mod common;
use ai_synth_backend::services::llm::mock::MockLlmProvider;
use ai_synth_backend::services::synthesis;
use std::sync::atomic::AtomicBool;
use std::sync::Arc;
use tokio::sync::watch;
use wiremock::matchers::{method, path};
use wiremock::{Mock, MockServer, ResponseTemplate};
async fn setup_mock_server() -> MockServer {
let server = MockServer::start().await;
// Source page with links to articles (for Phase 1 heuristic extraction)
let base = server.uri();
Mock::given(method("GET"))
.and(path("/blog"))
.respond_with(ResponseTemplate::new(200).set_body_string(format!(
r#"
Article One
Article Two
Article Three
"#
)))
.mount(&server)
.await;
// Article pages
for i in 1..=5 {
Mock::given(method("GET"))
.and(path(format!("/article-{}", i)))
.respond_with(ResponseTemplate::new(200).set_body_string(format!(
r#"
Test Article {i}
This is the content of test article {i} about artificial intelligence.
"#
)))
.mount(&server)
.await;
}
server
}
async fn setup_user_with_settings(
app: &common::TestApp,
categories: Vec<&str>,
max_items: i32,
) -> (uuid::Uuid, String, uuid::Uuid) {
let email = format!("pipeline-{}@test.com", uuid::Uuid::new_v4());
let (user_id, session) = app.create_authenticated_user(&email).await;
let settings = serde_json::json!({
"max_articles_per_source": 10,
"max_links_per_source": 8,
"use_brave_search": false,
"article_history_days": 90,
"batch_size": 5,
"source_extraction_window": 3,
"search_agent_behavior": "",
"ai_provider": "",
"ai_model": "",
"ai_model_websearch": "",
"rate_limit_max_requests": null,
"rate_limit_time_window_seconds": null
});
let (status, _) = app.put_with_session("/api/v1/settings", &settings, &session).await;
assert_eq!(status.as_u16(), 200, "Settings save should succeed");
// Create a theme for the pipeline
let categories_json: Vec = categories.iter().map(|c| serde_json::json!(c)).collect();
let theme_body = serde_json::json!({
"name": "Test Theme",
"theme": "Intelligence Artificielle",
"categories": categories_json,
"max_items_per_category": max_items,
"max_age_days": 365,
"summary_length": 3
});
let (theme_status, theme_resp) = app.post_with_session("/api/v1/themes", &theme_body, &session).await;
assert_eq!(theme_status.as_u16(), 201, "Theme creation should succeed");
let theme_id: uuid::Uuid = theme_resp["id"].as_str().unwrap().parse().unwrap();
(user_id, session, theme_id)
}
fn make_progress_channel() -> (Arc>, watch::Receiver) {
let (tx, rx) = watch::channel(synthesis::ProgressEvent::Progress {
step: "init".into(),
message: "Starting...".into(),
percent: 0,
});
(Arc::new(tx), rx)
}
#[tokio::test]
async fn phase1_heuristic_extraction_classifies_articles() {
let app = common::TestApp::new().await;
let mock_server = setup_mock_server().await;
let (user_id, session, theme_id) = setup_user_with_settings(&app, vec!["AI News"], 4).await;
// Add a source pointing to wiremock (same host as article URLs)
let source_url = format!("{}/blog", mock_server.uri());
let source = serde_json::json!({"title": "Test Source", "url": source_url, "theme_id": theme_id.to_string()});
let (status, _) = app.post_with_session("/api/v1/sources", &source, &session).await;
assert!(status.is_success());
let mock_provider = MockLlmProvider::new()
.with_default_category("AI News")
.into_arc();
let job_id = uuid::Uuid::new_v4();
let (tx, _rx) = make_progress_channel();
let state = ai_synth_backend::app_state::AppState::new(
app.config.clone(), app.pool.clone(), reqwest::Client::new(),
);
let result = synthesis::run_generation_inner(
job_id, &state, user_id, theme_id, &tx, Some(mock_provider), &AtomicBool::new(false),
).await;
assert!(result.is_ok(), "Generation should succeed: {:?}", result.err());
let synthesis_id = result.unwrap();
// Verify synthesis was saved with articles
let row: (serde_json::Value,) = sqlx::query_as(
"SELECT sections FROM syntheses WHERE id = $1"
)
.bind(synthesis_id)
.fetch_one(&app.pool)
.await
.expect("Synthesis should exist");
let sections: Vec = serde_json::from_value(row.0).unwrap();
assert!(!sections.is_empty(), "Should have at least one section");
let first_section = §ions[0];
assert_eq!(first_section["title"], "AI News");
let items = first_section["items"].as_array().unwrap();
assert!(!items.is_empty(), "AI News section should have articles");
// Verify article history was recorded
let history_count: (i64,) = sqlx::query_as(
"SELECT COUNT(*) FROM article_history WHERE user_id = $1 AND job_id = $2"
)
.bind(user_id)
.bind(job_id)
.fetch_one(&app.pool)
.await
.unwrap();
assert!(history_count.0 > 0, "Article history should have entries");
}
#[tokio::test]
async fn phase2_search_fills_gaps_when_no_sources() {
let app = common::TestApp::new().await;
let mock_server = setup_mock_server().await;
// No sources -- Phase 1 produces nothing
let (user_id, _session, theme_id) = setup_user_with_settings(&app, vec!["AI News"], 2).await;
let mock_provider = MockLlmProvider::new()
.with_default_category("AI News")
.with_search_urls(vec![
format!("{}/article-1", mock_server.uri()),
format!("{}/article-2", mock_server.uri()),
])
.into_arc();
let job_id = uuid::Uuid::new_v4();
let (tx, _rx) = make_progress_channel();
let state = ai_synth_backend::app_state::AppState::new(
app.config.clone(), app.pool.clone(), reqwest::Client::new(),
);
let result = synthesis::run_generation_inner(
job_id, &state, user_id, theme_id, &tx, Some(mock_provider), &AtomicBool::new(false),
).await;
assert!(result.is_ok(), "Generation should succeed: {:?}", result.err());
// Verify synthesis has articles from Phase 2
let synthesis_id = result.unwrap();
let row: (serde_json::Value,) = sqlx::query_as(
"SELECT sections FROM syntheses WHERE id = $1"
)
.bind(synthesis_id)
.fetch_one(&app.pool)
.await
.unwrap();
let sections: Vec = serde_json::from_value(row.0).unwrap();
assert!(!sections.is_empty(), "Should have sections from Phase 2 search");
}
#[tokio::test]
async fn category_overflow_spills_to_autre() {
let app = common::TestApp::new().await;
let mock_server = setup_mock_server().await;
// max_items_per_category=1, but LLM classifies all articles to "AI News"
let (user_id, session, theme_id) = setup_user_with_settings(&app, vec!["AI News"], 1).await;
let source_url = format!("{}/blog", mock_server.uri());
let source = serde_json::json!({"title": "Test Source", "url": source_url, "theme_id": theme_id.to_string()});
app.post_with_session("/api/v1/sources", &source, &session).await;
let mock_provider = MockLlmProvider::new()
.with_default_category("AI News")
.into_arc();
let job_id = uuid::Uuid::new_v4();
let (tx, _rx) = make_progress_channel();
let state = ai_synth_backend::app_state::AppState::new(
app.config.clone(), app.pool.clone(), reqwest::Client::new(),
);
let result = synthesis::run_generation_inner(
job_id, &state, user_id, theme_id, &tx, Some(mock_provider), &AtomicBool::new(false),
).await;
assert!(result.is_ok(), "Generation should succeed");
let synthesis_id = result.unwrap();
let row: (serde_json::Value,) = sqlx::query_as(
"SELECT sections FROM syntheses WHERE id = $1"
)
.bind(synthesis_id)
.fetch_one(&app.pool)
.await
.unwrap();
let sections: Vec = serde_json::from_value(row.0).unwrap();
// With max_items_per_category=1 and 3 articles all classified as "AI News":
// - 1 goes to AI News
// - Overflow goes to Autre
let ai_section = sections.iter().find(|s| s["title"] == "AI News");
let divers_section = sections.iter().find(|s| s["title"] == "Divers");
assert!(ai_section.is_some(), "Should have AI News section");
let ai_items = ai_section.unwrap()["items"].as_array().unwrap();
assert_eq!(ai_items.len(), 1, "AI News should have exactly 1 item (max)");
if sections.len() > 1 {
assert!(divers_section.is_some(), "Overflow should create Divers section");
}
}
// ── GAP-03: Brave Search pipeline path ────────────────────────────────
//
// The Brave Search code path (`use_brave_search: true`) cannot be tested in
// integration tests without a real Brave API key. The pipeline calls
// `resolve_brave_key` which decrypts a key stored via `user_api_keys`, so we
// would need working AES-256-GCM encryption round-tripping in the test harness.
// The LLM-based web search fallback (Phase 2 without Brave) is already covered
// by `phase2_search_fills_gaps_when_no_sources`.
// ── GAP-05: Source diversity cap ──────────────────────────────────────
#[tokio::test]
async fn source_diversity_limits_articles_per_source() {
let app = common::TestApp::new().await;
let mock_server = setup_mock_server().await;
// Create user with default settings first, then override max_articles_per_source
let email = format!("diversity-{}@test.com", uuid::Uuid::new_v4());
let (user_id, session) = app.create_authenticated_user(&email).await;
// Create theme
let theme_body = serde_json::json!({
"name": "Diversity Theme",
"theme": "Intelligence Artificielle",
"categories": ["AI News"],
"max_items_per_category": 10,
"max_age_days": 365,
"summary_length": 1
});
let (theme_status, theme_resp) = app
.post_with_session("/api/v1/themes", &theme_body, &session)
.await;
assert_eq!(theme_status.as_u16(), 201, "Theme creation should succeed");
let theme_id: uuid::Uuid = theme_resp["id"].as_str().unwrap().parse().unwrap();
// Update settings: max_articles_per_source = 1 so only 1 article from
// the mock server domain is accepted (the source page has 3 links).
let settings = serde_json::json!({
"max_articles_per_source": 1,
"max_links_per_source": 8,
"use_brave_search": false,
"article_history_days": 0,
"batch_size": 1,
"source_extraction_window": 3,
"search_agent_behavior": "",
"ai_provider": "",
"ai_model": "",
"ai_model_websearch": "",
"rate_limit_max_requests": null,
"rate_limit_time_window_seconds": null
});
let (settings_status, _) = app
.put_with_session("/api/v1/settings", &settings, &session)
.await;
assert_eq!(settings_status.as_u16(), 200, "Settings save should succeed");
// Add source pointing to mock server
let source_url = format!("{}/blog", mock_server.uri());
let source = serde_json::json!({
"title": "Diversity Source",
"url": source_url,
"theme_id": theme_id.to_string()
});
let (source_status, _) = app
.post_with_session("/api/v1/sources", &source, &session)
.await;
assert!(source_status.is_success(), "Source creation should succeed");
// Run pipeline
let mock_provider = MockLlmProvider::new()
.with_default_category("AI News")
.into_arc();
let job_id = uuid::Uuid::new_v4();
let (tx, _rx) = make_progress_channel();
let state = ai_synth_backend::app_state::AppState::new(
app.config.clone(),
app.pool.clone(),
reqwest::Client::new(),
);
let result = synthesis::run_generation_inner(
job_id, &state, user_id, theme_id, &tx, Some(mock_provider), &AtomicBool::new(false),
)
.await;
assert!(result.is_ok(), "Generation should succeed: {:?}", result.err());
// Verify that some articles were filtered by source diversity
let diversity_count: (i64,) = sqlx::query_as(
"SELECT COUNT(*) FROM article_history WHERE user_id = $1 AND job_id = $2 AND status = 'filtered_diversity'"
)
.bind(user_id)
.bind(job_id)
.fetch_one(&app.pool)
.await
.unwrap();
// The mock source page has 3 article links, all from the same domain.
// With max_articles_per_source=1, at least 2 should be filtered.
assert!(
diversity_count.0 > 0,
"Should have diversity-filtered articles (got 0)"
);
// Verify the synthesis only contains 1 article (the cap)
let synthesis_id = result.unwrap();
let row: (serde_json::Value,) =
sqlx::query_as("SELECT sections FROM syntheses WHERE id = $1")
.bind(synthesis_id)
.fetch_one(&app.pool)
.await
.unwrap();
let sections: Vec = serde_json::from_value(row.0).unwrap();
let total_items: usize = sections
.iter()
.filter_map(|s| s["items"].as_array())
.map(|items| items.len())
.sum();
assert_eq!(
total_items, 1,
"With max_articles_per_source=1, only 1 article should appear in the synthesis"
);
}
// ── GAP-07: Article history dedup across syntheses ────────────────────
#[tokio::test]
async fn article_history_dedup_prevents_repeat_articles() {
let app = common::TestApp::new().await;
let mock_server = setup_mock_server().await;
let email = format!("dedup-{}@test.com", uuid::Uuid::new_v4());
let (user_id, session) = app.create_authenticated_user(&email).await;
// Create theme
let theme_body = serde_json::json!({
"name": "Dedup Theme",
"theme": "Intelligence Artificielle",
"categories": ["AI News"],
"max_items_per_category": 10,
"max_age_days": 365,
"summary_length": 1
});
let (theme_status, theme_resp) = app
.post_with_session("/api/v1/themes", &theme_body, &session)
.await;
assert_eq!(theme_status.as_u16(), 201, "Theme creation should succeed");
let theme_id: uuid::Uuid = theme_resp["id"].as_str().unwrap().parse().unwrap();
// Settings with article_history_days > 0 to enable dedup
let settings = serde_json::json!({
"max_articles_per_source": 10,
"max_links_per_source": 8,
"use_brave_search": false,
"article_history_days": 90,
"batch_size": 5,
"source_extraction_window": 3,
"search_agent_behavior": "",
"ai_provider": "",
"ai_model": "",
"ai_model_websearch": "",
"rate_limit_max_requests": null,
"rate_limit_time_window_seconds": null
});
let (settings_status, _) = app
.put_with_session("/api/v1/settings", &settings, &session)
.await;
assert_eq!(settings_status.as_u16(), 200, "Settings save should succeed");
// Add source
let source_url = format!("{}/blog", mock_server.uri());
let source = serde_json::json!({
"title": "Dedup Source",
"url": source_url,
"theme_id": theme_id.to_string()
});
let (source_status, _) = app
.post_with_session("/api/v1/sources", &source, &session)
.await;
assert!(source_status.is_success(), "Source creation should succeed");
// ── First generation ──────────────────────────────────────────────
let mock1 = MockLlmProvider::new()
.with_default_category("AI News")
.into_arc();
let job1 = uuid::Uuid::new_v4();
let (tx1, _rx1) = make_progress_channel();
let state1 = ai_synth_backend::app_state::AppState::new(
app.config.clone(),
app.pool.clone(),
reqwest::Client::new(),
);
let result1 = synthesis::run_generation_inner(
job1, &state1, user_id, theme_id, &tx1, Some(mock1), &AtomicBool::new(false),
)
.await;
assert!(result1.is_ok(), "First generation should succeed: {:?}", result1.err());
// Verify first run produced articles
let used_count_1: (i64,) = sqlx::query_as(
"SELECT COUNT(*) FROM article_history WHERE user_id = $1 AND job_id = $2 AND status = 'used'"
)
.bind(user_id)
.bind(job1)
.fetch_one(&app.pool)
.await
.unwrap();
assert!(used_count_1.0 > 0, "First run should produce used articles");
// ── Second generation — same source, same articles ────────────────
let mock2 = MockLlmProvider::new()
.with_default_category("AI News")
.into_arc();
let job2 = uuid::Uuid::new_v4();
let (tx2, _rx2) = make_progress_channel();
let state2 = ai_synth_backend::app_state::AppState::new(
app.config.clone(),
app.pool.clone(),
reqwest::Client::new(),
);
// The second run scrapes the same URLs, which are already in article_history.
// They should be filtered out as "filtered_history".
let _result2 = synthesis::run_generation_inner(
job2, &state2, user_id, theme_id, &tx2, Some(mock2), &AtomicBool::new(false),
)
.await;
// The second run may succeed (empty synthesis) or fail (no valid articles).
// Either way, history-dedup entries must exist.
let dedup_count: (i64,) = sqlx::query_as(
"SELECT COUNT(*) FROM article_history WHERE user_id = $1 AND job_id = $2 AND status = 'filtered_history'"
)
.bind(user_id)
.bind(job2)
.fetch_one(&app.pool)
.await
.unwrap();
assert!(
dedup_count.0 > 0,
"Second run should have history-deduped articles (got 0)"
);
}
// ── GAP-08: Preferred source ordering ─────────────────────────────────
//
// The pipeline places preferred sources before non-preferred ones in the
// processing order (`ordered_sources = [preferred, non_preferred].concat()`).
// Within each wave, articles from preferred source URLs are also placed first
// before being shuffled independently within their group (preferred_urls and
// other_urls are each shuffled separately).
//
// Deterministic ordering at the individual article level cannot be guaranteed
// because articles within the preferred group are randomly shuffled.
//
// What we can verify deterministically:
// - With a preferred source and a non-preferred source, both contribute
// articles to the synthesis (i.e., preferred ordering does not prevent
// non-preferred sources from being processed).
// - The article_history table records entries from both sources.
//
// A test that tries to assert "article from preferred source appears before
// article from non-preferred source in the synthesis" would be flaky due to
// the intentional shuffle within each group.
#[tokio::test]
async fn preferred_sources_processed_first() {
let app = common::TestApp::new().await;
// Set up two source pages on the mock server:
// /blog-a (preferred) with /article-pref
// /blog-b (non-preferred) with /article-norm
let server = wiremock::MockServer::start().await;
let base = server.uri();
// Preferred source page
Mock::given(method("GET"))
.and(path("/blog-a"))
.respond_with(ResponseTemplate::new(200).set_body_string(format!(
r#"Preferred Article"#
)))
.mount(&server)
.await;
// Non-preferred source page
Mock::given(method("GET"))
.and(path("/blog-b"))
.respond_with(ResponseTemplate::new(200).set_body_string(format!(
r#"Normal Article"#
)))
.mount(&server)
.await;
// Article pages
Mock::given(method("GET"))
.and(path("/article-pref"))
.respond_with(ResponseTemplate::new(200).set_body_string(
r#"Preferred Article
This is a preferred article about artificial intelligence research.
"#
.to_string(),
))
.mount(&server)
.await;
Mock::given(method("GET"))
.and(path("/article-norm"))
.respond_with(ResponseTemplate::new(200).set_body_string(
r#"Normal Article
This is a normal article about machine learning and AI systems.
"#
.to_string(),
))
.mount(&server)
.await;
let email = format!("preferred-order-{}@test.com", uuid::Uuid::new_v4());
let (user_id, session) = app.create_authenticated_user(&email).await;
// Create theme
let theme_body = serde_json::json!({
"name": "Preferred Test Theme",
"theme": "Intelligence Artificielle",
"categories": ["AI News"],
"max_items_per_category": 10,
"max_age_days": 365,
"summary_length": 1
});
let (theme_status, theme_resp) = app
.post_with_session("/api/v1/themes", &theme_body, &session)
.await;
assert_eq!(theme_status.as_u16(), 201, "Theme creation should succeed");
let theme_id: uuid::Uuid = theme_resp["id"].as_str().unwrap().parse().unwrap();
// Settings: batch_size=1, source_extraction_window=10 so both sources
// are processed in a single wave, article_history_days=90 to enable tracing
let settings = serde_json::json!({
"max_articles_per_source": 10,
"max_links_per_source": 8,
"use_brave_search": false,
"article_history_days": 90,
"batch_size": 1,
"source_extraction_window": 10,
"search_agent_behavior": "",
"ai_provider": "",
"ai_model": "",
"ai_model_websearch": "",
"rate_limit_max_requests": null,
"rate_limit_time_window_seconds": null
});
let (settings_status, _) = app
.put_with_session("/api/v1/settings", &settings, &session)
.await;
assert_eq!(settings_status.as_u16(), 200, "Settings save should succeed");
// Create source A (will be marked preferred)
let source_a_body = serde_json::json!({
"title": "Source A (preferred)",
"url": format!("{}/blog-a", base),
"theme_id": theme_id.to_string()
});
let (status_a, resp_a) = app
.post_with_session("/api/v1/sources", &source_a_body, &session)
.await;
assert!(status_a.is_success(), "Source A creation should succeed");
let source_a_id = resp_a["id"].as_str().expect("Source A should have an id");
// Create source B (non-preferred)
let source_b_body = serde_json::json!({
"title": "Source B (normal)",
"url": format!("{}/blog-b", base),
"theme_id": theme_id.to_string()
});
let (status_b, _) = app
.post_with_session("/api/v1/sources", &source_b_body, &session)
.await;
assert!(status_b.is_success(), "Source B creation should succeed");
// Mark source A as preferred
let pref_body = serde_json::json!({ "source_ids": [source_a_id], "theme_id": theme_id });
let (pref_status, _) = app
.put_with_session("/api/v1/sources/preferred", &pref_body, &session)
.await;
assert_eq!(pref_status.as_u16(), 200, "Setting preferred sources should succeed");
// Run the pipeline
let mock_provider = MockLlmProvider::new()
.with_default_category("AI News")
.into_arc();
let job_id = uuid::Uuid::new_v4();
let (tx, _rx) = make_progress_channel();
let state = ai_synth_backend::app_state::AppState::new(
app.config.clone(),
app.pool.clone(),
reqwest::Client::new(),
);
let result = synthesis::run_generation_inner(
job_id,
&state,
user_id,
theme_id,
&tx,
Some(mock_provider),
&AtomicBool::new(false),
)
.await;
assert!(result.is_ok(), "Generation should succeed: {:?}", result.err());
// Verify both sources contributed articles to article_history.
// This confirms the preferred-first ordering does not prevent non-preferred
// sources from being processed. Asserting the exact order within the
// synthesis is not done here because articles within each group are
// randomly shuffled by the pipeline.
let used_count: (i64,) = sqlx::query_as(
"SELECT COUNT(*) FROM article_history WHERE user_id = $1 AND job_id = $2 AND status = 'used'",
)
.bind(user_id)
.bind(job_id)
.fetch_one(&app.pool)
.await
.unwrap();
assert!(
used_count.0 >= 2,
"Both preferred and non-preferred source articles should appear in history (got {})",
used_count.0
);
// Verify the synthesis has articles from both sources
let synthesis_id = result.unwrap();
let row: (serde_json::Value,) =
sqlx::query_as("SELECT sections FROM syntheses WHERE id = $1")
.bind(synthesis_id)
.fetch_one(&app.pool)
.await
.unwrap();
let sections: Vec = serde_json::from_value(row.0).unwrap();
let total_items: usize = sections
.iter()
.filter_map(|s| s["items"].as_array())
.map(|items| items.len())
.sum();
assert!(
total_items >= 2,
"Synthesis should contain articles from both sources (got {})",
total_items
);
}