Merge pull request 'feat: fruit search by name/synonym and type filter (story #06)' (#6) from feature/06-fruit-search into main

Reviewed-on: #6

.claude/CLAUDE.md

@@ -0,0 +1,11 @@
+## Taiga
+- TAIGA_URL: https://taiga.db-extern.de
+- TAIGA_PROJECT_SLUG: obstsortendatenbank
+
+## Stories
+Before implementing a story: read its spec file in docs/planning/specs/ first.
+
+## Branching
+Stories developed sequentially; prior story branch may not be merged to main yet.
+Before creating feature branch: run `git branch -r | grep feature/` — if predecessor
+story branch exists on remote and is unmerged, base new branch on it, not main.

Makefile

@@ -34,7 +34,7 @@ run-frontend:
 test:
 	cd backend && go test ./...
 	cd frontend && npm run test -- --run
-	cd scripts && python3 -m unittest import_fruits_test -v
+	cd scripts && python3 -m unittest import_fruits_test import_publications_test -v
 
 ## fmt: format all code
 fmt:

README.md

@@ -6,7 +6,8 @@ A full-stack fruit-variety database (Go + Vue 3).
 
 - Users can view a "Hello, OSDB!" landing page that confirms the backend is reachable via the Vite proxy.
 - Users can create, view, edit, and delete fruit varieties, including managing synonyms and uploading images.
-- Administrators can bulk-import the legacy fruit database from XML using `scripts/import_fruits.py`.
+- Administrators can bulk-import the legacy fruit database from XML using `scripts/import_fruits.py`, and then import all publications (cover images, linked fruits, PDFs, fruit images) using `scripts/import_publications.py`.
+- Users can search fruits by name or synonym (case-insensitive, debounced) and filter by type or combined-type alias (e.g. "Birnen- und Quittensorten") from the fruit list.
 - Users can manage publications (books, catalogues) with cover images, linked fruits, per-fruit PDF descriptions, and fruit images; fruit detail pages show publication descriptions and images.
 
 ## Quick Start

backend/internal/domain/fruit.go

@@ -2,6 +2,14 @@ package domain
 
 import "time"
 
+// FruitTypeAliases maps combined-type alias labels to the enum values they expand to.
+var FruitTypeAliases = map[string][]string{
+	"Birnen- und Quittensorten":  {"Birnensorten", "Quittensorten"},
+	"Aprikosen und Pfirsiche":    {"Aprikosen", "Pfirsiche"},
+	"Mirabellen und Reineclauden": {"Mirabellen", "Renekloden"},
+	"Pflaumen und Zwetschen":     {"Pflaumen", "Zwetschen"},
+}
+
 type Fruit struct {
 	ID          int          `json:"id"`
 	Name        string       `json:"name"`

backend/internal/handler/fruit_handler.go

@@ -21,7 +21,7 @@ var (
 // FruitRepository is the consumer-defined interface the handler depends on.
 // The pg implementation in the repository package satisfies this structurally.
 type FruitRepository interface {
-	List(ctx context.Context, limit, offset int) ([]domain.Fruit, int, error)
+	List(ctx context.Context, limit, offset int, name string, types []string) ([]domain.Fruit, int, error)
 	Get(ctx context.Context, id int) (domain.Fruit, error)
 	Create(ctx context.Context, dto domain.FruitWriteDTO) (domain.Fruit, error)
 	Update(ctx context.Context, id int, dto domain.FruitWriteDTO) (domain.Fruit, error)
@@ -107,7 +107,18 @@ func (h *FruitHandler) List(c echo.Context) error {
 		}
 	}
 
-	fruits, total, err := h.repo.List(c.Request().Context(), limit, offset)
+	name := c.QueryParam("name")
+	var types []string
+	if typeParam := c.QueryParam("type"); typeParam != "" {
+		if expanded, ok := domain.FruitTypeAliases[typeParam]; ok {
+			types = expanded
+		} else if _, ok := validFruitTypes[typeParam]; ok {
+			types = []string{typeParam}
+		}
+		// unknown typeParam → types stays nil → no filter applied
+	}
+
+	fruits, total, err := h.repo.List(c.Request().Context(), limit, offset, name, types)
 	if err != nil {
 		return c.JSON(http.StatusInternalServerError, map[string]string{"error": "internal server error"})
 	}

backend/internal/handler/fruit_handler_test.go

@@ -52,9 +52,32 @@ func newFakeRepo() *fakeRepo {
 	}
 }
 
-func (r *fakeRepo) List(_ context.Context, limit, offset int) ([]domain.Fruit, int, error) {
+func (r *fakeRepo) List(_ context.Context, limit, offset int, name string, types []string) ([]domain.Fruit, int, error) {
+	nameLower := strings.ToLower(name)
+	typeSet := make(map[string]struct{}, len(types))
+	for _, t := range types {
+		typeSet[t] = struct{}{}
+	}
 	all := make([]domain.Fruit, 0, len(r.fruits))
 	for _, f := range r.fruits {
+		if name != "" {
+			nameMatch := strings.Contains(strings.ToLower(f.Name), nameLower)
+			synMatch := false
+			for _, s := range f.Synonyms {
+				if strings.Contains(strings.ToLower(s), nameLower) {
+					synMatch = true
+					break
+				}
+			}
+			if !nameMatch && !synMatch {
+				continue
+			}
+		}
+		if len(types) > 0 {
+			if _, ok := typeSet[f.FruitType]; !ok {
+				continue
+			}
+		}
 		all = append(all, f)
 	}
 	total := len(all)
@@ -241,6 +264,79 @@ func TestFruitList_WithItems(t *testing.T) {
 	}
 }
 
+func TestFruitList_FilterByName(t *testing.T) {
+	repo := newFakeRepo()
+	repo.fruits[1] = domain.Fruit{ID: 1, Name: "Boskop", OSDBNumber: "A001", FruitType: "Apfelsorten", Synonyms: []string{}, Images: []domain.FruitImage{}}
+	repo.fruits[2] = domain.Fruit{ID: 2, Name: "Cox Orange", OSDBNumber: "A002", FruitType: "Apfelsorten", Synonyms: []string{}, Images: []domain.FruitImage{}}
+	h := handler.NewFruitHandler(repo)
+	e := newEcho()
+	req := httptest.NewRequest(http.MethodGet, "/api/v1/fruits?name=Boskop", nil)
+	rec := httptest.NewRecorder()
+	c := e.NewContext(req, rec)
+	if err := h.List(c); err != nil {
+		t.Fatal(err)
+	}
+	if rec.Code != http.StatusOK {
+		t.Fatalf("want 200 got %d", rec.Code)
+	}
+	var resp domain.FruitListResponse
+	json.Unmarshal(rec.Body.Bytes(), &resp)
+	if resp.Total != 1 {
+		t.Fatalf("want 1 result got %d", resp.Total)
+	}
+	if resp.Items[0].Name != "Boskop" {
+		t.Fatalf("want Boskop got %s", resp.Items[0].Name)
+	}
+}
+
+func TestFruitList_FilterByType(t *testing.T) {
+	repo := newFakeRepo()
+	repo.fruits[1] = domain.Fruit{ID: 1, Name: "Boskop", OSDBNumber: "A001", FruitType: "Apfelsorten", Synonyms: []string{}, Images: []domain.FruitImage{}}
+	repo.fruits[2] = domain.Fruit{ID: 2, Name: "Williams", OSDBNumber: "B001", FruitType: "Birnensorten", Synonyms: []string{}, Images: []domain.FruitImage{}}
+	h := handler.NewFruitHandler(repo)
+	e := newEcho()
+	req := httptest.NewRequest(http.MethodGet, "/api/v1/fruits?type=Apfelsorten", nil)
+	rec := httptest.NewRecorder()
+	c := e.NewContext(req, rec)
+	if err := h.List(c); err != nil {
+		t.Fatal(err)
+	}
+	if rec.Code != http.StatusOK {
+		t.Fatalf("want 200 got %d", rec.Code)
+	}
+	var resp domain.FruitListResponse
+	json.Unmarshal(rec.Body.Bytes(), &resp)
+	if resp.Total != 1 {
+		t.Fatalf("want 1 result got %d", resp.Total)
+	}
+	if resp.Items[0].FruitType != "Apfelsorten" {
+		t.Fatalf("want Apfelsorten got %s", resp.Items[0].FruitType)
+	}
+}
+
+func TestFruitList_AliasExpandsToMultipleTypes(t *testing.T) {
+	repo := newFakeRepo()
+	repo.fruits[1] = domain.Fruit{ID: 1, Name: "Williams", OSDBNumber: "B001", FruitType: "Birnensorten", Synonyms: []string{}, Images: []domain.FruitImage{}}
+	repo.fruits[2] = domain.Fruit{ID: 2, Name: "Quitte", OSDBNumber: "Q001", FruitType: "Quittensorten", Synonyms: []string{}, Images: []domain.FruitImage{}}
+	repo.fruits[3] = domain.Fruit{ID: 3, Name: "Boskop", OSDBNumber: "A001", FruitType: "Apfelsorten", Synonyms: []string{}, Images: []domain.FruitImage{}}
+	h := handler.NewFruitHandler(repo)
+	e := newEcho()
+	req := httptest.NewRequest(http.MethodGet, "/api/v1/fruits?type=Birnen-+und+Quittensorten", nil)
+	rec := httptest.NewRecorder()
+	c := e.NewContext(req, rec)
+	if err := h.List(c); err != nil {
+		t.Fatal(err)
+	}
+	if rec.Code != http.StatusOK {
+		t.Fatalf("want 200 got %d", rec.Code)
+	}
+	var resp domain.FruitListResponse
+	json.Unmarshal(rec.Body.Bytes(), &resp)
+	if resp.Total != 2 {
+		t.Fatalf("want 2 results (Birnensorten + Quittensorten) got %d", resp.Total)
+	}
+}
+
 // -- Get --
 
 func TestFruitGet_Found(t *testing.T) {

backend/internal/repository/fruit_repo.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"strings"
 
 	"github.com/jackc/pgx/v5"
 	"github.com/jackc/pgx/v5/pgconn"
@@ -33,15 +34,35 @@ func mapPgError(err error) error {
 	return err
 }
 
-func (r *FruitRepo) List(ctx context.Context, limit, offset int) ([]domain.Fruit, int, error) {
+func escapeLike(s string) string {
+	s = strings.ReplaceAll(s, `\`, `\\`)
+	s = strings.ReplaceAll(s, `%`, `\%`)
+	s = strings.ReplaceAll(s, `_`, `\_`)
+	return s
+}
+
+func (r *FruitRepo) List(ctx context.Context, limit, offset int, name string, types []string) ([]domain.Fruit, int, error) {
+	escaped := escapeLike(name)
+	countRow := r.pool.QueryRow(ctx,
+		`SELECT COUNT(DISTINCT f.id)
+		 FROM fruits f
+		 LEFT JOIN fruit_synonyms fs ON fs.fruit_id = f.id
+		 WHERE ($1 = '' OR f.name ILIKE '%' || $1 || '%' ESCAPE '\' OR fs.synonym ILIKE '%' || $1 || '%' ESCAPE '\')
+		   AND ($2::fruit_type[] IS NULL OR f.fruit_type = ANY($2::fruit_type[]))`,
+		escaped, types)
 	var total int
-	if err := r.pool.QueryRow(ctx, "SELECT COUNT(*) FROM fruits").Scan(&total); err != nil {
+	if err := countRow.Scan(&total); err != nil {
 		return nil, 0, err
 	}
 
 	rows, err := r.pool.Query(ctx,
-		`SELECT id, name, osdb_number, comment, fruit_type, created_at, updated_at
+		`SELECT DISTINCT f.id, f.name, f.osdb_number, f.comment, f.fruit_type, f.created_at, f.updated_at
-		 FROM fruits ORDER BY id LIMIT $1 OFFSET $2`, limit, offset)
+		 FROM fruits f
+		 LEFT JOIN fruit_synonyms fs ON fs.fruit_id = f.id
+		 WHERE ($1 = '' OR f.name ILIKE '%' || $1 || '%' ESCAPE '\' OR fs.synonym ILIKE '%' || $1 || '%' ESCAPE '\')
+		   AND ($2::fruit_type[] IS NULL OR f.fruit_type = ANY($2::fruit_type[]))
+		 ORDER BY f.name, f.id LIMIT $3 OFFSET $4`,
+		escaped, types, limit, offset)
 	if err != nil {
 		return nil, 0, err
 	}

backend/internal/repository/fruit_repo_integration_test.go

@@ -97,7 +97,7 @@ func TestFruitRepoIntegration(t *testing.T) {
 	}
 
 	// List
-	fruits, total, err := repo.List(ctx, 50, 0)
+	fruits, total, err := repo.List(ctx, 50, 0, "", nil)
 	if err != nil {
 		t.Fatalf("List: %v", err)
 	}

frontend/src/api/fruits.test.ts

@@ -19,18 +19,17 @@ type MockResponse = {
 const fetchMock = vi.fn((url: string, init?: RequestInit): Promise<MockResponse> => {
   const method = init?.method ?? 'GET'
 
-  if (url === '/api/v1/fruits?limit=50&offset=0' && method === 'GET') {
+  if (url.startsWith('/api/v1/fruits?') && method === 'GET') {
+    const params = new URLSearchParams(url.split('?')[1])
     return Promise.resolve({
       ok: true,
       status: 200,
-      json: async () => ({ items: [], total: 0, limit: 50, offset: 0 }),
+      json: async () => ({
-    })
+        items: [],
-  }
+        total: 0,
-  if (url === '/api/v1/fruits?limit=10&offset=20' && method === 'GET') {
+        limit: Number(params.get('limit') ?? 50),
-    return Promise.resolve({
+        offset: Number(params.get('offset') ?? 0),
-      ok: true,
+      }),
-      status: 200,
-      json: async () => ({ items: [], total: 0, limit: 10, offset: 20 }),
     })
   }
   if (url === '/api/v1/fruits/1' && method === 'GET') {
@@ -99,9 +98,19 @@ describe('listFruits', () => {
   })
 
   it('passes custom limit and offset', async () => {
-    const result = await listFruits(10, 20)
+    const result = await listFruits({ limit: 10, offset: 20 })
     expect(result.offset).toBe(20)
   })
+
+  it('appends name param when provided', async () => {
+    await listFruits({ name: 'Boskop' })
+    expect(fetchMock).toHaveBeenLastCalledWith(expect.stringContaining('name=Boskop'))
+  })
+
+  it('appends type param when provided', async () => {
+    await listFruits({ type: 'Apfelsorten' })
+    expect(fetchMock).toHaveBeenLastCalledWith(expect.stringContaining('type=Apfelsorten'))
+  })
 })
 
 describe('getFruit', () => {

frontend/src/api/fruits.ts

@@ -77,8 +77,17 @@ async function checkOk(res: Response): Promise<Response> {
   return res
 }
 
-export async function listFruits(limit = 50, offset = 0): Promise<FruitListResponse> {
+export async function listFruits(params?: {
-  const res = await fetch(`/api/v1/fruits?limit=${limit}&offset=${offset}`)
+  limit?: number
+  offset?: number
+  name?: string
+  type?: string
+}): Promise<FruitListResponse> {
+  const { limit = 50, offset = 0, name, type } = params ?? {}
+  let url = `/api/v1/fruits?limit=${limit}&offset=${offset}`
+  if (name) url += `&name=${encodeURIComponent(name)}`
+  if (type) url += `&type=${encodeURIComponent(type)}`
+  const res = await fetch(url)
   return (await checkOk(res)).json()
 }
 

frontend/src/stores/fruitStore.test.ts

@@ -115,4 +115,15 @@ describe('fruitStore', () => {
     expect(store.fruits.find((f) => f.id === 1)).toBeUndefined()
     expect(store.total).toBe(1)
   })
+
+  it('setSearch updates search state and resets offset', async () => {
+    const store = useFruitStore()
+    await store.fetchFruits()
+    store.offset = 50
+    await store.setSearch('Boskop', 'Apfelsorten')
+    expect(store.searchName).toBe('Boskop')
+    expect(store.searchType).toBe('Apfelsorten')
+    expect(store.offset).toBe(0)
+    expect(fetchMock).toHaveBeenLastCalledWith(expect.stringContaining('name=Boskop'))
+  })
 })

frontend/src/stores/fruitStore.ts

@@ -18,12 +18,19 @@ export const useFruitStore = defineStore('fruit', () => {
   const current = ref<Fruit | null>(null)
   const loading = ref(false)
   const error = ref<string | null>(null)
+  const searchName = ref('')
+  const searchType = ref('')
 
   async function fetchFruits(lim = limit.value, off = offset.value) {
     loading.value = true
     error.value = null
     try {
-      const resp = await listFruits(lim, off)
+      const resp = await listFruits({
+        limit: lim,
+        offset: off,
+        name: searchName.value || undefined,
+        type: searchType.value || undefined,
+      })
       fruits.value = resp.items
       total.value = resp.total
       limit.value = resp.limit
@@ -35,6 +42,13 @@ export const useFruitStore = defineStore('fruit', () => {
     }
   }
 
+  async function setSearch(name: string, type: string) {
+    searchName.value = name
+    searchType.value = type
+    offset.value = 0
+    await fetchFruits(limit.value, 0)
+  }
+
   async function fetchFruit(id: number) {
     loading.value = true
     error.value = null
@@ -71,5 +85,5 @@ export const useFruitStore = defineStore('fruit', () => {
     if (current.value?.id === id) current.value = null
   }
 
-  return { fruits, total, limit, offset, current, loading, error, fetchFruits, fetchFruit, create, update, remove }
+  return { fruits, total, limit, offset, current, loading, error, searchName, searchType, fetchFruits, fetchFruit, create, update, remove, setSearch }
 })

frontend/src/views/FruitList.vue

@@ -1,14 +1,46 @@
 <script setup lang="ts">
-import { onMounted, computed } from 'vue'
+import { onMounted, onUnmounted, computed, ref } from 'vue'
 import { RouterLink } from 'vue-router'
 import { useFruitStore } from '../stores/fruitStore'
 
+const SEARCH_TYPES = [
+  { label: '(Alle)', value: '' },
+  { label: 'Apfelsorten', value: 'Apfelsorten' },
+  { label: 'Birnensorten', value: 'Birnensorten' },
+  { label: 'Quittensorten', value: 'Quittensorten' },
+  { label: 'Birnen- und Quittensorten', value: 'Birnen- und Quittensorten' },
+  { label: 'Aprikosen', value: 'Aprikosen' },
+  { label: 'Pfirsiche', value: 'Pfirsiche' },
+  { label: 'Aprikosen und Pfirsiche', value: 'Aprikosen und Pfirsiche' },
+  { label: 'Mirabellen', value: 'Mirabellen' },
+  { label: 'Renekloden', value: 'Renekloden' },
+  { label: 'Mirabellen und Reineclauden', value: 'Mirabellen und Reineclauden' },
+  { label: 'Pflaumen', value: 'Pflaumen' },
+  { label: 'Zwetschen', value: 'Zwetschen' },
+  { label: 'Pflaumen und Zwetschen', value: 'Pflaumen und Zwetschen' },
+  { label: 'Sauerkirschen', value: 'Sauerkirschen' },
+  { label: 'Süßkirschen', value: 'Süßkirschen' },
+  { label: 'Brombeeren', value: 'Brombeeren' },
+  { label: 'Erdbeeren', value: 'Erdbeeren' },
+  { label: 'Himbeeren', value: 'Himbeeren' },
+  { label: 'Johannisbeeren', value: 'Johannisbeeren' },
+  { label: 'Stachelbeeren', value: 'Stachelbeeren' },
+  { label: 'Wein', value: 'Wein' },
+]
+
 const store = useFruitStore()
+const nameInput = ref(store.searchName)
+const typeSelect = ref(store.searchType)
+let debounceTimer: ReturnType<typeof setTimeout> | null = null
 
 onMounted(async () => {
   await store.fetchFruits()
 })
 
+onUnmounted(() => {
+  if (debounceTimer) clearTimeout(debounceTimer)
+})
+
 const hasPrev = computed(() => store.offset > 0)
 const hasNext = computed(() => store.offset + store.limit < store.total)
 
@@ -19,6 +51,23 @@ async function prev() {
 async function next() {
   await store.fetchFruits(store.limit, store.offset + store.limit)
 }
+
+function onNameInput() {
+  if (debounceTimer) clearTimeout(debounceTimer)
+  debounceTimer = setTimeout(() => {
+    store.setSearch(nameInput.value, typeSelect.value)
+  }, 300)
+}
+
+function onNameEnter() {
+  if (debounceTimer) clearTimeout(debounceTimer)
+  store.setSearch(nameInput.value, typeSelect.value)
+}
+
+function onTypeChange() {
+  if (debounceTimer) clearTimeout(debounceTimer)
+  store.setSearch(nameInput.value, typeSelect.value)
+}
 </script>
 
 <template>
@@ -33,14 +82,24 @@ async function next() {
       </RouterLink>
     </div>
 
-    <!-- Search placeholder (story #06) -->
+    <div class="mb-4 flex gap-3">
-    <div class="mb-4">
       <input
+        v-model="nameInput"
         type="text"
-        placeholder="Suche... (kommt in Story #06)"
+        placeholder="Name oder Synonym suchen..."
-        disabled
+        class="flex-1 border border-gray-300 rounded px-3 py-2 focus:outline-none focus:ring-2 focus:ring-green-500"
-        class="w-full border border-gray-300 rounded px-3 py-2 bg-gray-100 text-gray-400 cursor-not-allowed"
+        @input="onNameInput"
+        @keydown.enter="onNameEnter"
       />
+      <select
+        v-model="typeSelect"
+        class="border border-gray-300 rounded px-3 py-2 focus:outline-none focus:ring-2 focus:ring-green-500"
+        @change="onTypeChange"
+      >
+        <option v-for="opt in SEARCH_TYPES" :key="opt.value" :value="opt.value">
+          {{ opt.label }}
+        </option>
+      </select>
     </div>
 
     <div v-if="store.loading" class="text-gray-500">Wird geladen...</div>

scripts/README_import.md

@@ -0,0 +1,37 @@
+# Import Scripts
+
+## Prerequisites
+
+- Postgres running and migrated (`make migrate-up`)
+- `DATABASE_URL` set, e.g.:
+  ```
+  export DATABASE_URL=postgres://user:pass@localhost:5432/osdb
+  ```
+- `03-data/` directory present at repo root
+- `psycopg2` installed (`pip install psycopg2-binary`)
+
+---
+
+## Run Order
+
+Scripts must be run in order — publications import depends on fruits being present.
+
+### 1. Import fruits
+
+```bash
+DATABASE_URL=... python3 scripts/import_fruits.py
+```
+
+Imports fruits, synonyms, and fruit images from `03-data/obstsorten.xml`.
+Idempotent: upserts fruits by `osdb_number`.
+
+### 2. Import publications
+
+```bash
+DATABASE_URL=... python3 scripts/import_publications.py
+```
+
+Imports publications from `03-data/osws.xml` (only `<obj>` elements with `<osw>=1`).
+For each publication: upserts the row, loads cover image, links fruits found by
+filesystem scan of `03-data/osdb/{pubId}/`, and imports PDFs and fruit images.
+Idempotent: clears and re-inserts linked data on re-run.

scripts/import_publications.py

@@ -0,0 +1,215 @@
+#!/usr/bin/env python3
+"""
+Import publications from 03-data/osws.xml into the OSDB database.
+
+Usage:
+    DATABASE_URL=postgres://... python3 scripts/import_publications.py
+
+Idempotent: upserts publication by osdb_pub_id; deletes and re-inserts
+linked fruits, descriptions, and fruit images per publication on re-run.
+"""
+
+import os
+import re
+import sys
+import xml.etree.ElementTree as ET
+from pathlib import Path
+
+import psycopg2
+
+DATA_DIR = os.path.join(os.path.dirname(os.path.abspath(__file__)), "..", "03-data")
+
+
+def parse_publications(xml_path: str) -> list:
+    """
+    Parse osws.xml and return list of publication dicts for <obj> elements with <osw>=1.
+    Each dict: id, title, author (None if "."), img_path (relative to data root, or None).
+    """
+    tree = ET.parse(xml_path)
+    root = tree.getroot()
+    pubs = []
+    for obj in root.findall("obj"):
+        if (obj.findtext("osw") or "").strip() != "1":
+            continue
+        author_raw = (obj.findtext("author") or "").strip()
+        img = obj.findtext("img")
+        pubs.append({
+            "id": (obj.findtext("id") or "").strip(),
+            "title": (obj.findtext("name") or "").strip(),
+            "author": author_raw if author_raw and author_raw != "." else None,
+            "img_path": img.strip() if img else None,
+        })
+    return pubs
+
+
+def scan_pub_dir(pub_dir: str, pub_id: str) -> dict:
+    """
+    Scan pub_dir for files matching {osdb_number}_{pub_id}_s0.jpg and {osdb_number}_{pub_id}.pdf.
+    Returns dict mapping osdb_number → {img_path?: str, pdf_path?: str}.
+    Ignores _tn.jpg thumbnails and any other files.
+    """
+    d = Path(pub_dir)
+    if not d.is_dir():
+        return {}
+
+    result = {}
+    img_re = re.compile(rf"^(.+)_{re.escape(pub_id)}_s0\.jpg$")
+    pdf_re = re.compile(rf"^(.+)_{re.escape(pub_id)}\.pdf$")
+
+    for f in d.iterdir():
+        m = img_re.match(f.name)
+        if m:
+            result.setdefault(m.group(1), {})["img_path"] = str(f)
+            continue
+        m = pdf_re.match(f.name)
+        if m:
+            result.setdefault(m.group(1), {})["pdf_path"] = str(f)
+
+    return result
+
+
+def import_publications(conn, data_dir: str = None) -> dict:
+    """
+    Import all publications from osws.xml into the database.
+    Returns counts: publications, covers, fruits_linked, pdfs, images, skipped_fruits.
+    """
+    if data_dir is None:
+        data_dir = DATA_DIR
+
+    data_path = Path(data_dir)
+    pubs = parse_publications(str(data_path / "osws.xml"))
+
+    counts = {
+        "publications": 0,
+        "covers": 0,
+        "fruits_linked": 0,
+        "pdfs": 0,
+        "images": 0,
+        "skipped_fruits": 0,
+    }
+
+    with conn.cursor() as cur:
+        for pub in pubs:
+            pub_id = pub["id"]
+
+            # Load cover image bytes (skip silently if file missing or unreadable)
+            cover_data = None
+            if pub["img_path"]:
+                cover_path = data_path / pub["img_path"]
+                try:
+                    cover_data = cover_path.read_bytes()
+                    counts["covers"] += 1
+                except OSError:
+                    pass
+
+            # Upsert publication row
+            cur.execute(
+                """
+                INSERT INTO publications (title, author, osdb_pub_id, image_data, updated_at)
+                VALUES (%s, %s, %s, %s, NOW())
+                ON CONFLICT (osdb_pub_id) DO UPDATE
+                  SET title      = EXCLUDED.title,
+                      author     = EXCLUDED.author,
+                      image_data = EXCLUDED.image_data,
+                      updated_at = NOW()
+                RETURNING id
+                """,
+                (
+                    pub["title"],
+                    pub["author"],
+                    pub_id,
+                    psycopg2.Binary(cover_data) if cover_data else None,
+                ),
+            )
+            db_pub_id = cur.fetchone()[0]
+            counts["publications"] += 1
+
+            # Scan filesystem for linked fruits
+            pub_dir = str(data_path / "osdb" / pub_id)
+            fruit_files = scan_pub_dir(pub_dir, pub_id)
+
+            # Resolve osdb_numbers → DB fruit IDs; warn and skip unknowns
+            linked = {}
+            for osdb_number, files in fruit_files.items():
+                cur.execute("SELECT id FROM fruits WHERE osdb_number = %s", (osdb_number,))
+                row = cur.fetchone()
+                if row is None:
+                    print(
+                        f"  WARNING: osdb_number not in fruits table, skipping: {osdb_number}",
+                        file=sys.stderr,
+                    )
+                    counts["skipped_fruits"] += 1
+                    continue
+                linked[osdb_number] = {"db_id": row[0], **files}
+
+            # Idempotent: clear previous linked data for this publication
+            cur.execute("DELETE FROM publication_fruit_images WHERE publication_id = %s", (db_pub_id,))
+            cur.execute("DELETE FROM publication_descriptions  WHERE publication_id = %s", (db_pub_id,))
+            cur.execute("DELETE FROM publication_fruits        WHERE publication_id = %s", (db_pub_id,))
+
+            # Re-insert linked fruits, PDFs, and images
+            for osdb_number, info in linked.items():
+                fruit_db_id = info["db_id"]
+
+                cur.execute(
+                    "INSERT INTO publication_fruits (publication_id, fruit_id) VALUES (%s, %s)",
+                    (db_pub_id, fruit_db_id),
+                )
+                counts["fruits_linked"] += 1
+
+                if "pdf_path" in info:
+                    pdf_data = Path(info["pdf_path"]).read_bytes()
+                    cur.execute(
+                        """
+                        INSERT INTO publication_descriptions (publication_id, fruit_id, pdf_data)
+                        VALUES (%s, %s, %s)
+                        """,
+                        (db_pub_id, fruit_db_id, psycopg2.Binary(pdf_data)),
+                    )
+                    counts["pdfs"] += 1
+
+                if "img_path" in info:
+                    img_data = Path(info["img_path"]).read_bytes()
+                    filename = Path(info["img_path"]).name
+                    cur.execute(
+                        """
+                        INSERT INTO publication_fruit_images
+                          (publication_id, fruit_id, filename, data, image_type)
+                        VALUES (%s, %s, %s, %s, 'fruit')
+                        """,
+                        (db_pub_id, fruit_db_id, filename, psycopg2.Binary(img_data)),
+                    )
+                    counts["images"] += 1
+
+    conn.commit()
+    return counts
+
+
+def main():
+    database_url = os.environ.get("DATABASE_URL")
+    if not database_url:
+        print("Error: DATABASE_URL environment variable not set", file=sys.stderr)
+        sys.exit(1)
+
+    print("Connecting to database...")
+    conn = psycopg2.connect(database_url)
+    try:
+        print("Importing publications...")
+        counts = import_publications(conn)
+        print("\nDone.")
+        print(f"  Publications upserted: {counts['publications']}")
+        print(f"  Cover images loaded:   {counts['covers']}")
+        print(f"  Fruits linked:         {counts['fruits_linked']}")
+        print(f"  PDFs imported:         {counts['pdfs']}")
+        print(f"  Fruit images imported: {counts['images']}")
+        print(f"  Fruits skipped:        {counts['skipped_fruits']}")
+    except Exception as e:
+        print(f"Error: {e}", file=sys.stderr)
+        conn.rollback()
+        sys.exit(1)
+    finally:
+        conn.close()
+
+
+if __name__ == "__main__":
+    main()

scripts/import_publications_test.py

@@ -0,0 +1,236 @@
+import os
+import tempfile
+import unittest
+from pathlib import Path
+from unittest.mock import MagicMock, call, patch
+
+from import_publications import parse_publications, scan_pub_dir, import_publications
+
+MINIMAL_XML_STR = """<?xml version="1.0" encoding="iso-8859-1" ?>
+<root>
+  <obj>
+    <id>skip_me</id>
+    <name>Not a publication</name>
+    <author>Nobody</author>
+    <osw>0</osw>
+  </obj>
+  <obj>
+    <id>ber</id>
+    <name>Bernisches Stammregister</name>
+    <author>.</author>
+    <img>osdb/ber/ber_s0.jpg</img>
+    <osw>1</osw>
+  </obj>
+  <obj>
+    <id>cal</id>
+    <name>Obst- und Beeren</name>
+    <author>Calwer</author>
+    <img>osdb/cal/cal_s0.jpg</img>
+    <osw>1</osw>
+  </obj>
+  <obj>
+    <id>noc</id>
+    <name>No Cover</name>
+    <author>Someone</author>
+    <osw>1</osw>
+  </obj>
+</root>
+"""
+MINIMAL_XML = MINIMAL_XML_STR.encode("iso-8859-1")
+
+
+def _write_xml(tmp_dir, content=MINIMAL_XML):
+    p = Path(tmp_dir) / "osws.xml"
+    p.write_bytes(content)
+    return str(p)
+
+
+class TestParsePublications(unittest.TestCase):
+    def setUp(self):
+        self.tmp = tempfile.mkdtemp()
+        self.xml_path = _write_xml(self.tmp)
+
+    def test_excludes_osw_not_1(self):
+        pubs = parse_publications(self.xml_path)
+        ids = [p["id"] for p in pubs]
+        self.assertNotIn("skip_me", ids)
+
+    def test_includes_osw_1(self):
+        pubs = parse_publications(self.xml_path)
+        ids = [p["id"] for p in pubs]
+        self.assertIn("ber", ids)
+        self.assertIn("cal", ids)
+
+    def test_author_dot_becomes_none(self):
+        pubs = parse_publications(self.xml_path)
+        ber = next(p for p in pubs if p["id"] == "ber")
+        self.assertIsNone(ber["author"])
+
+    def test_author_string_preserved(self):
+        pubs = parse_publications(self.xml_path)
+        cal = next(p for p in pubs if p["id"] == "cal")
+        self.assertEqual(cal["author"], "Calwer")
+
+    def test_img_path_present(self):
+        pubs = parse_publications(self.xml_path)
+        cal = next(p for p in pubs if p["id"] == "cal")
+        self.assertEqual(cal["img_path"], "osdb/cal/cal_s0.jpg")
+
+    def test_img_path_absent_is_none(self):
+        pubs = parse_publications(self.xml_path)
+        noc = next(p for p in pubs if p["id"] == "noc")
+        self.assertIsNone(noc["img_path"])
+
+    def test_title_mapped(self):
+        pubs = parse_publications(self.xml_path)
+        cal = next(p for p in pubs if p["id"] == "cal")
+        self.assertEqual(cal["title"], "Obst- und Beeren")
+
+
+class TestScanPubDir(unittest.TestCase):
+    def setUp(self):
+        self.tmp = tempfile.mkdtemp()
+
+    def test_returns_empty_for_missing_directory(self):
+        result = scan_pub_dir(os.path.join(self.tmp, "nonexistent"), "xyz")
+        self.assertEqual(result, {})
+
+    def test_extracts_osdb_number_from_image(self):
+        d = Path(self.tmp)
+        (d / "apfel_ber_s0.jpg").write_bytes(b"img")
+        result = scan_pub_dir(str(d), "ber")
+        self.assertIn("apfel", result)
+        self.assertEqual(result["apfel"]["img_path"], str(d / "apfel_ber_s0.jpg"))
+
+    def test_extracts_osdb_number_from_pdf(self):
+        d = Path(self.tmp)
+        (d / "birne_cal.pdf").write_bytes(b"pdf")
+        result = scan_pub_dir(str(d), "cal")
+        self.assertIn("birne", result)
+        self.assertEqual(result["birne"]["pdf_path"], str(d / "birne_cal.pdf"))
+
+    def test_combines_img_and_pdf_for_same_fruit(self):
+        d = Path(self.tmp)
+        (d / "apfel_ber_s0.jpg").write_bytes(b"img")
+        (d / "apfel_ber.pdf").write_bytes(b"pdf")
+        result = scan_pub_dir(str(d), "ber")
+        self.assertIn("img_path", result["apfel"])
+        self.assertIn("pdf_path", result["apfel"])
+
+    def test_ignores_thumbnail_files(self):
+        d = Path(self.tmp)
+        (d / "apfel_ber_tn.jpg").write_bytes(b"tn")
+        result = scan_pub_dir(str(d), "ber")
+        self.assertEqual(result, {})
+
+    def test_multi_segment_osdb_number(self):
+        d = Path(self.tmp)
+        (d / "berner_grauechapfel_ber_s0.jpg").write_bytes(b"img")
+        result = scan_pub_dir(str(d), "ber")
+        self.assertIn("berner_grauechapfel", result)
+
+
+class TestImportPublications(unittest.TestCase):
+    def _make_data_dir(self, pubs):
+        """Build a temp data dir with osws.xml and osdb/<id>/ dirs."""
+        tmp = tempfile.mkdtemp()
+        lines = ['<?xml version="1.0" encoding="iso-8859-1" ?>', "<root>"]
+        for p in pubs:
+            lines.append("  <obj>")
+            lines.append(f"    <id>{p['id']}</id>")
+            lines.append(f"    <name>{p['name']}</name>")
+            author = p.get("author", "Test")
+            lines.append(f"    <author>{author}</author>")
+            if p.get("img"):
+                lines.append(f"    <img>{p['img']}</img>")
+            lines.append("    <osw>1</osw>")
+            lines.append("  </obj>")
+        lines.append("</root>")
+        Path(tmp, "osws.xml").write_bytes("\n".join(lines).encode("iso-8859-1"))
+
+        for p in pubs:
+            pub_dir = Path(tmp, "osdb", p["id"])
+            pub_dir.mkdir(parents=True, exist_ok=True)
+            for fruit in p.get("fruits", []):
+                if fruit.get("img"):
+                    (pub_dir / f"{fruit['osdb_number']}_{p['id']}_s0.jpg").write_bytes(b"\x89PNG")
+                if fruit.get("pdf"):
+                    (pub_dir / f"{fruit['osdb_number']}_{p['id']}.pdf").write_bytes(b"%PDF")
+            if p.get("cover_bytes"):
+                cover_rel = p["img"]
+                cover_path = Path(tmp, cover_rel)
+                cover_path.parent.mkdir(parents=True, exist_ok=True)
+                cover_path.write_bytes(p["cover_bytes"])
+        return tmp
+
+    def _make_conn(self, pub_db_id=42, fruit_db_ids=None):
+        """Return a mock psycopg2 connection."""
+        conn = MagicMock()
+        cur = MagicMock()
+        conn.cursor.return_value.__enter__.return_value = cur
+
+        # fetchone side_effect: first call = pub id; subsequent = fruit lookups.
+        # None in fruit_db_ids → bare None (simulates no row found); int → (int,) tuple.
+        fruit_db_ids = fruit_db_ids or []
+        side_effects = [(pub_db_id,)] + [((fid,) if fid is not None else None) for fid in fruit_db_ids]
+        cur.fetchone.side_effect = side_effects
+        return conn, cur
+
+    def test_missing_cover_image_does_not_abort_import(self):
+        data_dir = self._make_data_dir([
+            {"id": "ber", "name": "Test", "img": "osdb/ber/ber_s0.jpg"},
+        ])
+        # cover file NOT created → missing
+        conn, cur = self._make_conn(pub_db_id=1)
+        counts = import_publications(conn, data_dir)
+        self.assertEqual(counts["publications"], 1)
+        self.assertEqual(counts["covers"], 0)
+
+    def test_fruit_not_in_db_is_skipped_with_warning(self):
+        data_dir = self._make_data_dir([
+            {"id": "cal", "name": "Cal", "fruits": [
+                {"osdb_number": "unknown_fruit", "img": True},
+            ]},
+        ])
+        conn, cur = self._make_conn(pub_db_id=5, fruit_db_ids=[None])
+        counts = import_publications(conn, data_dir)
+        self.assertEqual(counts["skipped_fruits"], 1)
+        self.assertEqual(counts["fruits_linked"], 0)
+
+    def test_happy_path_upserts_pub_links_fruits_imports_files(self):
+        data_dir = self._make_data_dir([
+            {
+                "id": "pom",
+                "name": "Pomologie",
+                "author": "Diel",
+                "img": "osdb/pom/pom_s0.jpg",
+                "cover_bytes": b"\x89PNG",
+                "fruits": [
+                    {"osdb_number": "apfelsorte_alpha", "img": True, "pdf": True},
+                ],
+            }
+        ])
+        conn, cur = self._make_conn(pub_db_id=7, fruit_db_ids=[99])
+        counts = import_publications(conn, data_dir)
+        self.assertEqual(counts["publications"], 1)
+        self.assertEqual(counts["covers"], 1)
+        self.assertEqual(counts["fruits_linked"], 1)
+        self.assertEqual(counts["pdfs"], 1)
+        self.assertEqual(counts["images"], 1)
+        self.assertEqual(counts["skipped_fruits"], 0)
+
+    def test_osw_not_1_never_imported(self):
+        tmp = tempfile.mkdtemp()
+        xml = b"""<?xml version="1.0" encoding="iso-8859-1"?>
+<root>
+  <obj><id>skip</id><name>Skip</name><author>X</author><osw>0</osw></obj>
+</root>"""
+        Path(tmp, "osws.xml").write_bytes(xml)
+        conn, cur = self._make_conn()
+        counts = import_publications(conn, tmp)
+        self.assertEqual(counts["publications"], 0)
+        cur.execute.assert_not_called()
+
+
+if __name__ == "__main__":
+    unittest.main()