feat: extract actors, directors, screenwriters from Wikipedia API

- Extract directors from infobox 'Directed by' field/list
- Extract screenwriters from infobox 'Screenplay by' list
- Extract actors from Cast section list (first link = person name)
- Upsert into people table, link via who table (profession: actor=1, director=2, screenwriter=3)
- Track processed entries with has_people flag column
- Consumer inserts people and marks has_people=1 on success

2 files changed, 184 insertions, 7 deletions

diff --git a/src/wikiarticle.go b/src/wikiarticle.go
index 5b891b6..cef9f02 100644
--- a/src/wikiarticle.go
+++ b/src/wikiarticle.go
@@ -14,15 +14,22 @@ import (
 
 var wikiArticleClient = &http.Client{Timeout: 60 * time.Second}
 
+// wikiPerson carries a person extracted from a wiki article.
+type wikiPerson struct {
+	Name       string
+	Profession int // 1=actor, 2=director, 3=screenwriter
+}
+
 // wikiArticleEntry holds extracted fields from a wiki article API response.
 type wikiArticleEntry struct {
-	Description string
-	Year        int
-	PosterURL   string
-	Synopsis    string
-	License     string
-	LicenseURL  string
+	Description  string
+	Year         int
+	PosterURL    string
+	Synopsis     string
+	License      string
+	LicenseURL   string
 	NumAccolades int
+	People       []wikiPerson
 }
 
 func (a *App) queryWikiArticle(name string) (wikiArticleEntry, int, error) {
@@ -84,6 +91,9 @@ func (a *App) queryWikiArticle(name string) (wikiArticleEntry, int, error) {
 	// year, poster_url from infobox
 	entry.Year, entry.PosterURL = extractInfoboxData(article)
 
+	// people from infobox and sections
+	entry.People = extractPeople(article)
+
 	// license
 	if licList, ok := article["license"]; ok {
 		if arr, ok := licList.([]interface{}); ok && len(arr) > 0 {
@@ -204,3 +214,131 @@ func extractAccolades(article map[string]interface{}) int {
 	}
 	return total
 }
+
+// extractPeople extracts actors, directors, and screenwriters from the article.
+func extractPeople(article map[string]interface{}) []wikiPerson {
+	var people []wikiPerson
+
+	// Directors and screenwriters from infobox
+	ibParts := getInfoboxParts(article)
+	for _, p := range ibParts {
+		fp, ok := p.(map[string]interface{})
+		if !ok {
+			continue
+		}
+		name := fp["name"]
+		if name == "Directed by" {
+			people = append(people, extractPersonFromField(fp, 2)...)
+		}
+		if name == "Screenplay by" {
+			people = append(people, extractPersonFromField(fp, 3)...)
+		}
+	}
+
+	// Actors from Cast section
+	for _, sec := range getSections(article) {
+		s, ok := sec.(map[string]interface{})
+		if !ok || s["name"] != "Cast" {
+			continue
+		}
+		for _, part := range s["has_parts"].([]interface{}) {
+			p, ok := part.(map[string]interface{})
+			if !ok || p["type"] != "list" {
+				continue
+			}
+			for _, item := range p["has_parts"].([]interface{}) {
+				item, ok := item.(map[string]interface{})
+				if !ok {
+					continue
+				}
+				if link, ok := getFirstPersonLink(item); ok {
+					people = append(people, wikiPerson{Name: link, Profession: 1})
+				}
+			}
+			break // only first list in Cast section
+		}
+		break // only Cast section
+	}
+
+	return people
+}
+
+// getInfoboxParts returns the inner parts of the infobox section.
+func getInfoboxParts(article map[string]interface{}) []interface{} {
+	infoboxes, ok := article["infoboxes"].([]interface{})
+	if !ok || len(infoboxes) == 0 {
+		return nil
+	}
+	ib, ok := infoboxes[0].(map[string]interface{})
+	if !ok {
+		return nil
+	}
+	parts, ok := ib["has_parts"].([]interface{})
+	if !ok || len(parts) == 0 {
+		return nil
+	}
+	section, ok := parts[0].(map[string]interface{})
+	if !ok {
+		return nil
+	}
+	return section["has_parts"].([]interface{})
+}
+
+// getSections returns sections from the article.
+func getSections(article map[string]interface{}) []interface{} {
+	sections, ok := article["sections"].([]interface{})
+	if !ok {
+		return nil
+	}
+	return sections
+}
+
+// extractPersonFromField extracts a person name from a field or list.
+func extractPersonFromField(fp map[string]interface{}, profession int) []wikiPerson {
+	var people []wikiPerson
+	switch fp["type"] {
+	case "field":
+		if val, ok := fp["value"]; ok && val != nil {
+			name := fmt.Sprintf("%v", val)
+			if name != "" {
+				people = append(people, wikiPerson{Name: name, Profession: profession})
+			}
+		}
+	case "list":
+		for _, item := range fp["has_parts"].([]interface{}) {
+			item, ok := item.(map[string]interface{})
+			if !ok {
+				continue
+			}
+			// Prefer link text over value
+			if link, ok := getFirstPersonLink(item); ok {
+				people = append(people, wikiPerson{Name: link, Profession: profession})
+			} else if val := item["value"]; val != nil {
+				name := fmt.Sprintf("%v", val)
+				if name != "" {
+					people = append(people, wikiPerson{Name: name, Profession: profession})
+				}
+			}
+		}
+	}
+	return people
+}
+
+// getFirstPersonLink extracts the first link text from a list item.
+func getFirstPersonLink(item map[string]interface{}) (string, bool) {
+	links, ok := item["links"].([]interface{})
+	if !ok || len(links) == 0 {
+		return "", false
+	}
+	link, ok := links[0].(map[string]interface{})
+	if !ok {
+		return "", false
+	}
+	if text, ok := link["text"]; ok && text != nil {
+		s := fmt.Sprintf("%v", text)
+		if s != "" {
+			return s, true
+		}
+	}
+	return "", false
+}
diff --git a/src/wikidata.go b/src/wikidata.go
index e6c2e0e..59c8188 100644
--- a/src/wikidata.go
+++ b/src/wikidata.go
@@ -126,7 +126,7 @@ func (a *App) getExistingWikiArticles() ([]existingWikiArticle, error) {
 		WHERE wiki_article IS NOT NULL
 		AND wiki_status_code != 404
 		AND (synopsis IS NULL OR description IS NULL OR year IS NULL
-			OR poster_url IS NULL OR license IS NULL OR license_url IS NULL OR num_accolades IS NULL)
+			OR poster_url IS NULL OR license IS NULL OR license_url IS NULL OR num_accolades IS NULL OR has_people = 0)
 	`)
 	if err != nil {
 		return nil, fmt.Errorf("query existing wiki articles: %w", err)
@@ -261,10 +261,49 @@ func (a *App) wikiDataConsumer(artCh <-chan wikiArticleFetch, done chan<- struct
 				entry.License, entry.LicenseURL, entry.NumAccolades, art.imdbID)
 			updated++
 		}
+
+		// Insert people (actors, directors, screenwriters)
+		if statusCode == 200 && len(entry.People) > 0 {
+			if err := a.insertWikiPeople(art.imdbID, entry.People); err != nil {
+				log.Printf("insert people error %s: %v", art.imdbID, err)
+			}
+			a.DB.Exec(`UPDATE imdb SET has_people = 1 WHERE imdb_id = ?`, art.imdbID)
+		}
 	}
 
 	log.Printf("fetchWikiArticlesData: %d updated, %d skipped (non-200)", updated, skipped)
 }
+
+// insertWikiPeople upserts people into people/who tables.
+func (a *App) insertWikiPeople(imdbID string, people []wikiPerson) error {
+	// Get the DB row id for this imdb_id
+	var dbID int
+	if err := a.DB.Get(&dbID, `SELECT id FROM imdb WHERE imdb_id = ?`, imdbID); err != nil {
+		return fmt.Errorf("lookup imdb id for %s: %w", imdbID, err)
+	}
+
+	for _, p := range people {
+		// Upsert person
+		var personID int
+		err := a.DB.Get(&personID, `SELECT id FROM people WHERE name = ?`, p.Name)
+		if err != nil {
+			// Insert new person
+			result, err := a.DB.Exec(`INSERT INTO people (name) VALUES (?)`, p.Name)
+			if err != nil {
+				continue // skip on conflict
+			}
+			if id, err := result.LastInsertId(); err == nil {
+				personID = int(id)
+			}
+		}
+
+		// Insert who relationship (unique on imdb_id, people_id, profession_id)
+		a.DB.Exec(`INSERT IGNORE INTO who (imdb_id, people_id, profession_id) VALUES (?, ?, ?)`,
+			dbID, personID, p.Profession)
+	}
+
+	return nil
+}
 // fetchWikiArticlesData fetches wiki article data from the custom server for all
 // entries that have a wiki_article but need data extraction. Callable independently.
 func (a *App) fetchWikiArticlesData() error {