mirror of
https://github.com/perstarkse/minne.git
synced 2026-05-29 19:00:51 +02:00
test: cover system settings sync, validation, and ingestion prompts
Add tests for embedding provider sync, patch isolation, typed backend serde, and DB-backed ingestion prompts.
This commit is contained in:
@@ -499,6 +499,181 @@ mod tests {
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_patch_leaves_unmentioned_fields_unchanged() -> anyhow::Result<()> {
|
||||
let db = SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.with_context(|| "Failed to start in-memory surrealdb".to_string())?;
|
||||
db.apply_migrations()
|
||||
.await
|
||||
.with_context(|| "Failed to apply migrations".to_string())?;
|
||||
|
||||
let original = SystemSettings::get_current(&db)
|
||||
.await
|
||||
.with_context(|| "Failed to get system settings".to_string())?;
|
||||
let sentinel = "custom-query-prompt-sentinel".to_string();
|
||||
|
||||
let patched = SystemSettingsPatch {
|
||||
query_system_prompt: Some(sentinel.clone()),
|
||||
..Default::default()
|
||||
}
|
||||
.apply(&db)
|
||||
.await
|
||||
.with_context(|| "Failed to patch query prompt".to_string())?;
|
||||
|
||||
assert_eq!(patched.query_system_prompt, sentinel);
|
||||
assert_eq!(patched.ingestion_system_prompt, original.ingestion_system_prompt);
|
||||
assert_eq!(patched.query_model, original.query_model);
|
||||
assert_eq!(
|
||||
patched.registrations_enabled,
|
||||
original.registrations_enabled
|
||||
);
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_update_rejects_empty_model_name() -> anyhow::Result<()> {
|
||||
let db = SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.with_context(|| "Failed to start in-memory surrealdb".to_string())?;
|
||||
db.apply_migrations()
|
||||
.await
|
||||
.with_context(|| "Failed to apply migrations".to_string())?;
|
||||
|
||||
let mut invalid_settings = SystemSettings::get_current(&db)
|
||||
.await
|
||||
.with_context(|| "Failed to get system settings".to_string())?;
|
||||
invalid_settings.query_model = " ".to_string();
|
||||
|
||||
let result = SystemSettings::update(&db, invalid_settings).await;
|
||||
assert!(matches!(result, Err(AppError::Validation(_))));
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_update_normalizes_record_id() -> anyhow::Result<()> {
|
||||
let db = SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.with_context(|| "Failed to start in-memory surrealdb".to_string())?;
|
||||
db.apply_migrations()
|
||||
.await
|
||||
.with_context(|| "Failed to apply migrations".to_string())?;
|
||||
|
||||
let mut settings = SystemSettings::get_current(&db)
|
||||
.await
|
||||
.with_context(|| "Failed to get system settings".to_string())?;
|
||||
settings.id = "wrong-id".to_string();
|
||||
|
||||
let updated = SystemSettings::update(&db, settings)
|
||||
.await
|
||||
.with_context(|| "Failed to update settings".to_string())?;
|
||||
assert_eq!(updated.id, SystemSettings::RECORD_ID);
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_update_preserves_embedding_backend() -> anyhow::Result<()> {
|
||||
use crate::utils::embedding::EmbeddingProvider;
|
||||
|
||||
let db = SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.with_context(|| "Failed to start in-memory surrealdb".to_string())?;
|
||||
db.apply_migrations()
|
||||
.await
|
||||
.with_context(|| "Failed to apply migrations".to_string())?;
|
||||
|
||||
let provider = EmbeddingProvider::new_hashed(384)
|
||||
.with_context(|| "Failed to create hashed embedding provider".to_string())?;
|
||||
SystemSettings::sync_from_embedding_provider(&db, &provider)
|
||||
.await
|
||||
.with_context(|| "Failed to sync embedding provider".to_string())?;
|
||||
|
||||
let synced = SystemSettings::get_current(&db)
|
||||
.await
|
||||
.with_context(|| "Failed to get synced settings".to_string())?;
|
||||
assert_eq!(synced.embedding_backend, Some(EmbeddingBackend::Hashed));
|
||||
|
||||
let mut tampered = synced;
|
||||
tampered.embedding_backend = Some(EmbeddingBackend::OpenAI);
|
||||
let updated = SystemSettings::update(&db, tampered)
|
||||
.await
|
||||
.with_context(|| "Failed to update settings".to_string())?;
|
||||
|
||||
assert_eq!(updated.embedding_backend, Some(EmbeddingBackend::Hashed));
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_sync_from_embedding_provider_updates_mismatched_settings() -> anyhow::Result<()> {
|
||||
use crate::utils::embedding::EmbeddingProvider;
|
||||
|
||||
let db = SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.with_context(|| "Failed to start in-memory surrealdb".to_string())?;
|
||||
db.apply_migrations()
|
||||
.await
|
||||
.with_context(|| "Failed to apply migrations".to_string())?;
|
||||
|
||||
let provider = EmbeddingProvider::new_hashed(384)
|
||||
.with_context(|| "Failed to create hashed embedding provider".to_string())?;
|
||||
let (settings, changed) = SystemSettings::sync_from_embedding_provider(&db, &provider)
|
||||
.await
|
||||
.with_context(|| "Failed to sync embedding provider".to_string())?;
|
||||
|
||||
assert!(changed);
|
||||
assert_eq!(settings.embedding_backend, Some(EmbeddingBackend::Hashed));
|
||||
assert_eq!(settings.embedding_dimensions, 384);
|
||||
|
||||
let persisted = SystemSettings::get_current(&db)
|
||||
.await
|
||||
.with_context(|| "Failed to reload synced settings".to_string())?;
|
||||
assert_eq!(persisted.embedding_backend, Some(EmbeddingBackend::Hashed));
|
||||
assert_eq!(persisted.embedding_dimensions, 384);
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_sync_from_embedding_provider_is_noop_when_already_synced() -> anyhow::Result<()> {
|
||||
use crate::utils::embedding::EmbeddingProvider;
|
||||
|
||||
let db = SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.with_context(|| "Failed to start in-memory surrealdb".to_string())?;
|
||||
db.apply_migrations()
|
||||
.await
|
||||
.with_context(|| "Failed to apply migrations".to_string())?;
|
||||
|
||||
let provider = EmbeddingProvider::new_hashed(384)
|
||||
.with_context(|| "Failed to create hashed embedding provider".to_string())?;
|
||||
SystemSettings::sync_from_embedding_provider(&db, &provider)
|
||||
.await
|
||||
.with_context(|| "Failed to initial sync".to_string())?;
|
||||
|
||||
let (_, changed) = SystemSettings::sync_from_embedding_provider(&db, &provider)
|
||||
.await
|
||||
.with_context(|| "Failed to repeat sync".to_string())?;
|
||||
assert!(!changed);
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_sync_rejects_provider_dimension_above_u32_max() -> anyhow::Result<()> {
|
||||
use crate::utils::embedding::EmbeddingProvider;
|
||||
|
||||
let db = SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.with_context(|| "Failed to start in-memory surrealdb".to_string())?;
|
||||
db.apply_migrations()
|
||||
.await
|
||||
.with_context(|| "Failed to apply migrations".to_string())?;
|
||||
|
||||
let provider = EmbeddingProvider::new_hashed((u32::MAX as usize) + 1)
|
||||
.with_context(|| "Failed to create oversized hashed provider".to_string())?;
|
||||
let result = SystemSettings::sync_from_embedding_provider(&db, &provider).await;
|
||||
assert!(matches!(result, Err(AppError::Validation(_))));
|
||||
Ok(())
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_migration_after_changing_embedding_length() -> anyhow::Result<()> {
|
||||
let db = SurrealDbClient::memory("test", &Uuid::new_v4().to_string())
|
||||
|
||||
@@ -453,3 +453,72 @@ pub async fn generate_embedding_with_params(
|
||||
|
||||
Ok(embedding)
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::{EmbeddingBackend, ParseEmbeddingBackendError};
|
||||
use crate::storage::types::system_settings::SystemSettings;
|
||||
use serde_json::json;
|
||||
|
||||
#[test]
|
||||
fn embedding_backend_as_str_matches_serde_names() {
|
||||
assert_eq!(EmbeddingBackend::OpenAI.as_str(), "openai");
|
||||
assert_eq!(EmbeddingBackend::FastEmbed.as_str(), "fastembed");
|
||||
assert_eq!(EmbeddingBackend::Hashed.as_str(), "hashed");
|
||||
|
||||
assert_eq!(
|
||||
serde_json::to_string(&EmbeddingBackend::FastEmbed).expect("serialize"),
|
||||
"\"fastembed\""
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn embedding_backend_deserializes_lowercase_values() {
|
||||
let openai: EmbeddingBackend = serde_json::from_str("\"openai\"").expect("openai");
|
||||
let fastembed: EmbeddingBackend = serde_json::from_str("\"fastembed\"").expect("fastembed");
|
||||
let hashed: EmbeddingBackend = serde_json::from_str("\"hashed\"").expect("hashed");
|
||||
|
||||
assert_eq!(openai, EmbeddingBackend::OpenAI);
|
||||
assert_eq!(fastembed, EmbeddingBackend::FastEmbed);
|
||||
assert_eq!(hashed, EmbeddingBackend::Hashed);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn embedding_backend_from_str_accepts_aliases() {
|
||||
assert_eq!(
|
||||
"fast-embed".parse::<EmbeddingBackend>().expect("fast-embed"),
|
||||
EmbeddingBackend::FastEmbed
|
||||
);
|
||||
assert_eq!(
|
||||
"FASTEMBED".parse::<EmbeddingBackend>().expect("FASTEMBED"),
|
||||
EmbeddingBackend::FastEmbed
|
||||
);
|
||||
assert!(matches!(
|
||||
"unknown-backend".parse::<EmbeddingBackend>(),
|
||||
Err(ParseEmbeddingBackendError { .. })
|
||||
));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn system_settings_deserializes_embedding_backend_field() {
|
||||
let value = json!({
|
||||
"id": "current",
|
||||
"registrations_enabled": true,
|
||||
"require_email_verification": false,
|
||||
"query_model": "gpt-4o-mini",
|
||||
"processing_model": "gpt-4o-mini",
|
||||
"embedding_model": "text-embedding-3-small",
|
||||
"embedding_dimensions": 1536,
|
||||
"embedding_backend": "hashed",
|
||||
"query_system_prompt": "query",
|
||||
"ingestion_system_prompt": "ingestion",
|
||||
"image_processing_model": "gpt-4o-mini",
|
||||
"image_processing_prompt": "image",
|
||||
"voice_processing_model": "whisper-1",
|
||||
});
|
||||
|
||||
let settings: SystemSettings =
|
||||
serde_json::from_value(value).expect("deserialize system settings");
|
||||
assert_eq!(settings.embedding_backend, Some(EmbeddingBackend::Hashed));
|
||||
}
|
||||
}
|
||||
|
||||
@@ -348,3 +348,84 @@ fn truncate_for_embedding(text: &str, max_chars: usize) -> String {
|
||||
truncated.push('…');
|
||||
truncated
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use std::sync::Arc;
|
||||
|
||||
use anyhow::Context;
|
||||
use async_openai::{config::OpenAIConfig, types::ChatCompletionRequestMessage, Client};
|
||||
use common::{
|
||||
storage::{
|
||||
db::SurrealDbClient,
|
||||
store::StorageManager,
|
||||
types::system_settings::SystemSettingsPatch,
|
||||
},
|
||||
utils::{
|
||||
config::{AppConfig, StorageKind},
|
||||
embedding::EmbeddingProvider,
|
||||
},
|
||||
};
|
||||
use uuid::Uuid;
|
||||
|
||||
use super::DefaultPipelineServices;
|
||||
|
||||
fn system_prompt_from_request(
|
||||
request: &async_openai::types::CreateChatCompletionRequest,
|
||||
) -> String {
|
||||
let ChatCompletionRequestMessage::System(system) = &request.messages[0] else {
|
||||
panic!("expected first message to be system");
|
||||
};
|
||||
match &system.content {
|
||||
async_openai::types::ChatCompletionRequestSystemMessageContent::Text(text) => {
|
||||
text.clone()
|
||||
}
|
||||
other => panic!("unexpected system message content: {other:?}"),
|
||||
}
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn prepare_llm_request_uses_ingestion_prompt_from_system_settings(
|
||||
) -> anyhow::Result<()> {
|
||||
const SENTINEL: &str = "ingestion-prompt-sentinel-from-db";
|
||||
|
||||
let db = Arc::new(
|
||||
SurrealDbClient::memory("test_ns", &Uuid::new_v4().to_string())
|
||||
.await
|
||||
.context("start in-memory db")?,
|
||||
);
|
||||
db.apply_migrations().await.context("apply migrations")?;
|
||||
SystemSettingsPatch {
|
||||
ingestion_system_prompt: Some(SENTINEL.to_string()),
|
||||
..Default::default()
|
||||
}
|
||||
.apply(&db)
|
||||
.await
|
||||
.context("patch ingestion prompt")?;
|
||||
|
||||
let config = AppConfig {
|
||||
storage: StorageKind::Memory,
|
||||
..Default::default()
|
||||
};
|
||||
let storage = StorageManager::new(&config).await.context("storage manager")?;
|
||||
let openai_client = Arc::new(Client::with_config(OpenAIConfig::default()));
|
||||
let embedding_provider = Arc::new(EmbeddingProvider::new_hashed(384)?);
|
||||
|
||||
let services = DefaultPipelineServices::new(
|
||||
db,
|
||||
openai_client,
|
||||
config,
|
||||
None,
|
||||
storage,
|
||||
embedding_provider,
|
||||
);
|
||||
|
||||
let request = services
|
||||
.prepare_llm_request("notes", None, "hello world", &[])
|
||||
.await
|
||||
.context("prepare llm request")?;
|
||||
|
||||
assert_eq!(system_prompt_from_request(&request), SENTINEL);
|
||||
Ok(())
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user