mirror of
https://github.com/fawney19/Aether.git
synced 2026-09-01 17:00:21 +08:00
fix(admin): 补齐 key 自动获取模型的即时刷新场景 (#295)
- 新增 key 且开启自动获取时立即抓取并写回 allowed_models - 自动获取已开启时修改包含/排除规则后立即刷新 allowed_models - 补充创建与过滤规则变更场景的控制层回归测试
This commit is contained in:
fawney19
@@ -1,7 +1,7 @@
|
||||
use crate::handlers::admin::provider::shared::paths::admin_provider_id_for_keys;
|
||||
use crate::handlers::admin::provider::shared::payloads::AdminProviderKeyCreateRequest;
|
||||
use crate::handlers::admin::request::{AdminAppState, AdminRequestContext};
|
||||
use crate::GatewayError;
|
||||
use crate::{model_fetch::perform_model_fetch_for_key, GatewayError};
|
||||
use axum::{
|
||||
body::{Body, Bytes},
|
||||
http,
|
||||
@@ -63,6 +63,32 @@ pub(super) async fn maybe_handle(
|
||||
let Some(created) = state.create_provider_catalog_key(&record).await? else {
|
||||
return Ok(None);
|
||||
};
|
||||
let key_id = created.id.clone();
|
||||
let created = if created.auto_fetch_models {
|
||||
let summary =
|
||||
perform_model_fetch_for_key(state.as_ref(), &provider.id, &created.id).await?;
|
||||
if summary.succeeded == 0 {
|
||||
let detail = state
|
||||
.read_provider_catalog_keys_by_ids(std::slice::from_ref(&key_id))
|
||||
.await?
|
||||
.into_iter()
|
||||
.next()
|
||||
.and_then(|key| key.last_models_fetch_error)
|
||||
.unwrap_or_else(|| "未获取到可用上游模型".to_string());
|
||||
return Err(GatewayError::Internal(format!(
|
||||
"开启自动获取模型后同步上游模型失败: {detail}"
|
||||
)));
|
||||
}
|
||||
|
||||
state
|
||||
.read_provider_catalog_keys_by_ids(std::slice::from_ref(&key_id))
|
||||
.await?
|
||||
.into_iter()
|
||||
.next()
|
||||
.unwrap_or(created)
|
||||
} else {
|
||||
created
|
||||
};
|
||||
let now_unix_secs = SystemTime::now()
|
||||
.duration_since(UNIX_EPOCH)
|
||||
.ok()
|
||||
|
||||
@@ -81,8 +81,12 @@ pub(super) async fn maybe_handle(
|
||||
let Some(updated) = state.update_provider_catalog_key(&updated_record).await? else {
|
||||
return Ok(None);
|
||||
};
|
||||
let should_overwrite_allowed_models_immediately =
|
||||
!existing_key.auto_fetch_models && updated.auto_fetch_models;
|
||||
let auto_fetch_filters_changed = existing_key.model_include_patterns
|
||||
!= updated.model_include_patterns
|
||||
|| existing_key.model_exclude_patterns != updated.model_exclude_patterns;
|
||||
// 自动获取开启后，调整过滤规则也要立即刷新 allowed_models。
|
||||
let should_overwrite_allowed_models_immediately = updated.auto_fetch_models
|
||||
&& (!existing_key.auto_fetch_models || auto_fetch_filters_changed);
|
||||
let updated = if should_overwrite_allowed_models_immediately {
|
||||
let summary =
|
||||
perform_model_fetch_for_key(state.as_ref(), &provider.id, &updated.id).await?;
|
||||
|
||||
@@ -282,6 +282,114 @@ async fn gateway_creates_admin_provider_key_locally_with_trusted_admin_principal
|
||||
upstream_handle.abort();
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn gateway_fetches_allowed_models_immediately_when_creating_key_with_auto_fetch() {
|
||||
let execution_runtime_hits = Arc::new(Mutex::new(0usize));
|
||||
let execution_runtime_hits_clone = Arc::clone(&execution_runtime_hits);
|
||||
let execution_runtime = Router::new().route(
|
||||
"/v1/execute/sync",
|
||||
any(move |Json(plan): Json<ExecutionPlan>| {
|
||||
let execution_runtime_hits_inner = Arc::clone(&execution_runtime_hits_clone);
|
||||
async move {
|
||||
*execution_runtime_hits_inner
|
||||
.lock()
|
||||
.expect("mutex should lock") += 1;
|
||||
assert_eq!(plan.url, "https://api.openai.example/v1/models");
|
||||
assert_eq!(
|
||||
plan.headers.get("authorization").map(String::as_str),
|
||||
Some("Bearer sk-created-openai")
|
||||
);
|
||||
Json(json!({
|
||||
"request_id": "req-create-key-auto-fetch",
|
||||
"status_code": 200,
|
||||
"headers": {
|
||||
"content-type": "application/json"
|
||||
},
|
||||
"body": {
|
||||
"json_body": {
|
||||
"data": [
|
||||
{"id": "gpt-5-mini"},
|
||||
{"id": "gpt-4.1-mini"}
|
||||
]
|
||||
}
|
||||
}
|
||||
}))
|
||||
}
|
||||
}),
|
||||
);
|
||||
let (execution_runtime_url, execution_runtime_handle) = start_server(execution_runtime).await;
|
||||
|
||||
let provider_catalog_repository = Arc::new(InMemoryProviderCatalogReadRepository::seed(
|
||||
vec![sample_provider("provider-openai", "openai", 10)],
|
||||
vec![sample_endpoint(
|
||||
"endpoint-openai-chat",
|
||||
"provider-openai",
|
||||
"openai:chat",
|
||||
"https://api.openai.example",
|
||||
)],
|
||||
vec![],
|
||||
));
|
||||
|
||||
let gateway = build_router_with_state(
|
||||
build_state_with_execution_runtime_override(execution_runtime_url)
|
||||
.with_data_state_for_tests(
|
||||
GatewayDataState::with_provider_catalog_repository_for_tests(
|
||||
provider_catalog_repository.clone(),
|
||||
)
|
||||
.with_encryption_key_for_tests(DEVELOPMENT_ENCRYPTION_KEY),
|
||||
),
|
||||
);
|
||||
let (gateway_url, gateway_handle) = start_server(gateway).await;
|
||||
|
||||
let response = reqwest::Client::new()
|
||||
.post(format!(
|
||||
"{gateway_url}/api/admin/endpoints/providers/provider-openai/keys"
|
||||
))
|
||||
.header(crate::constants::GATEWAY_HEADER, "rust-phase3b")
|
||||
.header(TRUSTED_ADMIN_USER_ID_HEADER, "admin-user-123")
|
||||
.header(TRUSTED_ADMIN_USER_ROLE_HEADER, "admin")
|
||||
.header(TRUSTED_ADMIN_SESSION_ID_HEADER, "session-123")
|
||||
.json(&json!({
|
||||
"api_formats": ["openai:chat"],
|
||||
"api_key": "sk-created-openai",
|
||||
"name": "created key with auto fetch",
|
||||
"auto_fetch_models": true
|
||||
}))
|
||||
.send()
|
||||
.await
|
||||
.expect("request should succeed");
|
||||
|
||||
assert_eq!(response.status(), StatusCode::OK);
|
||||
let payload: serde_json::Value = response.json().await.expect("json body should parse");
|
||||
assert_eq!(payload["provider_id"], "provider-openai");
|
||||
assert_eq!(payload["name"], "created key with auto fetch");
|
||||
assert_eq!(payload["auto_fetch_models"], true);
|
||||
assert_eq!(
|
||||
payload["allowed_models"],
|
||||
json!(["gpt-4.1-mini", "gpt-5-mini"])
|
||||
);
|
||||
assert_eq!(payload["last_models_fetch_error"], serde_json::Value::Null);
|
||||
assert_eq!(
|
||||
*execution_runtime_hits.lock().expect("mutex should lock"),
|
||||
1
|
||||
);
|
||||
|
||||
let keys = provider_catalog_repository
|
||||
.list_keys_by_provider_ids(&["provider-openai".to_string()])
|
||||
.await
|
||||
.expect("keys should read");
|
||||
assert_eq!(keys.len(), 1);
|
||||
assert_eq!(keys[0].name, "created key with auto fetch");
|
||||
assert!(keys[0].auto_fetch_models);
|
||||
assert_eq!(
|
||||
keys[0].allowed_models,
|
||||
Some(json!(["gpt-4.1-mini", "gpt-5-mini"]))
|
||||
);
|
||||
|
||||
gateway_handle.abort();
|
||||
execution_runtime_handle.abort();
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn gateway_reveals_admin_provider_key_locally_with_trusted_admin_principal() {
|
||||
let upstream_hits = Arc::new(Mutex::new(0usize));
|
||||
@@ -947,6 +1055,220 @@ async fn gateway_fetches_allowed_models_immediately_when_enabling_auto_fetch_fro
|
||||
execution_runtime_handle.abort();
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn gateway_refreshes_allowed_models_when_updating_include_patterns_with_auto_fetch_enabled() {
|
||||
let execution_runtime_hits = Arc::new(Mutex::new(0usize));
|
||||
let execution_runtime_hits_clone = Arc::clone(&execution_runtime_hits);
|
||||
let execution_runtime = Router::new().route(
|
||||
"/v1/execute/sync",
|
||||
any(move |Json(plan): Json<ExecutionPlan>| {
|
||||
let execution_runtime_hits_inner = Arc::clone(&execution_runtime_hits_clone);
|
||||
async move {
|
||||
*execution_runtime_hits_inner
|
||||
.lock()
|
||||
.expect("mutex should lock") += 1;
|
||||
assert_eq!(plan.url, "https://api.openai.example/v1/models");
|
||||
Json(json!({
|
||||
"request_id": "req-update-key-include-patterns",
|
||||
"status_code": 200,
|
||||
"headers": {
|
||||
"content-type": "application/json"
|
||||
},
|
||||
"body": {
|
||||
"json_body": {
|
||||
"data": [
|
||||
{"id": "gpt-5"},
|
||||
{"id": "gpt-4.1"},
|
||||
{"id": "claude-3.7-sonnet"}
|
||||
]
|
||||
}
|
||||
}
|
||||
}))
|
||||
}
|
||||
}),
|
||||
);
|
||||
let (execution_runtime_url, execution_runtime_handle) = start_server(execution_runtime).await;
|
||||
|
||||
let mut key = sample_key(
|
||||
"key-openai-a",
|
||||
"provider-openai",
|
||||
"openai:chat",
|
||||
"sk-test-a",
|
||||
);
|
||||
key.auto_fetch_models = true;
|
||||
key.allowed_models = Some(json!(["gpt-4.1", "gpt-5", "claude-3.7-sonnet"]));
|
||||
key.model_include_patterns = Some(json!(["gpt-*", "claude-*"]));
|
||||
|
||||
let provider_catalog_repository = Arc::new(InMemoryProviderCatalogReadRepository::seed(
|
||||
vec![sample_provider("provider-openai", "openai", 10)],
|
||||
vec![sample_endpoint(
|
||||
"endpoint-openai-chat",
|
||||
"provider-openai",
|
||||
"openai:chat",
|
||||
"https://api.openai.example",
|
||||
)],
|
||||
vec![key],
|
||||
));
|
||||
|
||||
let gateway = build_router_with_state(
|
||||
build_state_with_execution_runtime_override(execution_runtime_url)
|
||||
.with_data_state_for_tests(
|
||||
GatewayDataState::with_provider_catalog_repository_for_tests(
|
||||
provider_catalog_repository.clone(),
|
||||
)
|
||||
.with_encryption_key_for_tests(DEVELOPMENT_ENCRYPTION_KEY),
|
||||
),
|
||||
);
|
||||
let (gateway_url, gateway_handle) = start_server(gateway).await;
|
||||
|
||||
let response = reqwest::Client::new()
|
||||
.put(format!(
|
||||
"{gateway_url}/api/admin/endpoints/keys/key-openai-a"
|
||||
))
|
||||
.header(crate::constants::GATEWAY_HEADER, "rust-phase3b")
|
||||
.header(TRUSTED_ADMIN_USER_ID_HEADER, "admin-user-123")
|
||||
.header(TRUSTED_ADMIN_USER_ROLE_HEADER, "admin")
|
||||
.header(TRUSTED_ADMIN_SESSION_ID_HEADER, "session-123")
|
||||
.json(&json!({
|
||||
"model_include_patterns": ["gpt-5*"]
|
||||
}))
|
||||
.send()
|
||||
.await
|
||||
.expect("request should succeed");
|
||||
|
||||
assert_eq!(response.status(), StatusCode::OK);
|
||||
let payload: serde_json::Value = response.json().await.expect("json body should parse");
|
||||
assert_eq!(payload["auto_fetch_models"], true);
|
||||
assert_eq!(payload["model_include_patterns"], json!(["gpt-5*"]));
|
||||
assert_eq!(payload["allowed_models"], json!(["gpt-5"]));
|
||||
assert_eq!(
|
||||
*execution_runtime_hits.lock().expect("mutex should lock"),
|
||||
1
|
||||
);
|
||||
|
||||
let reloaded = provider_catalog_repository
|
||||
.list_keys_by_ids(&["key-openai-a".to_string()])
|
||||
.await
|
||||
.expect("keys should read");
|
||||
assert_eq!(reloaded.len(), 1);
|
||||
assert!(reloaded[0].auto_fetch_models);
|
||||
assert_eq!(reloaded[0].model_include_patterns, Some(json!(["gpt-5*"])));
|
||||
assert_eq!(reloaded[0].allowed_models, Some(json!(["gpt-5"])));
|
||||
|
||||
gateway_handle.abort();
|
||||
execution_runtime_handle.abort();
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn gateway_refreshes_allowed_models_when_updating_exclude_patterns_with_auto_fetch_enabled() {
|
||||
let execution_runtime_hits = Arc::new(Mutex::new(0usize));
|
||||
let execution_runtime_hits_clone = Arc::clone(&execution_runtime_hits);
|
||||
let execution_runtime = Router::new().route(
|
||||
"/v1/execute/sync",
|
||||
any(move |Json(plan): Json<ExecutionPlan>| {
|
||||
let execution_runtime_hits_inner = Arc::clone(&execution_runtime_hits_clone);
|
||||
async move {
|
||||
*execution_runtime_hits_inner
|
||||
.lock()
|
||||
.expect("mutex should lock") += 1;
|
||||
assert_eq!(plan.url, "https://api.openai.example/v1/models");
|
||||
Json(json!({
|
||||
"request_id": "req-update-key-exclude-patterns",
|
||||
"status_code": 200,
|
||||
"headers": {
|
||||
"content-type": "application/json"
|
||||
},
|
||||
"body": {
|
||||
"json_body": {
|
||||
"data": [
|
||||
{"id": "gpt-5"},
|
||||
{"id": "gpt-4.1"},
|
||||
{"id": "claude-3.7-sonnet"}
|
||||
]
|
||||
}
|
||||
}
|
||||
}))
|
||||
}
|
||||
}),
|
||||
);
|
||||
let (execution_runtime_url, execution_runtime_handle) = start_server(execution_runtime).await;
|
||||
|
||||
let mut key = sample_key(
|
||||
"key-openai-a",
|
||||
"provider-openai",
|
||||
"openai:chat",
|
||||
"sk-test-a",
|
||||
);
|
||||
key.auto_fetch_models = true;
|
||||
key.allowed_models = Some(json!(["gpt-5", "gpt-4.1", "claude-3.7-sonnet"]));
|
||||
key.model_exclude_patterns = Some(json!(["claude-*"]));
|
||||
|
||||
let provider_catalog_repository = Arc::new(InMemoryProviderCatalogReadRepository::seed(
|
||||
vec![sample_provider("provider-openai", "openai", 10)],
|
||||
vec![sample_endpoint(
|
||||
"endpoint-openai-chat",
|
||||
"provider-openai",
|
||||
"openai:chat",
|
||||
"https://api.openai.example",
|
||||
)],
|
||||
vec![key],
|
||||
));
|
||||
|
||||
let gateway = build_router_with_state(
|
||||
build_state_with_execution_runtime_override(execution_runtime_url)
|
||||
.with_data_state_for_tests(
|
||||
GatewayDataState::with_provider_catalog_repository_for_tests(
|
||||
provider_catalog_repository.clone(),
|
||||
)
|
||||
.with_encryption_key_for_tests(DEVELOPMENT_ENCRYPTION_KEY),
|
||||
),
|
||||
);
|
||||
let (gateway_url, gateway_handle) = start_server(gateway).await;
|
||||
|
||||
let response = reqwest::Client::new()
|
||||
.put(format!(
|
||||
"{gateway_url}/api/admin/endpoints/keys/key-openai-a"
|
||||
))
|
||||
.header(crate::constants::GATEWAY_HEADER, "rust-phase3b")
|
||||
.header(TRUSTED_ADMIN_USER_ID_HEADER, "admin-user-123")
|
||||
.header(TRUSTED_ADMIN_USER_ROLE_HEADER, "admin")
|
||||
.header(TRUSTED_ADMIN_SESSION_ID_HEADER, "session-123")
|
||||
.json(&json!({
|
||||
"model_exclude_patterns": ["gpt-4*"]
|
||||
}))
|
||||
.send()
|
||||
.await
|
||||
.expect("request should succeed");
|
||||
|
||||
assert_eq!(response.status(), StatusCode::OK);
|
||||
let payload: serde_json::Value = response.json().await.expect("json body should parse");
|
||||
assert_eq!(payload["auto_fetch_models"], true);
|
||||
assert_eq!(payload["model_exclude_patterns"], json!(["gpt-4*"]));
|
||||
assert_eq!(
|
||||
payload["allowed_models"],
|
||||
json!(["claude-3.7-sonnet", "gpt-5"])
|
||||
);
|
||||
assert_eq!(
|
||||
*execution_runtime_hits.lock().expect("mutex should lock"),
|
||||
1
|
||||
);
|
||||
|
||||
let reloaded = provider_catalog_repository
|
||||
.list_keys_by_ids(&["key-openai-a".to_string()])
|
||||
.await
|
||||
.expect("keys should read");
|
||||
assert_eq!(reloaded.len(), 1);
|
||||
assert!(reloaded[0].auto_fetch_models);
|
||||
assert_eq!(reloaded[0].model_exclude_patterns, Some(json!(["gpt-4*"])));
|
||||
assert_eq!(
|
||||
reloaded[0].allowed_models,
|
||||
Some(json!(["claude-3.7-sonnet", "gpt-5"]))
|
||||
);
|
||||
|
||||
gateway_handle.abort();
|
||||
execution_runtime_handle.abort();
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn gateway_rejects_admin_provider_key_update_when_api_key_duplicates_existing_key() {
|
||||
let upstream_hits = Arc::new(Mutex::new(0usize));
|
||||
|
||||
Reference in New Issue
Block a user