Gulum/GerbilManager
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(rennmausakte): Datenherkunft/Nachverfolgung pro Rennmaus

Browse Source 
Neuer „Datenherkunft"-Button in der Rennmausakte öffnet einen Dialog, der zeigt,
aus welchen Quellen der Eintrag erzeugt wurde: Quelldateien (Stammbäume +
Wurfchronik), Anzahl zusammengeführter Datensätze, Eltern-Herleitung
(Methode/Konfidenz) und Hinweise (z. B. „per manueller Entscheidung zugeordnet",
„Konflikt gelöst", „aus N Datensätzen zusammengeführt").

Import: build_provenance() in merge_and_resolve.py sammelt die Herkunft über alle
deduplizierten Datensätze und schreibt sie als Provenance-JSON je Tier in
resolved_import.json. Backend: Gerbil.Provenance (nullable text) + Migration
AddGerbilProvenance, gemappt im Ingest und im GerbilDto zurückgegeben. Frontend:
ProvenanceDialog + Typen + Strings.

Tests: Ingest-Round-trip (vorhanden/abwesend), e2e provenance.spec.ts.
dotnet(212)/vitest(129)/playwright/tsc/eslint grün.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Gulum
2026-06-22 15:45:15 +02:00
parent b9e5b031c4
commit 9c98c37d2a
16 changed files with 2126 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
GerbilManager.Tests/IngestResolvedServiceTests.cs
View File

@@ -84,7 +84,8 @@ namespace GerbilManager.Tests
CharacterTraits = new string[] {},
CharacterNote = (string?)null,
IsDeaf = false,
IsResident = true
IsResident = true,
Provenance = (string?)"{\"sourceFiles\":[\"Stammbaum von Papa.xlsx\",\"Wurfchronik-Detail.docx\"],\"mergedRecordCount\":2,\"fromWurfchronik\":true,\"notes\":[\"aus 2 Datensätzen zusammengeführt\"]}"
},
new
{
@@ -111,7 +112,8 @@ namespace GerbilManager.Tests
CharacterTraits = new string[] {},
CharacterNote = (string?)null,
IsDeaf = false,
IsResident = true
IsResident = true,
Provenance = (string?)null
}
},
GerbilPhotos = new[]
@@ -176,6 +178,12 @@ namespace GerbilManager.Tests
Assert.Equal(mother.Id, litter.MotherId);
Assert.Equal(father.Id, photo.GerbilId);
Assert.Equal(father.OriginContactId, mother.OriginContactId);
// Provenance round-trips verbatim through ingest (and stays null when absent).
Assert.NotNull(father.Provenance);
Assert.Contains("Stammbaum von Papa.xlsx", father.Provenance);
Assert.Contains("mergedRecordCount", father.Provenance);
Assert.Null(mother.Provenance);
}
}
}

3
GerbilManagerWebAPI/Dtos/ApiDtos.cs
View File

@@ -32,7 +32,8 @@ namespace GerbilManagerWebAPI.Dtos
bool? IsDeaf,
bool IsResident,
string? ProfilePhotoUrl,
bool IsCastrated);
bool IsCastrated,
string? Provenance);
public record LitterDto(
Guid Id,

3
GerbilManagerWebAPI/Endpoints/GerbilEndpoints.cs
View File

@@ -137,6 +137,7 @@ namespace GerbilManagerWebAPI.Endpoints
g.Id, g.Name, g.Gender, g.Status, g.LitterId, g.OriginContactId, g.ReceiverContactId,
g.EnclosureId, g.ColorVarietyId, g.DateOfBirth, g.DateOfDeath, g.CauseOfDeath,
g.GoHomeDate, g.Genotype, g.SpottingType, g.Notes, g.ImportSource, g.ExternalRef, g.OriginBreeder,
g.CharacterTraits, g.CharacterNote, g.IsDeaf, g.IsResident, profilePhotoUrl, g.IsCastrated);
g.CharacterTraits, g.CharacterNote, g.IsDeaf, g.IsResident, profilePhotoUrl, g.IsCastrated,
g.Provenance);
}
}

1
GerbilManagerWebAPI/Import/IngestResolvedService.cs
View File

@@ -160,6 +160,7 @@ namespace GerbilManagerWebAPI.Import
ImportSource = g.ImportSource,
ExternalRef = g.ExternalRef,
RawImportData = g.RawImportData,
Provenance = g.Provenance,
OriginBreeder = g.OriginBreeder,
NameSearch = g.NameSearch,
CharacterTraits = g.CharacterTraits,

1539
GerbilManagerWebAPI/Migrations/20260622133758_AddGerbilProvenance.Designer.cs generated Normal file
View File

File diff suppressed because it is too large Load Diff

28
GerbilManagerWebAPI/Migrations/20260622133758_AddGerbilProvenance.cs Normal file
View File

@@ -0,0 +1,28 @@
using Microsoft.EntityFrameworkCore.Migrations;
#nullable disable
namespace GerbilManagerWebAPI.Migrations
{
/// <inheritdoc />
public partial class AddGerbilProvenance : Migration
{
/// <inheritdoc />
protected override void Up(MigrationBuilder migrationBuilder)
{
migrationBuilder.AddColumn<string>(
name: "Provenance",
table: "Gerbils",
type: "text",
nullable: true);
}
/// <inheritdoc />
protected override void Down(MigrationBuilder migrationBuilder)
{
migrationBuilder.DropColumn(
name: "Provenance",
table: "Gerbils");
}
}
}

3
GerbilManagerWebAPI/Migrations/ApplicationContextModelSnapshot.cs
View File

@@ -912,6 +912,9 @@ namespace GerbilManagerWebAPI.Migrations
b.Property<Guid?>("OriginContactId")
.HasColumnType("uuid");
b.Property<string>("Provenance")
.HasColumnType("text");
b.Property<string>("RawImportData")
.HasColumnType("text");

7
GerbilManagerWebAPI/Models/Gerbil.cs
View File

@@ -49,6 +49,13 @@ namespace GerbilManagerWebAPI.Models
public string? ImportSource { get; set; }
public string? ExternalRef { get; set; }
/// <summary>Data-provenance / traceability for the import: a JSON object describing
/// WHICH source information produced this entry (sourceFiles, mergedRecordCount,
/// fromWurfchronik, parentMethod/parentConfidence, notes). Written by the Python
/// merge_and_resolve step and surfaced read-only in the Rennmausakte
/// ("Nachverfolgungsinformationen"). Null = manually-added animal / no import data.</summary>
public string? Provenance { get; set; }
/// <summary>Raw import payload preserved verbatim (rawGenotype + unmappedTokens like
/// the Uw locus / WFNZ markers) so nothing from the spreadsheets is lost. JSON text.</summary>
public string? RawImportData { get; set; }

9
gerbil-manager-web/e2e/mock-data.ts
View File

@@ -122,6 +122,15 @@ export function seedDb(): MockDb {
enclosureId: 'enc-gross',
originContactId: 'con-meier',
originBreeder: 'Clan-Kleine-Chaoten',
// NACHVERFOLGUNG: Datenherkunft (vom Import erzeugter JSON-String).
provenance: JSON.stringify({
sourceFiles: ['Stammbaum von Krümel.xlsx', 'Wurfchronik-Detail.docx'],
mergedRecordCount: 2,
fromWurfchronik: true,
parentMethod: 'chart-position',
parentConfidence: 'medium',
notes: ['aus 2 Datensätzen zusammengeführt'],
}),
},
{ ...gerbil('fridolin', 'Fridolin', 'male', '2023-05-01', 'w-fridolin', 'cv-schwarz', 'aa CC DD EE GG PP spsp rere'), enclosureId: 'enc-gross', originBreeder: 'Zoohandlung Meier', profilePhotoUrl: 'data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==' },
gerbil('luna', 'Luna', 'female', '2023-08-15', 'w-luna', 'cv-gold', 'AA CC DD EE GG pp spsp rere'),

46
gerbil-manager-web/e2e/provenance.spec.ts Normal file
View File

@@ -0,0 +1,46 @@
/** NACHVERFOLGUNG: "Datenherkunft"-Button in der Tierakte öffnet den
* Nachverfolgungs-Dialog mit Quelldateien, Datensatzanzahl und Hinweisen. */
import { de, expect, skipUnlessMock, test } from './fixtures'
const p = de.provenance
test('Tierakte: "Datenherkunft"-Button öffnet den Nachverfolgungs-Dialog', async ({ page }) => {
skipUnlessMock()
await page.goto('/rennmaeuse/kruemel')
// Button in den Aktionen öffnet den Dialog.
await page.getByRole('button', { name: p.button }).click()
const dialog = page.getByRole('dialog', { name: p.dialogTitle })
await expect(dialog).toBeVisible()
// Quelldateien werden angezeigt.
await expect(dialog).toContainText(p.sourceFilesTitle)
await expect(dialog).toContainText('Stammbaum von Krümel.xlsx')
await expect(dialog).toContainText('Wurfchronik-Detail.docx')
// Zusammenführungs-Info + Wurfchronik-Hinweis.
await expect(dialog).toContainText(p.mergedCount(2))
await expect(dialog).toContainText(p.fromWurfchronik)
// Eltern-Herkunft (übersetzte Methode/Konfidenz).
await expect(dialog).toContainText(p.methods['chart-position'])
await expect(dialog).toContainText(p.confidences.medium)
// Hinweis aus dem Import.
await expect(dialog).toContainText('aus 2 Datensätzen zusammengeführt')
// Schließen (Footer-Button — Header-Button trägt dasselbe aria-label).
await dialog.locator('.provenance__actions').getByRole('button', { name: p.close }).click()
await expect(dialog).not.toBeVisible()
})
test('Tierakte: Tier ohne Importdaten zeigt "Keine Herkunftsdaten"', async ({ page }) => {
skipUnlessMock()
// Luna hat keine provenance im Mock → Leer-Zustand.
await page.goto('/rennmaeuse/luna')
await page.getByRole('button', { name: p.button }).click()
const dialog = page.getByRole('dialog', { name: p.dialogTitle })
await expect(dialog).toBeVisible()
await expect(dialog).toContainText(p.empty)
})

26
gerbil-manager-web/src/api/types.ts
View File

@@ -62,6 +62,32 @@ export interface Gerbil {
profilePhotoUrl?: string | null
spottingType?: string | null
isCastrated?: boolean
/**
* NACHVERFOLGUNG: Datenherkunft des Import-Eintrags als JSON-String (vom
* Python-Merge erzeugt; siehe {@link GerbilProvenance} für die geparste Form).
* null = manuell angelegtes Tier / keine Importdaten.
*/
provenance?: string | null
}
/**
* NACHVERFOLGUNG: geparste Datenherkunft eines importierten Tiers (JSON aus
* {@link Gerbil.provenance}). Beschreibt, WELCHE Quellinformationen den
* Datenbankeintrag erzeugt haben.
*/
export interface GerbilProvenance {
/** Alle Quelldateien (Stammbäume / Wurfchronik), die beigetragen haben. */
sourceFiles: string[]
/** Anzahl der Rohdatensätze, die zu diesem Tier zusammengeführt wurden. */
mergedRecordCount: number
/** Hat ein Wurfchronik-Treffer beigetragen? */
fromWurfchronik: boolean
/** Wie die Eltern abgeleitet wurden (z. B. "chart-position", "decision"). */
parentMethod?: string
/** Konfidenz der Eltern-Ableitung (z. B. "low" | "medium" | "high"). */
parentConfidence?: string
/** Menschlich lesbare Herkunftshinweise (deutsch). */
notes: string[]
}
/** Payload for POST /gerbils. */

157
gerbil-manager-web/src/components/ProvenanceDialog.tsx Normal file
View File

@@ -0,0 +1,157 @@
/**
* NACHVERFOLGUNG: "Nachverfolgungsinformationen" — read-only modal showing the
* data provenance of an imported gerbil (which source files contributed, how many
* raw records were merged, how the parents were derived, and human-readable notes).
*
* The provenance arrives as a JSON string on the Gerbil DTO (produced by the Python
* merge step). We parse it here defensively; anything unparseable / absent is shown
* as "Keine Herkunftsdaten vorhanden." Mirrors the ReportErrorDialog modal pattern.
*/
import { useEffect } from 'react'
import { de } from '../strings/de'
import type { GerbilProvenance } from '../api/types'
import './provenanceDialog.css'
interface ProvenanceDialogProps {
open: boolean
onClose: () => void
/** Raw provenance JSON string from Gerbil.provenance (null = no import data). */
provenance?: string | null
}
/** Parse the JSON provenance string defensively; null on absence / parse error. */
function parseProvenance(raw?: string | null): GerbilProvenance | null {
if (!raw || !raw.trim()) return null
try {
const p = JSON.parse(raw) as Partial<GerbilProvenance>
return {
sourceFiles: Array.isArray(p.sourceFiles) ? p.sourceFiles : [],
mergedRecordCount: typeof p.mergedRecordCount === 'number' ? p.mergedRecordCount : 0,
fromWurfchronik: Boolean(p.fromWurfchronik),
parentMethod: p.parentMethod,
parentConfidence: p.parentConfidence,
notes: Array.isArray(p.notes) ? p.notes : [],
}
} catch {
return null
}
}
export default function ProvenanceDialog({ open, onClose, provenance }: ProvenanceDialogProps) {
// Close on Escape (only while mounted).
useEffect(() => {
if (!open) return
const onKey = (e: KeyboardEvent) => {
if (e.key === 'Escape') onClose()
}
window.addEventListener('keydown', onKey)
return () => window.removeEventListener('keydown', onKey)
}, [open, onClose])
if (!open) return null
const t = de.provenance
const p = parseProvenance(provenance)
const method = p?.parentMethod ? (t.methods[p.parentMethod] ?? p.parentMethod) : null
const confidence = p?.parentConfidence
? (t.confidences[p.parentConfidence] ?? p.parentConfidence)
: null
return (
<div className="provenance__backdrop" onClick={onClose} role="presentation">
<div
className="provenance__dialog"
role="dialog"
aria-modal="true"
aria-labelledby="provenance-title"
onClick={(e) => e.stopPropagation()}
>
<div className="provenance__header">
<h2 id="provenance-title" className="provenance__title">
{t.dialogTitle}
</h2>
<button
type="button"
className="provenance__close"
onClick={onClose}
aria-label={t.close}
>
✕
</button>
</div>
{!p ? (
<p className="provenance__empty">{t.empty}</p>
) : (
<>
<p className="provenance__intro">{t.intro}</p>
<section className="provenance__section">
<div className="provenance__section-title">
{t.mergedTitle}
</div>
<p className="provenance__merged">{t.mergedCount(p.mergedRecordCount)}</p>
{p.fromWurfchronik && (
<p className="provenance__wurfchronik">📖 {t.fromWurfchronik}</p>
)}
</section>
<section className="provenance__section">
<div className="provenance__section-title">
{t.sourceFilesTitle} — {t.sourceFilesCount(p.sourceFiles.length)}
</div>
{p.sourceFiles.length === 0 ? (
<p className="provenance__empty">{t.empty}</p>
) : (
<ul className="provenance__files">
{p.sourceFiles.map((f) => (
<li key={f} className="provenance__file">
{f}
</li>
))}
</ul>
)}
</section>
{(method || confidence) && (
<section className="provenance__section">
<div className="provenance__section-title">{t.parentsTitle}</div>
<dl className="provenance__kvlist">
{method && (
<div className="provenance__kv">
<dt>{t.parentMethodLabel}</dt>
<dd>{method}</dd>
</div>
)}
{confidence && (
<div className="provenance__kv">
<dt>{t.parentConfidenceLabel}</dt>
<dd>{confidence}</dd>
</div>
)}
</dl>
</section>
)}
{p.notes.length > 0 && (
<section className="provenance__section">
<div className="provenance__section-title">{t.notesTitle}</div>
<ul className="provenance__notes">
{p.notes.map((n) => (
<li key={n}>{n}</li>
))}
</ul>
</section>
)}
</>
)}
<div className="provenance__actions">
<button type="button" className="btn btn--primary" onClick={onClose}>
{t.close}
</button>
</div>
</div>
</div>
)
}

157
gerbil-manager-web/src/components/provenanceDialog.css Normal file
View File

@@ -0,0 +1,157 @@
/* NACHVERFOLGUNG: "Nachverfolgungsinformationen" modal dialog. Mirrors the
reportErrorDialog modal pattern. */
.provenance__backdrop {
position: fixed;
inset: 0;
z-index: 200;
display: flex;
align-items: center;
justify-content: center;
padding: 1rem;
background: rgba(43, 33, 25, 0.45);
}
.provenance__dialog {
width: 100%;
max-width: 32rem;
max-height: calc(100vh - 2rem);
overflow-y: auto;
background: var(--color-bg, #fff);
color: var(--color-text);
border-radius: 14px;
box-shadow: 0 12px 40px rgba(43, 33, 25, 0.32);
padding: 1.25rem;
animation: provenanceIn 0.18s cubic-bezier(0.22, 1, 0.36, 1);
}
@keyframes provenanceIn {
from {
opacity: 0;
transform: translateY(12px);
}
to {
opacity: 1;
transform: none;
}
}
.provenance__header {
display: flex;
align-items: center;
justify-content: space-between;
gap: 0.5rem;
margin-bottom: 0.5rem;
}
.provenance__title {
margin: 0;
font-size: 1.15rem;
}
.provenance__close {
background: none;
border: none;
font: inherit;
font-size: 1.1rem;
line-height: 1;
cursor: pointer;
color: var(--color-text-muted, #666);
padding: 0.25rem 0.4rem;
border-radius: 6px;
}
.provenance__close:hover {
background: var(--color-border, #eee);
}
.provenance__intro {
margin: 0 0 0.85rem;
font-size: 0.9rem;
color: var(--color-text-muted, #666);
}
.provenance__empty {
margin: 0.4rem 0;
font-size: 0.9rem;
color: var(--color-text-muted, #888);
font-style: italic;
}
.provenance__section {
margin-top: 0.9rem;
padding: 0.6rem 0.75rem;
border: 1px solid var(--color-border, #e5e5e5);
border-radius: 8px;
background: var(--color-surface-2, rgba(0, 0, 0, 0.03));
}
.provenance__section-title {
font-size: 0.78rem;
font-weight: 600;
text-transform: uppercase;
letter-spacing: 0.03em;
color: var(--color-text-muted, #888);
margin-bottom: 0.4rem;
}
.provenance__merged {
margin: 0;
font-size: 0.9rem;
}
.provenance__wurfchronik {
margin: 0.4rem 0 0;
font-size: 0.85rem;
color: var(--color-text-muted, #666);
}
.provenance__files {
margin: 0;
padding-left: 1.1rem;
display: grid;
gap: 0.2rem;
}
.provenance__file {
font-size: 0.85rem;
word-break: break-word;
}
.provenance__kvlist {
margin: 0;
display: grid;
gap: 0.25rem;
}
.provenance__kv {
display: flex;
gap: 0.5rem;
font-size: 0.85rem;
}
.provenance__kv dt {
flex: 0 0 6rem;
font-weight: 600;
color: var(--color-text-muted, #777);
}
.provenance__kv dd {
margin: 0;
min-width: 0;
word-break: break-word;
}
.provenance__notes {
margin: 0;
padding-left: 1.1rem;
display: grid;
gap: 0.2rem;
font-size: 0.85rem;
}
.provenance__actions {
display: flex;
justify-content: flex-end;
gap: 0.5rem;
margin-top: 1.1rem;
}

11
gerbil-manager-web/src/pages/GerbilDetailPage.tsx
View File

@@ -14,6 +14,7 @@ import GerbilHealthTab from '../components/GerbilHealthTab'
import GerbilPhotosTab from '../components/GerbilPhotosTab'
import GerbilProfilePhoto from '../components/GerbilProfilePhoto'
import GerbilWeightTab from '../components/GerbilWeightTab'
import ProvenanceDialog from '../components/ProvenanceDialog'
import ReportErrorDialog from '../components/ReportErrorDialog'
import { useGerbilName } from '../components/breederSuffix'
import { useToast } from '../components/toast'
@@ -59,6 +60,7 @@ export default function GerbilDetailPage() {
const { id = '' } = useParams()
const [tab, setTab] = useState<DetailTab>('photos')
const [reportOpen, setReportOpen] = useState(false)
const [provenanceOpen, setProvenanceOpen] = useState(false)
const gerbil = useApi(() => getGerbil(id), [id])
const forSale = useMutation(() => updateGerbil(id, { status: 'ForSale' }))
@@ -258,6 +260,9 @@ export default function GerbilDetailPage() {
{de.pages.vertraege.wizard.title}
</Link>
)}
<button type="button" className="ak-btn" onClick={() => setProvenanceOpen(true)}>
{de.provenance.button}
</button>
<button type="button" className="ak-btn" onClick={() => setReportOpen(true)}>
{de.feedback.button}
</button>
@@ -482,6 +487,12 @@ export default function GerbilDetailPage() {
onClose={() => setReportOpen(false)}
context={{ context: 'gerbil-detail', gerbilId: g.id, entityName: g.name || de.pages.gerbils.nameless }}
/>
<ProvenanceDialog
open={provenanceOpen}
onClose={() => setProvenanceOpen(false)}
provenance={g.provenance}
/>
</section>
)
}

31
gerbil-manager-web/src/strings/de.ts
View File

@@ -878,6 +878,37 @@ export const de = {
idCopied: 'ID kopiert',
idCopyFailed: 'ID konnte nicht kopiert werden.',
},
/** NACHVERFOLGUNG: Datenherkunft eines importierten Tiers (Rennmausakte). */
provenance: {
/** Button-Beschriftung in den Akten-Aktionen. */
button: 'Datenherkunft',
dialogTitle: 'Nachverfolgungsinformationen',
intro: 'Woher stammen die Daten dieses Eintrags? Diese Angaben werden beim Import automatisch erfasst.',
/** Überschriften / Feldbeschriftungen. */
sourceFilesTitle: 'Quelldateien',
sourceFilesCount: (n: number) => (n === 1 ? 'aus 1 Quelle' : `aus ${n} Quellen`),
mergedTitle: 'Datensätze',
mergedCount: (n: number) =>
n === 1 ? 'aus 1 Datensatz' : `zusammengeführt aus ${n} Datensätzen`,
fromWurfchronik: 'Enthält Angaben aus der Wurfchronik',
parentsTitle: 'Eltern-Herkunft',
parentMethodLabel: 'Methode',
parentConfidenceLabel: 'Konfidenz',
/** Übersetzungen für die technischen Methoden-/Konfidenz-Werte. */
methods: {
'chart-position': 'Position im Stammbaum',
decision: 'Manuelle Entscheidung',
} as Record<string, string>,
confidences: {
low: 'niedrig',
medium: 'mittel',
high: 'hoch',
} as Record<string, string>,
notesTitle: 'Hinweise',
/** Wenn ein Tier keine Importdaten hat (manuell angelegt). */
empty: 'Keine Herkunftsdaten vorhanden.',
close: 'Schließen',
},
common: {
loading: 'Lädt …',
retry: 'Erneut versuchen',

99
tools/import/merge_and_resolve.py
View File

@@ -1520,6 +1520,8 @@ def main():
"IsResident": a_id in stammbaum_resident_ids,
"parentRefs": a.get("parentRefs", []),
"_photos": a.get("photos", []),
"_conflict": bool(a.get("conflict")),
"_resolved_by_decision": bool(a.get("resolvedByDecision")),
"_old_scoped_litter_id": scoped_litter_id,
"_eff_dob": dob_val or "2010-01-01",
"_birth_date": dob_val,
@@ -1678,6 +1680,80 @@ def main():
return True
def _record_source_files(g):
"""All distinct source files a single raw record drew from.
ImportSource is either a comma-joined Stammbaum file list, a single
Wurfchronik filename, or the per-litter filename; _filename is the
primary file. We union both so nothing is lost."""
files = set()
imp = g.get("ImportSource")
if imp:
for part in str(imp).split(","):
part = part.strip()
if part:
files.add(part)
fn = g.get("_filename")
if fn:
files.add(fn)
return files
def build_provenance(records, best_g, extra_notes=None):
"""Aggregate data-provenance across every raw record merged into one
resolved gerbil. Returns a JSON string (stored on the Gerbil entity as a
nullable text column) so the Rennmausakte can show where the entry came
from. `records` is the list of raw records that became this gerbil;
`best_g` is the chosen primary record."""
source_files = set()
from_wurfchronik = False
any_conflict = False
any_decision = False
for r in records:
source_files |= _record_source_files(r)
for f in _record_source_files(r):
if "wurfchronik" in f.lower():
from_wurfchronik = True
if r.get("_conflict"):
any_conflict = True
if r.get("_resolved_by_decision"):
any_decision = True
notes = []
merged_count = len(records)
if merged_count > 1:
notes.append(f"aus {merged_count} Datensätzen zusammengeführt")
if any_decision:
notes.append("per manueller Entscheidung zugeordnet")
if any_conflict:
notes.append("Konflikt per Entscheidung gelöst")
# Parent derivation: surface the strongest parentRef method/confidence
# the primary record carries (chart-position etc.).
parent_method = None
parent_confidence = None
for ref in best_g.get("parentRefs", []) or []:
if ref.get("method") and not parent_method:
parent_method = ref.get("method")
if ref.get("confidence") and not parent_confidence:
parent_confidence = ref.get("confidence")
if extra_notes:
for n in extra_notes:
if n and n not in notes:
notes.append(n)
prov = {
"sourceFiles": sorted(source_files),
"mergedRecordCount": merged_count,
"fromWurfchronik": from_wurfchronik,
"notes": notes,
}
if parent_method:
prov["parentMethod"] = parent_method
if parent_confidence:
prov["parentConfidence"] = parent_confidence
return json.dumps(prov, ensure_ascii=False)
# Group gerbils by name to perform deduplication
gerbil_groups = {}
for g in all_processed_gerbils:
@@ -1701,6 +1777,7 @@ def main():
if is_placeholder:
# Placeholders: do NOT merge, keep all separate
for g in group:
g["Provenance"] = build_provenance([g], g)
resolved_gerbils.append(g)
gerbil_id_map[g["Id"]] = g["Id"]
continue
@@ -1721,6 +1798,7 @@ def main():
for sub in sub_groups:
if len(sub) == 1:
g = sub[0]
g["Provenance"] = build_provenance([g], g)
resolved_gerbils.append(g)
gerbil_id_map[g["Id"]] = g["Id"]
continue
@@ -1806,6 +1884,7 @@ def main():
merged_notes.append(g["Notes"])
# Reconcile fields based on number of source files supporting them
conflict_notes = []
for field in ["DateOfBirth", "DateOfDeath", "Gender", "Genotype", "ColorVarietyId"]:
votes = {}
for g in sub:
@@ -1816,6 +1895,19 @@ def main():
votes[val] = votes.get(val, 0) + sources_count
if votes:
best_val = max(votes, key=votes.get)
# If the records disagreed on a field, the merge had to pick a
# winner — record that as a provenance note.
if len(votes) > 1:
FIELD_LABEL = {
"DateOfBirth": "Geburtsdatum",
"DateOfDeath": "Sterbedatum",
"Gender": "Geschlecht",
"Genotype": "Genotyp",
"ColorVarietyId": "Farbschlag",
}
conflict_notes.append(
f"Konflikt bei {FIELD_LABEL[field]} per Mehrheitsentscheidung gelöst"
)
best_g[field] = best_val
# Keep helper fields in sync if we changed DateOfBirth
if field == "DateOfBirth":
@@ -1824,12 +1916,13 @@ def main():
if merged_notes:
best_g["Notes"] = " | ".join(merged_notes)
best_g["_photos"] = merged_photos
best_g["Provenance"] = build_provenance(sub, best_g, extra_notes=conflict_notes)
# Print merge trace
print(f"Deduplicated same-animal name '{best_g['Name']}': merged {len(sub)} entries across files: {', '.join(sources)}")
resolved_gerbils.append(best_g)
gerbil_id_map[best_g["Id"]] = best_g["Id"]
@@ -1885,6 +1978,8 @@ def main():
del g["_birth_date"]
del g["_filename"]
del g["_old_id"]
g.pop("_conflict", None)
g.pop("_resolved_by_decision", None)
# Gather final valid gerbil IDs
valid_gerbil_ids = {g["Id"] for g in resolved_gerbils}