feat: extensible search engine with person entities (FTS5/tsvector)

This commit is contained in:
2026-05-12 18:45:24 +02:00
parent 763d622601
commit c6770659c5
45 changed files with 2421 additions and 86 deletions

View File

@@ -17,6 +17,7 @@ mod import_profile;
mod import_session;
mod migrations;
mod models;
mod persons;
mod profile;
mod users;
@@ -28,6 +29,7 @@ use models::{
pub use image_ref::{SqliteImageRefAdapter, create_image_ref};
pub use import_profile::SqliteImportProfileRepository;
pub use import_session::SqliteImportSessionRepository;
pub use persons::{SqlitePersonAdapter, create_person_adapter};
pub use profile::SqliteMovieProfileRepository;
pub use users::SqliteUserRepository;

View File

@@ -0,0 +1,195 @@
use async_trait::async_trait;
use domain::{
errors::DomainError,
models::{CastCredit, CrewCredit, ExternalPersonId, Person, PersonCredits, PersonId},
ports::{PersonCommand, PersonQuery},
value_objects::MovieId,
};
use sqlx::SqlitePool;
use std::sync::Arc;
pub struct SqlitePersonAdapter {
pool: SqlitePool,
}
impl SqlitePersonAdapter {
pub fn new(pool: SqlitePool) -> Self {
Self { pool }
}
}
pub fn create_person_adapter(pool: SqlitePool) -> (Arc<dyn PersonCommand>, Arc<dyn PersonQuery>) {
let adapter = Arc::new(SqlitePersonAdapter::new(pool));
(Arc::clone(&adapter) as Arc<dyn PersonCommand>, adapter as Arc<dyn PersonQuery>)
}
fn map_err(e: sqlx::Error) -> DomainError {
DomainError::InfrastructureError(e.to_string())
}
#[async_trait]
impl PersonCommand for SqlitePersonAdapter {
async fn upsert_batch(&self, persons: &[Person]) -> Result<(), DomainError> {
for person in persons {
let tmdb_id = person.external_id().tmdb_id();
sqlx::query(
"INSERT INTO persons (id, external_id, tmdb_person_id, name, known_for_department, profile_path)
VALUES (?, ?, ?, ?, ?, ?)
ON CONFLICT(id) DO UPDATE SET
external_id = excluded.external_id,
tmdb_person_id = excluded.tmdb_person_id,
name = excluded.name,
known_for_department = excluded.known_for_department,
profile_path = excluded.profile_path",
)
.bind(person.id().value().to_string())
.bind(person.external_id().value())
.bind(tmdb_id)
.bind(person.name())
.bind(person.known_for_department())
.bind(person.profile_path())
.execute(&self.pool)
.await
.map_err(map_err)?;
}
Ok(())
}
}
#[async_trait]
impl PersonQuery for SqlitePersonAdapter {
async fn get_by_id(&self, id: &PersonId) -> Result<Option<Person>, DomainError> {
let row = sqlx::query_as::<_, PersonRow>(
"SELECT id, external_id, name, known_for_department, profile_path FROM persons WHERE id = ?",
)
.bind(id.value().to_string())
.fetch_optional(&self.pool)
.await
.map_err(map_err)?;
Ok(row.map(PersonRow::into_person))
}
async fn get_by_external_id(&self, id: &ExternalPersonId) -> Result<Option<Person>, DomainError> {
let row = sqlx::query_as::<_, PersonRow>(
"SELECT id, external_id, name, known_for_department, profile_path FROM persons WHERE external_id = ?",
)
.bind(id.value())
.fetch_optional(&self.pool)
.await
.map_err(map_err)?;
Ok(row.map(PersonRow::into_person))
}
async fn get_credits(&self, id: &PersonId) -> Result<PersonCredits, DomainError> {
let person = self.get_by_id(id).await?.ok_or_else(|| {
DomainError::NotFound(format!("Person {} not found", id.value()))
})?;
let tmdb_id: Option<i64> = sqlx::query_scalar(
"SELECT tmdb_person_id FROM persons WHERE id = ?",
)
.bind(id.value().to_string())
.fetch_optional(&self.pool)
.await
.map_err(map_err)?
.flatten();
let Some(tmdb_id) = tmdb_id else {
return Ok(PersonCredits { person, cast: vec![], crew: vec![] });
};
let cast = sqlx::query_as::<_, CastRow>(
"SELECT m.id, m.title, m.release_year, mc.character, m.poster_path
FROM movie_cast mc
JOIN movies m ON m.id = mc.movie_id
WHERE mc.tmdb_person_id = ?
ORDER BY mc.billing_order",
)
.bind(tmdb_id)
.fetch_all(&self.pool)
.await
.map_err(map_err)?
.into_iter()
.map(|r| CastCredit {
movie_id: MovieId::from_uuid(uuid::Uuid::parse_str(&r.id).unwrap_or_default()),
title: r.title,
release_year: r.release_year.map(|y| y as u16),
character: r.character,
poster_path: r.poster_path,
})
.collect();
let crew = sqlx::query_as::<_, CrewRow>(
"SELECT m.id, m.title, m.release_year, mc.job, mc.department, m.poster_path
FROM movie_crew mc
JOIN movies m ON m.id = mc.movie_id
WHERE mc.tmdb_person_id = ?
ORDER BY m.title",
)
.bind(tmdb_id)
.fetch_all(&self.pool)
.await
.map_err(map_err)?
.into_iter()
.map(|r| CrewCredit {
movie_id: MovieId::from_uuid(uuid::Uuid::parse_str(&r.id).unwrap_or_default()),
title: r.title,
release_year: r.release_year.map(|y| y as u16),
job: r.job,
department: r.department,
poster_path: r.poster_path,
})
.collect();
Ok(PersonCredits { person, cast, crew })
}
}
// ── Row types ────────────────────────────────────────────────────────────────
#[derive(sqlx::FromRow)]
struct PersonRow {
id: String,
external_id: String,
name: String,
known_for_department: Option<String>,
profile_path: Option<String>,
}
impl PersonRow {
fn into_person(self) -> Person {
let ext = ExternalPersonId::new(self.external_id);
Person::new(
PersonId::from_uuid(uuid::Uuid::parse_str(&self.id).unwrap_or_default()),
ext,
self.name,
self.known_for_department,
self.profile_path,
)
}
}
#[derive(sqlx::FromRow)]
struct CastRow {
id: String,
title: String,
release_year: Option<i64>,
character: String,
poster_path: Option<String>,
}
#[derive(sqlx::FromRow)]
struct CrewRow {
id: String,
title: String,
release_year: Option<i64>,
job: String,
department: String,
poster_path: Option<String>,
}
#[cfg(test)]
#[path = "tests/persons.rs"]
mod tests;

View File

@@ -0,0 +1,126 @@
use super::super::persons::SqlitePersonAdapter;
use domain::{
errors::DomainError,
models::{ExternalPersonId, Person, PersonId},
ports::{PersonCommand, PersonQuery},
};
use sqlx::SqlitePool;
async fn pool_with_schema() -> SqlitePool {
let pool = SqlitePool::connect("sqlite::memory:").await.unwrap();
sqlx::query(
"CREATE TABLE persons (
id TEXT PRIMARY KEY, external_id TEXT NOT NULL UNIQUE,
tmdb_person_id INTEGER UNIQUE, name TEXT NOT NULL,
known_for_department TEXT, profile_path TEXT
)",
)
.execute(&pool)
.await
.unwrap();
sqlx::query(
"CREATE TABLE movies (id TEXT PRIMARY KEY, title TEXT NOT NULL,
release_year INTEGER, director TEXT, poster_path TEXT,
external_metadata_id TEXT)",
)
.execute(&pool)
.await
.unwrap();
sqlx::query(
"CREATE TABLE movie_cast (movie_id TEXT, tmdb_person_id INTEGER,
name TEXT, character TEXT, billing_order INTEGER, profile_path TEXT,
PRIMARY KEY (movie_id, tmdb_person_id))",
)
.execute(&pool)
.await
.unwrap();
sqlx::query(
"CREATE TABLE movie_crew (movie_id TEXT, tmdb_person_id INTEGER,
name TEXT, job TEXT, department TEXT, profile_path TEXT,
PRIMARY KEY (movie_id, tmdb_person_id, job))",
)
.execute(&pool)
.await
.unwrap();
pool
}
fn make_person(tmdb_id: i64, name: &str, dept: Option<&str>) -> Person {
let ext = ExternalPersonId::new(format!("tmdb:{tmdb_id}"));
Person::new(
PersonId::from_external(&ext),
ext,
name.to_string(),
dept.map(str::to_string),
None,
)
}
#[tokio::test]
async fn upsert_batch_inserts_persons() {
let pool = pool_with_schema().await;
let adapter = SqlitePersonAdapter::new(pool.clone());
let persons = vec![make_person(1, "Alice", Some("Acting")), make_person(2, "Bob", Some("Directing"))];
adapter.upsert_batch(&persons).await.unwrap();
let count: (i64,) = sqlx::query_as("SELECT COUNT(*) FROM persons")
.fetch_one(&pool).await.unwrap();
assert_eq!(count.0, 2);
}
#[tokio::test]
async fn upsert_batch_is_idempotent() {
let pool = pool_with_schema().await;
let adapter = SqlitePersonAdapter::new(pool.clone());
let persons = vec![make_person(1, "Alice", Some("Acting"))];
adapter.upsert_batch(&persons).await.unwrap();
adapter.upsert_batch(&persons).await.unwrap();
let count: (i64,) = sqlx::query_as("SELECT COUNT(*) FROM persons")
.fetch_one(&pool).await.unwrap();
assert_eq!(count.0, 1);
}
#[tokio::test]
async fn get_by_id_returns_person() {
let pool = pool_with_schema().await;
let adapter = SqlitePersonAdapter::new(pool.clone());
let p = make_person(42, "Charlie", Some("Acting"));
adapter.upsert_batch(&[p.clone()]).await.unwrap();
let found = adapter.get_by_id(p.id()).await.unwrap().unwrap();
assert_eq!(found.name(), "Charlie");
assert_eq!(found.external_id().value(), "tmdb:42");
}
#[tokio::test]
async fn get_by_id_returns_none_for_unknown() {
let pool = pool_with_schema().await;
let adapter = SqlitePersonAdapter::new(pool);
let ext = ExternalPersonId::new("tmdb:999");
let id = PersonId::from_external(&ext);
assert!(adapter.get_by_id(&id).await.unwrap().is_none());
}
#[tokio::test]
async fn get_credits_returns_cast_and_crew() {
let pool = pool_with_schema().await;
let adapter = SqlitePersonAdapter::new(pool.clone());
let p = make_person(7, "Diana", Some("Acting"));
adapter.upsert_batch(&[p.clone()]).await.unwrap();
sqlx::query("INSERT INTO movies VALUES ('m1', 'The Film', 2020, 'Dir', NULL, NULL)")
.execute(&pool).await.unwrap();
sqlx::query("INSERT INTO movie_cast VALUES ('m1', 7, 'Diana', 'Hero', 1, NULL)")
.execute(&pool).await.unwrap();
let credits = adapter.get_credits(p.id()).await.unwrap();
assert_eq!(credits.person.name(), "Diana");
assert_eq!(credits.cast.len(), 1);
assert_eq!(credits.cast[0].character, "Hero");
assert!(credits.crew.is_empty());
}