f"{OUT}/phases.npz", **{f"theta_{s1}_train": Dtr, f"theta_{s2}_train": Mtr, "theta_wed_train": Wtr,
+                        f"theta_{s1}_test": Dte, f"theta_{s2}_test": Mte, "theta_wed_test": Wte}, bodies=np.array(BODIES, dtype=object), slots=np.array([s1, s2], dtype=object),
                         y_train=tr[LABEL].to_numpy().astype(np.int8), id_test=te.id.to_numpy() if "id" in te else np.arange(len(te)),
                         plain_train=plain(tr), plain_test=plain(te),
-                        plain_names=np.array(["age_dad_at_start", "age_mom_at_start", "age_gap", "start_year", "start_is_jan1"], dtype=object),
-                        yr_train=np.column_stack([pd.to_numeric(tr.dob_dad.str[:4], errors="coerce").fillna(0),
-                                                  pd.to_numeric(tr.dob_mom.str[:4], errors="coerce").fillna(0)]).astype(np.int16))
+                        plain_names=np.array([f"age_{s1}_at_start", f"age_{s2}_at_start", "age_gap", "start_year", "start_is_jan1"], dtype=object),
+                        yr_train=np.column_stack([pd.to_numeric(d1.str[:4], errors="coerce").fillna(0),
+                                                  pd.to_numeric(d2.str[:4], errors="coerce").fillna(0)]).astype(np.int16))
     full = np.isfinite(Dtr).all(1) & np.isfinite(Mtr).all(1)
     log(f"wrote {OUT}/phases.npz · {len(BODIES)} bodies · train rows with BOTH natal charts complete: {full.sum():,} · "
         f"wedding sky complete: {np.isfinite(Wtr[:, :10]).all(1).sum():,}")