Merge feature/import-polish: decision-matching (name,zucht,dob), specific-wins dedup, parent-FK-backfill (DB-weit via allDbNormToGid), Skarlett-Parse-Fix [god-QA validated]

GerbilManager.Tests/ImportServiceTests.cs

@@ -321,6 +321,192 @@ namespace GerbilManager.Tests
             Assert.Equal("aa Ccchm ?? eef ?? ?? ?? ??", ImportService.ComposeGenotype(g));
         }
 
+        [Fact]
+        public async Task ParentFkBackfill_fills_null_litter_parent_on_reimport()
+        {
+            // Run 1: litter "Wurf A" has sire "Vater" (conflict=true — not loaded) and dam "Mutter"
+            //        (conflict=false — loaded). After run 1: litter.FatherId = null.
+            // Run 2: sire "Vater" now conflict=false → loaded as NEW in run 2. Backfill via
+            //        createdAnimalByName sets FatherId. (god steering point 3: run-2 path.)
+            var dir = Path.Combine(Path.GetTempPath(), "backfill-" + Guid.NewGuid().ToString("N"));
+            Directory.CreateDirectory(dir);
+            using var conn = new SqliteConnection("DataSource=:memory:");
+            conn.Open();
+            try
+            {
+                var littersJson = """
+                [{"id":"L-A","litterId":"A","date":"01.05.2023","damName":"Mutter [ZdkC]","sireName":"Vater [ZdkC]","totalBorn":3,"zuchtnummer":"","note":""}]
+                """;
+                // Run 1: Vater is in conflict -> not loaded
+                var animals1 = """
+                [
+                  {"id":"mutter","name":"Mutter [ZdkC]","dob":"01.01.2021","death":"","farbschlag":"","gender":"female","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false},
+                  {"id":"vater","name":"Vater [ZdkC]","dob":"02.02.2021","death":"","farbschlag":"","gender":"male","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":true},
+                  {"id":"kind","name":"Kind [ZdkC]","dob":"01.05.2023","death":"","farbschlag":"","gender":null,"zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false,
+                   "litterRef":{"litterId":"L-A","method":"geburtsdatum+eltern","confidence":"hoch"}}
+                ]
+                """;
+                File.WriteAllText(Path.Combine(dir, "litters.json"), littersJson);
+                File.WriteAllText(Path.Combine(dir, "animals.json"), animals1);
+
+                var opts = new DbContextOptionsBuilder<ApplicationContext>().UseSqlite(conn).Options;
+                using var db = new ApplicationContext(opts);
+                await db.Database.EnsureCreatedAsync();
+
+                var report1 = await new ImportService(db, dir, dir).RunAsync(execute: true);
+                Assert.Equal(0, report1.Litters.ParentFksBackfilled);
+                var litter1 = await db.Litters.SingleAsync(l => l.Name == "Wurf A");
+                Assert.Null(litter1.FatherId);              // Vater was quarantined -> null FK
+                Assert.NotNull(litter1.MotherId);           // Mutter was loaded -> set
+
+                // Run 2: Vater now conflict=false -> loaded as NEW animal in this run
+                var animals2 = """
+                [
+                  {"id":"mutter","name":"Mutter [ZdkC]","dob":"01.01.2021","death":"","farbschlag":"","gender":"female","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false},
+                  {"id":"vater","name":"Vater [ZdkC]","dob":"02.02.2021","death":"","farbschlag":"","gender":"male","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false},
+                  {"id":"kind","name":"Kind [ZdkC]","dob":"01.05.2023","death":"","farbschlag":"","gender":null,"zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false,
+                   "litterRef":{"litterId":"L-A","method":"geburtsdatum+eltern","confidence":"hoch"}}
+                ]
+                """;
+                File.WriteAllText(Path.Combine(dir, "animals.json"), animals2);
+
+                var report2 = await new ImportService(db, dir, dir).RunAsync(execute: true);
+                Assert.Equal(1, report2.Litters.ParentFksBackfilled);    // backfill happened
+                var vater = await db.Gerbils.SingleAsync(g => g.ExternalRef == "vater");
+                var litter2 = await db.Litters.SingleAsync(l => l.Name == "Wurf A");
+                Assert.Equal(vater.Id, litter2.FatherId);                // FK now set
+            }
+            finally
+            {
+                try { Directory.Delete(dir, recursive: true); } catch { }
+            }
+        }
+
+        [Fact]
+        public async Task ParentFkBackfill_uses_allDb_lookup_when_parent_not_in_current_loadable()
+        {
+            // god steering point 3: the main case — parent was loaded in a PREVIOUS run (not in
+            // the current run's animals.json at all). Backfill must find them via allDbNormToGid.
+            //
+            // Run 1: litter "Wurf C" + dam loaded, sire quarantined -> FatherId null.
+            // Run 2: sire loaded (new animal).
+            // Run 3: animals.json has ONLY the kind (sire absent from extract). Sire is in DB
+            //        from run 2 but NOT in the current run's loadable/createdAnimalByName.
+            //        Backfill must use allDbNormToGid to find him.
+            var dir = Path.Combine(Path.GetTempPath(), "backfill-db-" + Guid.NewGuid().ToString("N"));
+            Directory.CreateDirectory(dir);
+            using var conn = new SqliteConnection("DataSource=:memory:");
+            conn.Open();
+            try
+            {
+                var littersJson = """
+                [{"id":"L-C","litterId":"C","date":"10.06.2023","damName":"Dame [ZdkC]","sireName":"Herr [ZdkC]","totalBorn":2,"zuchtnummer":"","note":""}]
+                """;
+                // Run 1: sire quarantined
+                File.WriteAllText(Path.Combine(dir, "litters.json"), littersJson);
+                File.WriteAllText(Path.Combine(dir, "animals.json"), """
+                [
+                  {"id":"dame","name":"Dame [ZdkC]","dob":"05.05.2021","death":"","farbschlag":"","gender":"female","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false},
+                  {"id":"herr","name":"Herr [ZdkC]","dob":"06.06.2021","death":"","farbschlag":"","gender":"male","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":true}
+                ]
+                """);
+                var opts = new DbContextOptionsBuilder<ApplicationContext>().UseSqlite(conn).Options;
+                using var db = new ApplicationContext(opts);
+                await db.Database.EnsureCreatedAsync();
+                await new ImportService(db, dir, dir).RunAsync(execute: true);
+                Assert.Null((await db.Litters.SingleAsync(l => l.Name == "Wurf C")).FatherId);
+
+                // Run 2: sire now loaded
+                File.WriteAllText(Path.Combine(dir, "animals.json"), """
+                [
+                  {"id":"dame","name":"Dame [ZdkC]","dob":"05.05.2021","death":"","farbschlag":"","gender":"female","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false},
+                  {"id":"herr","name":"Herr [ZdkC]","dob":"06.06.2021","death":"","farbschlag":"","gender":"male","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false}
+                ]
+                """);
+                await new ImportService(db, dir, dir).RunAsync(execute: true);
+                var herrId = (await db.Gerbils.SingleAsync(g => g.ExternalRef == "herr")).Id;
+                // Run 2 itself may or may not backfill (depends on name normalization alignment).
+                // For the test we care about run 3.
+
+                // Run 3: sire NOT in animals.json at all (absent from new extract).
+                // litter still has FatherId=null if run 2 didn't backfill; if it did, we simulate
+                // by manually resetting FatherId to null so run 3 must fix it.
+                var litter3 = await db.Litters.SingleAsync(l => l.Name == "Wurf C");
+                litter3.FatherId = null;
+                await db.SaveChangesAsync();
+
+                File.WriteAllText(Path.Combine(dir, "animals.json"), """
+                [
+                  {"id":"dame","name":"Dame [ZdkC]","dob":"05.05.2021","death":"","farbschlag":"","gender":"female","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false}
+                ]
+                """);
+                // Run 3: sire absent from loadable (NOT in createdAnimalByName), but IS in DB.
+                var report3 = await new ImportService(db, dir, dir).RunAsync(execute: true);
+                Assert.Equal(1, report3.Litters.ParentFksBackfilled);   // allDbNormToGid path
+                var litter3After = await db.Litters.SingleAsync(l => l.Name == "Wurf C");
+                Assert.Equal(herrId, litter3After.FatherId);             // FK set from DB lookup
+            }
+            finally
+            {
+                try { Directory.Delete(dir, recursive: true); } catch { }
+            }
+        }
+
+        [Fact]
+        public async Task ParentFkBackfill_dry_run_counts_without_writing()
+        {
+            // Dry-run on a DB with an existing null-parent litter should predict the backfill count.
+            var dir = Path.Combine(Path.GetTempPath(), "backfill-dr-" + Guid.NewGuid().ToString("N"));
+            Directory.CreateDirectory(dir);
+            using var conn = new SqliteConnection("DataSource=:memory:");
+            conn.Open();
+            try
+            {
+                var littersJson = """
+                [{"id":"L-B","litterId":"B","date":"15.06.2023","damName":"Mami [ZdkC]","sireName":"Papi [ZdkC]","totalBorn":2,"zuchtnummer":"","note":""}]
+                """;
+                var animals1 = """
+                [
+                  {"id":"mami","name":"Mami [ZdkC]","dob":"03.03.2021","death":"","farbschlag":"","gender":"female","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":false},
+                  {"id":"papi","name":"Papi [ZdkC]","dob":"04.04.2021","death":"","farbschlag":"","gender":"male","zuchtCanon":"kleinechaote",
+                   "genotype":{"mapped8locus":{},"rawGenotype":"","unmappedTokens":[]},"conflict":true}
+                ]
+                """;
+                File.WriteAllText(Path.Combine(dir, "litters.json"), littersJson);
+                File.WriteAllText(Path.Combine(dir, "animals.json"), animals1);
+
+                var opts = new DbContextOptionsBuilder<ApplicationContext>().UseSqlite(conn).Options;
+                using var db = new ApplicationContext(opts);
+                await db.Database.EnsureCreatedAsync();
+                await new ImportService(db, dir, dir).RunAsync(execute: true);   // run 1
+
+                // Run 2 dry-run with papi un-quarantined
+                var animals2 = animals1.Replace("\"conflict\":true", "\"conflict\":false");
+                File.WriteAllText(Path.Combine(dir, "animals.json"), animals2);
+                var dry = await new ImportService(db, dir, dir).RunAsync(execute: false);
+
+                Assert.Equal(1, dry.Litters.ParentFksBackfilled);     // predicted but not written
+                var litter = await db.Litters.SingleAsync(l => l.Name == "Wurf B");
+                Assert.Null(litter.FatherId);                          // not written in dry-run
+            }
+            finally
+            {
+                try { Directory.Delete(dir, recursive: true); } catch { }
+            }
+        }
+
         [Theory]
         [InlineData("01.02.2020", 2020, 2, 1)]
         [InlineData("5.3.21", 2021, 3, 5)]

GerbilManagerWebAPI/Import/ImportModels.cs

@@ -86,7 +86,8 @@ namespace GerbilManagerWebAPI.Import
     public sealed record ResidencySummary(int Resident, int External, int FlippedByParentRule);
 
     public sealed record LitterSummary(int InSource, int Created, int AlreadyImported,
-        int DerivedFromChart = 0, int DerivedSkipped = 0, int ParentFksDropped = 0);
+        int DerivedFromChart = 0, int DerivedSkipped = 0, int ParentFksDropped = 0,
+        int ParentFksBackfilled = 0);
 
     public sealed record AnimalSummary(
         int InSource,

GerbilManagerWebAPI/Import/ImportService.cs

@@ -399,10 +399,67 @@ namespace GerbilManagerWebAPI.Import
                 await _db.SaveChangesAsync();
             }
 
+            // PARENT-FK BACKFILL (idempotent re-run): already-imported Wurfchronik litters that
+            // have null Father/MotherId because the parent was previously quarantined may now be
+            // resolvable. Two lookup sources — must check BOTH:
+            //   (a) createdAnimalByName: animals loaded/re-linked in THIS run (new or existing).
+            //   (b) allDbNormToGid: ALL gerbils already in the DB, for parents loaded in an
+            //       EARLIER run who are no longer in the current extract (e.g. alreadyImported
+            //       animals absent from this run's animals.json, or name normalization mismatch
+            //       between animals.json and the Wurfchronik sire/dam field).
+            // Counted for dry-run too; writes only when execute=true.
+            int parentFksBackfilled = 0;
+            {
+                // Build DB-wide normalized-name lookup (supplementary to createdAnimalByName).
+                var allDbNormToGid = existingRows
+                    .GroupBy(g => Normalize(StripZucht(g.Name)))
+                    .ToDictionary(grp => grp.Key, grp => grp.First().Id);
+
+                var existingWithNullParent = await _db.Litters
+                    .Where(l => l.FatherId == null || l.MotherId == null)
+                    .Select(l => new { l.Id, l.Name, l.FatherId, l.MotherId })
+                    .ToListAsync();
+                var sourceByName = litters
+                    .GroupBy(sl => $"Wurf {sl.LitterId}".Trim())
+                    .ToDictionary(g => g.Key, g => g.First());
+
+                Guid? ResolveParentForBackfill(string rawName)
+                {
+                    var n = Normalize(StripZucht(rawName));
+                    if (n.Length == 0) return null;
+                    if (createdAnimalByName.TryGetValue(n, out var fromLoadable) && persisted.Contains(fromLoadable))
+                        return fromLoadable;
+                    if (allDbNormToGid.TryGetValue(n, out var fromDb) && persisted.Contains(fromDb))
+                        return fromDb;
+                    return null;
+                }
+
+                foreach (var el in existingWithNullParent)
+                {
+                    if (!sourceByName.TryGetValue(el.Name, out var sl)) continue;
+                    var newF = el.FatherId == null ? ResolveParentForBackfill(sl.SireName) : null;
+                    var newM = el.MotherId == null ? ResolveParentForBackfill(sl.DamName) : null;
+                    if (newF is null && newM is null) continue;
+                    parentFksBackfilled++;
+                    if (execute)
+                    {
+                        var row = await _db.Litters.FirstOrDefaultAsync(l => l.Id == el.Id);
+                        if (row is not null)
+                        {
+                            if (newF is not null) row.FatherId = newF;
+                            if (newM is not null) row.MotherId = newM;
+                        }
+                    }
+                }
+                if (execute && parentFksBackfilled > 0) await _db.SaveChangesAsync();
+            }
+
             notes.Add("Quarantäne (kein Import): Konflikte + Stubs ohne Geburtsdatum + unsichere Wurf-Zuordnungen — warten auf die Prüfung durch die Züchterin.");
             if (parentLinksAdded > 0)
                 notes.Add($"Stammbaum-Diagramm: {parentLinksAdded} Tiere über Eltern-Verknüpfung einem (abgeleiteten) Wurf zugeordnet ({derivedLitters} abgeleitete Würfe).");
             notes.Add($"FK-Integrität: {litterParentFksDropped} Eltern-Verknüpfung(en) verworfen (Elternteil nicht ladbar), {derivedLittersSkipped} abgeleitete Würfe übersprungen (kein ladbares Elternteil). Bei 0/0 ist /import/execute FK-sicher.");
+            if (parentFksBackfilled > 0)
+                notes.Add($"Parent-FK-Backfill: {parentFksBackfilled} bereits importierte Würfe haben jetzt eine Eltern-Verknüpfung (Elternteil war zuvor in Quarantäne, jetzt geladen).");
             notes.Add($"Bestand/Herkunft: {residentTotal} im Bestand (Clan Kleine Chaoten), {externalTotal} externe Ahnen ({flippedByParentRule} davon über die Eltern-Regel als Bestand erkannt).");
             int conflictsResolvedByDecision = loadable.Count(a => a.ResolvedByDecision);
             if (conflictsResolvedByDecision > 0)
@@ -411,7 +468,7 @@ namespace GerbilManagerWebAPI.Import
 
             return new ImportReport(
                 Executed: execute,
-                Litters: new LitterSummary(litters.Count, littersCreated, littersExisting, derivedLitters, derivedLittersSkipped, litterParentFksDropped),
+                Litters: new LitterSummary(litters.Count, littersCreated, littersExisting, derivedLitters, derivedLittersSkipped, litterParentFksDropped, parentFksBackfilled),
                 Animals: new AnimalSummary(
                     animals.Count, animalsCreated, linked, fbMatched, fbUnmatched, animalsExisting,
                     new QuarantineSummary(conflicts, stubs, dateOnly, ambiguous, conflicts + stubs),

tools/import/conflict-decisions.json

@@ -81,7 +81,7 @@
     {
       "name": "Victoria Welby gen. Welby v.d. Kleinen Chaoten",
       "dob": "16.01.2023",
-      "decision": "E-locus = ee[f] (Fuchs). NOTE: this is the mother of animal 'C' (c-29042024) — un-quarantining her links C's second parent. Name kept in the merged record's v.d. spelling: extract.py decision matching uses norm_name (no v.d.<->von den fold) — workaround until the canon_pair matching fix lands.",
+      "decision": "E-locus = ee[f] (Fuchs). This is the mother of animal 'C' (c-29042024) — un-quarantining her links C's second parent. Name in v.d. spelling (workaround from Re-Import #2); both spellings now match after FIX-1 (canon_pair identity).",
       "genotype": "Aa CC D- ee[f] Gg pp Spsp [DP]",
       "source": "Julian 2026-06-06 — HUMANQUESTION D4"
     }

tools/import/extract.py

@@ -158,6 +158,10 @@ def parse_detail(text):
         tail = text[dob.end():]
         tail = re.sub(r"^\s*/?\+?\s?\d[\d.]*", "", tail)   # drop any /+death remnant
         tail = tail.lstrip(" ,").strip()
+        # FIX-4 (Skarlett): strip trailing "/ +YEAR" death-year artifacts leaked from compact
+        # chart cells (e.g. "… rere / +2018"). The DEATH regex still captures the year from
+        # the full cell text, so it appears as a death-date conflict — not a genotype conflict.
+        tail = re.sub(r"\s*/\s*\+\d{4}\s*$", "", tail).strip()
         if gt.looks_like_genotype(tail):
             geno = tail
     return (dob.group(1) if dob else "",
@@ -518,10 +522,11 @@ def _alleles_compatible(a, b):
     if a == b:
         return True
     if a == "?" or b == "?":
-        return False                       # unknown vs filled = contradiction (D- vs DD)
+        return True                        # specific-wins: unknown allele is compatible with any
+                                           # specified value (C- vs CC -> CC; G- vs Gg -> Gg)
     (ba, ma), (bb, mb) = _split_allele(a), _split_allele(b)
     if ba != bb:
-        return False                       # different base allele = real value diff (E vs e)
+        return False                       # different base allele = real value diff (E vs e, D vs d)
     return ma == "" or mb == ""            # same base, modifier present-vs-absent -> presence wins
 
 
@@ -622,9 +627,12 @@ def dedup(animals):
             if a.get("deaf") is not None:
                 deaf_seen.add(a["deaf"])
             tags_set.update(a.get("tags", []))
-        # pick the richest genotype (most mapped loci, then longest raw)
+        # pick the richest genotype: most mapped loci, then fewest unknowns ('?' alleles = specific
+        # wins, FIX-2), then longest raw string as final tiebreaker.
+        def _specificity(gd):
+            return sum(1 for pair in gd["mapped8locus"].values() for a in pair if a != "?")
         best = max((a["genotype"] for a in grp),
-                   key=lambda gd: (len(gd["mapped8locus"]), len(gd["rawGenotype"])))
+                   key=lambda gd: (len(gd["mapped8locus"]), _specificity(gd), len(gd["rawGenotype"])))
         out = {
             "id": slug(base["name"], base["dob"]),
             "name": base["name"],
@@ -887,21 +895,30 @@ def apply_dob_remaps(raw_animals, path):
     """PRE-dedup: a conflict-decision carrying `correctDob` marks a record as a DUPLICATE with a
     wrong birthdate — remap that raw record's DOB to correctDob so dedup MERGES it into the
     canonical same-named animal (e.g. Chelsea *15.10.2021 -> *02.04.2021). Match =
-    norm_name(name)+norm_dob(dob). Tolerates a missing/garbled file. Returns the remap count.
+    canon_pair(name)+(dob) with same Zucht-aware logic as apply_conflict_decisions (see there).
+    Tolerates a missing/garbled file. Returns the remap count.
     Must run BEFORE dedup (it changes the dedup identity). (god/HUMANQUESTION D — Dubletten.)"""
-    remaps = {}
+    remaps_full = {}   # (nameCanon, zuchtCanon, dob) -> correctDob  — decision carries Zucht
+    remaps_name = {}   # (nameCanon, dob) -> correctDob              — no Zucht in decision
     try:
         with open(path, encoding="utf-8") as fh:
             for r in (json.load(fh).get("resolutions") or []):
                 if r.get("correctDob"):
-                    remaps[(norm_name(r.get("name", "")), norm_dob(r.get("dob", "")))] = r["correctDob"]
+                    nc, zc = canon_pair(r.get("name", ""))
+                    dob = norm_dob(r.get("dob", ""))
+                    if zc:
+                        remaps_full[(nc, zc, dob)] = r["correctDob"]
+                    else:
+                        remaps_name[(nc, dob)] = r["correctDob"]
     except (OSError, ValueError):
         return 0
-    if not remaps:
+    if not remaps_full and not remaps_name:
         return 0
     n = 0
     for a in raw_animals:
-        new = remaps.get((norm_name(a.get("name", "")), norm_dob(a.get("dob", ""))))
+        nc, zc = canon_pair(a.get("name", ""))
+        dob = norm_dob(a.get("dob", ""))
+        new = remaps_full.get((nc, zc, dob)) or remaps_name.get((nc, dob))
         if new and a.get("dob") != new:
             a["dob"] = new
             n += 1
@@ -911,23 +928,36 @@ def apply_dob_remaps(raw_animals, path):
 def apply_conflict_decisions(merged, conflicts, path):
     """Consume human conflict resolutions (tools/import/conflict-decisions.json) so the wife's
     answers UN-QUARANTINE animals. Schema: {"resolutions":[{name, dob, decision, genotype?,
-    farbschlag?, source}]}. Match = norm_name(name)+norm_dob(dob) (same identity as dedup). A
-    matching animal: clear its conflict, mark resolvedByDecision; an explicit `genotype`
-    (breeder notation) is parsed and becomes authoritative, `farbschlag` overrides too. Tolerates
-    a missing/empty/garbled file. Returns the number of conflicts resolved. (god/HUMANQUESTION D.)"""
-    decisions = {}
+    farbschlag?, source}]}. Match = canon_pair(name)+(dob):
+      - When the decision name CARRIES a Zucht (zuchtCanon != ''), match on the FULL
+        (nameCanon, zuchtCanon, dob) triple — preserves the C3 rule that same name+DOB but
+        different Zucht = different animal.
+      - When the decision has NO Zucht, fall back to (nameCanon, dob) name-only match.
+    Both spellings v.d. / von den fold to the same canon. A matching animal: clear its conflict,
+    mark resolvedByDecision; an explicit `genotype` (breeder notation) is parsed and becomes
+    authoritative, `farbschlag` overrides too. Tolerates a missing/empty/garbled file.
+    Returns the number of conflicts resolved. (god/HUMANQUESTION D.)"""
+    decisions_full = {}   # (nameCanon, zuchtCanon, dob) -> r  — when decision carries a Zucht
+    decisions_name = {}   # (nameCanon, dob) -> r              — fallback, decision has no Zucht
     try:
         with open(path, encoding="utf-8") as fh:
             for r in (json.load(fh).get("resolutions") or []):
-                decisions[(norm_name(r.get("name", "")), norm_dob(r.get("dob", "")))] = r
+                nc, zc = canon_pair(r.get("name", ""))
+                dob = norm_dob(r.get("dob", ""))
+                if zc:
+                    decisions_full[(nc, zc, dob)] = r
+                else:
+                    decisions_name[(nc, dob)] = r
     except (OSError, ValueError):
         return 0
-    if not decisions:
+    if not decisions_full and not decisions_name:
         return 0
 
     resolved = 0
     for a in merged:
-        d = decisions.get((norm_name(a["name"]), norm_dob(a["dob"])))
+        nc, zc = canon_pair(a["name"])
+        dob = norm_dob(a["dob"])
+        d = decisions_full.get((nc, zc, dob)) or decisions_name.get((nc, dob))
         if not d:
             continue
         a["resolvedByDecision"] = True

tools/import/test_extract.py

@@ -102,6 +102,63 @@ check("apply_conflict_decisions returns resolved count", n == 2)
 check("missing decisions file tolerated (returns 0)",
       e.apply_conflict_decisions([], [], os.path.join(tempfile.gettempdir(), "does-not-exist.json")) == 0)
 
+# FIX-1: decision matching uses canon_pair identity -> 'von den' decision matches 'v.d.' record
+dec_vd = os.path.join(tempfile.gettempdir(), "decisions-vd.json")
+_json.dump({"resolutions": [
+    {"name": "Victoria Welby gen. Welby von den Kleinen Chaoten",  # written with 'von den'
+     "dob": "16.01.2023", "decision": "E-locus = ee[f]",
+     "genotype": "Aa CC D- ee[f] Gg pp Spsp", "source": "test"},
+]}, open(dec_vd, "w", encoding="utf-8"))
+merged_vd = [
+    {"id": "vw", "name": "Victoria Welby gen. Welby v.d. Kleinen Chaoten",  # record has 'v.d.'
+     "dob": "16.01.2023", "conflict": True, "farbschlag": "", "death": "",
+     "genotype": {"mapped8locus": {}, "rawGenotype": "", "unmappedTokens": []}},
+]
+conflicts_vd = [{"id": "vw"}]
+n_vd = e.apply_conflict_decisions(merged_vd, conflicts_vd, dec_vd)
+check("FIX-1: 'von den' decision matches 'v.d.' record (canon_pair identity)", n_vd == 1)
+check("FIX-1: conflict cleared for v.d. record", merged_vd[0]["conflict"] is False)
+# Also verify the workaround spelling (v.d. in decision) matches a 'von den' record
+_json.dump({"resolutions": [
+    {"name": "Victoria Welby gen. Welby v.d. Kleinen Chaoten",  # workaround: v.d. in decision
+     "dob": "16.01.2023", "decision": "E-locus = ee[f]",
+     "genotype": "Aa CC D- ee[f] Gg pp Spsp", "source": "test"},
+]}, open(dec_vd, "w", encoding="utf-8"))
+merged_vd2 = [
+    {"id": "vw2", "name": "Victoria Welby gen. Welby von den Kleinen Chaoten",  # record 'von den'
+     "dob": "16.01.2023", "conflict": True, "farbschlag": "", "death": "",
+     "genotype": {"mapped8locus": {}, "rawGenotype": "", "unmappedTokens": []}},
+]
+conflicts_vd2 = [{"id": "vw2"}]
+n_vd2 = e.apply_conflict_decisions(merged_vd2, conflicts_vd2, dec_vd)
+check("FIX-1: v.d. decision also matches 'von den' record (both spellings match)", n_vd2 == 1)
+try: os.remove(dec_vd)
+except OSError: pass
+
+# FIX-1 C3-rule: same name+DOB, two Zuchten -> decision hits ONLY the correct Zucht (C3 isolation)
+dec_c3 = os.path.join(tempfile.gettempdir(), "decisions-c3.json")
+_json.dump({"resolutions": [
+    # Decision only for Luna from ZdkC, NOT Luna from Black Forest
+    {"name": "Luna von den Kleinen Chaoten", "dob": "01.01.2020",
+     "decision": "D-locus = DD", "genotype": "aa CC DD ee gg PP spsp rere", "source": "test"},
+]}, open(dec_c3, "w", encoding="utf-8"))
+merged_c3 = [
+    {"id": "luna-kc", "name": "Luna von den Kleinen Chaoten", "dob": "01.01.2020",
+     "conflict": True, "farbschlag": "", "death": "",
+     "genotype": {"mapped8locus": {"D": ["D","?"]}, "rawGenotype": "D-", "unmappedTokens": []}},
+    {"id": "luna-bf", "name": "Luna of Black Forest", "dob": "01.01.2020",
+     "conflict": True, "farbschlag": "", "death": "",
+     "genotype": {"mapped8locus": {"D": ["D","?"]}, "rawGenotype": "D-", "unmappedTokens": []}},
+]
+conflicts_c3 = [{"id": "luna-kc"}, {"id": "luna-bf"}]
+n_c3 = e.apply_conflict_decisions(merged_c3, conflicts_c3, dec_c3)
+check("FIX-1 C3: decision hits only the correct Zucht (luna-kc resolved)", n_c3 == 1)
+check("FIX-1 C3: luna-kc conflict cleared (correct Zucht)", merged_c3[0]["conflict"] is False)
+check("FIX-1 C3: luna-bf conflict NOT cleared (different Zucht)", merged_c3[1]["conflict"] is True)
+check("FIX-1 C3: conflicts list has only luna-bf left", len(conflicts_c3) == 1 and conflicts_c3[0]["id"] == "luna-bf")
+try: os.remove(dec_c3)
+except OSError: pass
+
 # --- correctDob: a wrong-birthdate duplicate is remapped BEFORE dedup so it merges ---
 dec2 = os.path.join(tempfile.gettempdir(), "decisions-dob.json")
 _json.dump({"resolutions": [
@@ -128,23 +185,95 @@ except OSError: pass
 try: os.remove(dec_path)
 except OSError: pass
 
-# --- "presence wins" conflict rule (Julian) ---
-# present-vs-absent (whole locus or [f] modifier) is NOT a conflict; differing filled values are.
+# --- "presence wins" + "specific wins" conflict rules (Julian) ---
+# present-vs-absent (whole locus or [f] modifier) is NOT a conflict; differing FILLED values are.
+# FIX-2 (specific-wins): unknown allele '?' vs any specified value is also NOT a conflict —
+# the specific value wins (C- vs CC -> CC; G- vs Gg -> Gg; P? vs PP -> PP).
 check("spsp present vs locus absent -> no conflict",
       not e._genotype_conflict([{"Sp": ["sp", "sp"]}, {}]))
 check("ee[f] vs ee ([f] modifier present/absent) -> no conflict",
       not e._genotype_conflict([{"E": ["e", "e^f"]}, {"E": ["e", "e"]}]))
-check("DD vs D- (unknown vs filled) -> conflict",
-      e._genotype_conflict([{"D": ["D", "D"]}, {"D": ["D", "?"]}]))
+# FIX-2: '?' vs specified = specific wins (was: contradiction)
+check("FIX-2: DD vs D- (specific wins: DD wins) -> NOT conflict",
+      not e._genotype_conflict([{"D": ["D", "D"]}, {"D": ["D", "?"]}]))
+check("FIX-2: C- vs Cc[h] (specific wins: c^h wins) -> NOT conflict",
+      not e._genotype_conflict([{"C": ["C", "?"]}, {"C": ["C", "c^h"]}]))
+check("FIX-2: C- vs CC (specific wins: CC) -> NOT conflict",
+      not e._genotype_conflict([{"C": ["C", "?"]}, {"C": ["C", "C"]}]))
+check("FIX-2: G- vs Gg (specific wins) -> NOT conflict",
+      not e._genotype_conflict([{"G": ["G", "?"]}, {"G": ["G", "g"]}]))
+check("FIX-2: PP vs P? (specific wins: PP) -> NOT conflict",
+      not e._genotype_conflict([{"P": ["P", "P"]}, {"P": ["P", "?"]}]))
+# Genuine value contradictions (both alleles specified but different) still quarantine
 check("Ee vs ee (different base allele) -> conflict",
       e._genotype_conflict([{"E": ["E", "e"]}, {"E": ["e", "e"]}]))
-check("C- vs Cc[h] -> conflict",
-      e._genotype_conflict([{"C": ["C", "?"]}, {"C": ["C", "c^h"]}]))
-check("c[h] vs c[chm] (different modifiers) -> conflict",
+check("DD vs Dd (both specified, D vs d) -> conflict",
+      e._genotype_conflict([{"D": ["D", "D"]}, {"D": ["D", "d"]}]))
+check("PP vs Pp (both specified) -> conflict",
+      e._genotype_conflict([{"P": ["P", "P"]}, {"P": ["P", "p"]}]))
+check("c[h] vs c[chm] (different modifiers, both specified) -> conflict",
       not e._alleles_compatible("c^h", "c^chm"))
 check("identical genotypes -> no conflict",
       not e._genotype_conflict([{"A": ["A", "a"]}, {"A": ["A", "a"]}]))
 
+# FIX-2 MERGE: specific allele must survive the merge regardless of which variant comes first.
+# dedup() picks the most specific genotype (fewest '?' alleles); C- vs CC -> CC must win.
+def _minimal_animal(name, dob, mapped):
+    """Build a minimal raw animal dict suitable for dedup()."""
+    from genotype import parse as gparse
+    raw = " ".join(f"{l}{''.join(a)}" for l, pa in mapped.items() for a in [pa])
+    return {
+        "name": name, "dob": dob, "death": "", "gender": None,
+        "farbschlag": "", "breeder": "", "zucht": "", "parentRefs": [],
+        "photos": [], "sourceFiles": ["test.xlsx"], "tags": [],
+        "deaf": None, "conflict": False,
+        "genotype": {"mapped8locus": mapped, "rawGenotype": raw, "unmappedTokens": []},
+        "_gen": 0, "_col": 5, "_row": 10, "_file": "test.xlsx",
+        "_zucht": "",
+    }
+
+# Order A: C- first, CC second
+animals_merge_a = [
+    _minimal_animal("TestTier", "01.01.2020", {"C": ["C", "?"]}),  # C-
+    _minimal_animal("TestTier", "01.01.2020", {"C": ["C", "C"]}),  # CC
+]
+merged_ma, _, _, _ = e.dedup(animals_merge_a)
+check("FIX-2 merge A (C- first): result has CC not C-",
+      merged_ma[0]["genotype"]["mapped8locus"].get("C") == ["C", "C"])
+
+# Order B: CC first, C- second (must give same result)
+animals_merge_b = [
+    _minimal_animal("TestTier2", "02.02.2020", {"C": ["C", "C"]}),  # CC
+    _minimal_animal("TestTier2", "02.02.2020", {"C": ["C", "?"]}),  # C-
+]
+merged_mb, _, _, _ = e.dedup(animals_merge_b)
+check("FIX-2 merge B (CC first): result has CC not C-",
+      merged_mb[0]["genotype"]["mapped8locus"].get("C") == ["C", "C"])
+
+# G- vs Gg: Gg must win
+animals_merge_g = [
+    _minimal_animal("TestGGerbil", "03.03.2020", {"G": ["G", "?"]}),  # G-
+    _minimal_animal("TestGGerbil", "03.03.2020", {"G": ["G", "g"]}),  # Gg
+]
+merged_mg, _, _, _ = e.dedup(animals_merge_g)
+check("FIX-2 merge G (G- vs Gg): Gg wins",
+      merged_mg[0]["genotype"]["mapped8locus"].get("G") == ["G", "g"])
+
+# --- FIX-4: Skarlett parse artifact — trailing "/ +YEAR" stripped from geno, death captured ---
+dob4, death4, geno4 = e.parse_detail("Skarlett,*17.04.2016, aa C- DD ee Gg PP spsp rere / +2018")
+check("FIX-4: '/ +YEAR' artifact stripped from geno tail",
+      geno4 == "aa C- DD ee Gg PP spsp rere")
+check("FIX-4: death year still captured from full cell text",
+      death4 == "2018")
+check("FIX-4: DOB still correct",
+      dob4 == "17.04.2016")
+# Without artifact — must be unchanged
+dob5, death5, geno5 = e.parse_detail("*01.01.2020, aa C- DD ee Gg PP spsp rere")
+check("FIX-4: no artifact -> geno unchanged",
+      geno5 == "aa C- DD ee Gg PP spsp rere")
+check("FIX-4: no artifact -> no spurious death",
+      death5 == "")
+
 # --- name-bleed guard (a parent name is not a Farbschlag) ---
 check("v.d. name rejected", e.looks_like_animal_name("Tennessee von den Kleinen Chaoten"))
 check("gen.+v.d. name rejected", e.looks_like_animal_name("Victoria Welby gen. Welby v.d. Kleinen Chaoten"))