refactor: 大规模模块拆分与代码精简,新增 ai-pipeline/data-contracts 独立 crate

- 新增 aether-ai-pipeline 和 aether-data-contracts crate,将 pipeline 逻辑与数据契约从 gateway 中解耦
- 重构 admin handlers:拆分单体模块为 auth/billing/endpoint/features/model/observability/provider/system 等独立子模块
- 合并 chat/cli 重复代码路径:精简 conversion、finalize、planner 中的 sync/chat/cli 分支
- 重构 scheduler/executor/data 层,引入 facade 模式降低模块间耦合
- 移除冗余的 intent 模块,将 plan_fallback/policy/stream_path/sync_path 迁移至 executor
- 前端适配:调整 admin API 调用和 provider 模型测试对话框
This commit is contained in:
fawney19
2026-04-07 02:50:19 +08:00
parent 763ff03a7b
commit 5d96d6673b
732 changed files with 28589 additions and 20662 deletions
@@ -0,0 +1,9 @@
mod types;
pub use types::{
build_decision_trace, derive_request_candidate_final_status, DecisionTrace,
DecisionTraceCandidate, PublicHealthStatusCount, PublicHealthTimelineBucket,
RequestCandidateFinalStatus, RequestCandidateReadRepository, RequestCandidateRepository,
RequestCandidateStatus, RequestCandidateTrace, RequestCandidateWriteRepository,
StoredRequestCandidate, UpsertRequestCandidateRecord,
};
@@ -0,0 +1,522 @@
use std::collections::BTreeMap;
use async_trait::async_trait;
use crate::repository::provider_catalog::{
StoredProviderCatalogEndpoint, StoredProviderCatalogKey, StoredProviderCatalogProvider,
};
#[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
#[serde(rename_all = "snake_case")]
pub enum RequestCandidateStatus {
Available,
Unused,
Pending,
Streaming,
Success,
Failed,
Cancelled,
Skipped,
}
impl RequestCandidateStatus {
pub fn from_database(value: &str) -> Result<Self, crate::DataLayerError> {
match value.trim().to_ascii_lowercase().as_str() {
"available" => Ok(Self::Available),
"unused" => Ok(Self::Unused),
"pending" => Ok(Self::Pending),
"streaming" => Ok(Self::Streaming),
"success" => Ok(Self::Success),
"failed" => Ok(Self::Failed),
"cancelled" => Ok(Self::Cancelled),
"skipped" => Ok(Self::Skipped),
other => Err(crate::DataLayerError::UnexpectedValue(format!(
"unsupported request_candidates.status: {other}"
))),
}
}
pub fn is_attempted(self, started_at_unix_secs: Option<u64>) -> bool {
match self {
Self::Available | Self::Unused | Self::Skipped => false,
Self::Pending => started_at_unix_secs.is_some(),
Self::Streaming | Self::Success | Self::Failed | Self::Cancelled => true,
}
}
}
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct StoredRequestCandidate {
pub id: String,
pub request_id: String,
pub user_id: Option<String>,
pub api_key_id: Option<String>,
pub username: Option<String>,
pub api_key_name: Option<String>,
pub candidate_index: u32,
pub retry_index: u32,
pub provider_id: Option<String>,
pub endpoint_id: Option<String>,
pub key_id: Option<String>,
pub status: RequestCandidateStatus,
pub skip_reason: Option<String>,
pub is_cached: bool,
pub status_code: Option<u16>,
pub error_type: Option<String>,
pub error_message: Option<String>,
pub latency_ms: Option<u64>,
pub concurrent_requests: Option<u32>,
pub extra_data: Option<serde_json::Value>,
pub required_capabilities: Option<serde_json::Value>,
pub created_at_unix_secs: u64,
pub started_at_unix_secs: Option<u64>,
pub finished_at_unix_secs: Option<u64>,
}
impl StoredRequestCandidate {
#[allow(clippy::too_many_arguments)]
pub fn new(
id: String,
request_id: String,
user_id: Option<String>,
api_key_id: Option<String>,
username: Option<String>,
api_key_name: Option<String>,
candidate_index: i32,
retry_index: i32,
provider_id: Option<String>,
endpoint_id: Option<String>,
key_id: Option<String>,
status: RequestCandidateStatus,
skip_reason: Option<String>,
is_cached: bool,
status_code: Option<i32>,
error_type: Option<String>,
error_message: Option<String>,
latency_ms: Option<i32>,
concurrent_requests: Option<i32>,
extra_data: Option<serde_json::Value>,
required_capabilities: Option<serde_json::Value>,
created_at_unix_secs: i64,
started_at_unix_secs: Option<i64>,
finished_at_unix_secs: Option<i64>,
) -> Result<Self, crate::DataLayerError> {
let candidate_index = u32::try_from(candidate_index).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.candidate_index: {candidate_index}"
))
})?;
let retry_index = u32::try_from(retry_index).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.retry_index: {retry_index}"
))
})?;
let status_code = status_code
.map(|value| {
u16::try_from(value).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.status_code: {value}"
))
})
})
.transpose()?;
let latency_ms = latency_ms
.map(|value| {
u64::try_from(value).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.latency_ms: {value}"
))
})
})
.transpose()?;
let concurrent_requests = concurrent_requests
.map(|value| {
u32::try_from(value).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.concurrent_requests: {value}"
))
})
})
.transpose()?;
let created_at_unix_secs = u64::try_from(created_at_unix_secs).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.created_at_unix_secs: {created_at_unix_secs}"
))
})?;
let started_at_unix_secs = started_at_unix_secs
.map(|value| {
u64::try_from(value).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.started_at_unix_secs: {value}"
))
})
})
.transpose()?;
let finished_at_unix_secs = finished_at_unix_secs
.map(|value| {
u64::try_from(value).map_err(|_| {
crate::DataLayerError::UnexpectedValue(format!(
"invalid request_candidates.finished_at_unix_secs: {value}"
))
})
})
.transpose()?;
Ok(Self {
id,
request_id,
user_id,
api_key_id,
username,
api_key_name,
candidate_index,
retry_index,
provider_id,
endpoint_id,
key_id,
status,
skip_reason,
is_cached,
status_code,
error_type,
error_message,
latency_ms,
concurrent_requests,
extra_data,
required_capabilities,
created_at_unix_secs,
started_at_unix_secs,
finished_at_unix_secs,
})
}
}
#[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
#[serde(rename_all = "snake_case")]
pub enum RequestCandidateFinalStatus {
Success,
Failed,
Cancelled,
Streaming,
Pending,
}
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct RequestCandidateTrace {
pub request_id: String,
pub total_candidates: usize,
pub final_status: RequestCandidateFinalStatus,
pub total_latency_ms: u64,
pub candidates: Vec<StoredRequestCandidate>,
}
impl RequestCandidateTrace {
pub fn from_candidates(
request_id: impl Into<String>,
all_candidates: Vec<StoredRequestCandidate>,
attempted_only: bool,
) -> Option<Self> {
if all_candidates.is_empty() {
return None;
}
let candidates = if attempted_only {
all_candidates
.iter()
.filter(|candidate| {
candidate
.status
.is_attempted(candidate.started_at_unix_secs)
})
.cloned()
.collect::<Vec<_>>()
} else {
all_candidates.clone()
};
let total_latency_ms = candidates
.iter()
.filter(|candidate| {
matches!(
candidate.status,
RequestCandidateStatus::Success
| RequestCandidateStatus::Failed
| RequestCandidateStatus::Cancelled
) && candidate.latency_ms.is_some()
})
.map(|candidate| candidate.latency_ms.unwrap_or(0))
.sum();
let final_status_source = if attempted_only && candidates.is_empty() {
&all_candidates
} else {
&candidates
};
Some(Self {
request_id: request_id.into(),
total_candidates: candidates.len(),
final_status: derive_request_candidate_final_status(final_status_source),
total_latency_ms,
candidates,
})
}
}
pub fn derive_request_candidate_final_status(
candidates: &[StoredRequestCandidate],
) -> RequestCandidateFinalStatus {
let has_success = candidates.iter().any(|candidate| {
candidate.status == RequestCandidateStatus::Success
|| matches!(candidate.status_code, Some(status_code) if (200..300).contains(&status_code))
});
if has_success {
return RequestCandidateFinalStatus::Success;
}
if candidates
.iter()
.any(|candidate| candidate.status == RequestCandidateStatus::Streaming)
{
return RequestCandidateFinalStatus::Streaming;
}
if candidates
.iter()
.any(|candidate| candidate.status == RequestCandidateStatus::Pending)
{
return RequestCandidateFinalStatus::Pending;
}
let has_cancelled = candidates
.iter()
.any(|candidate| candidate.status == RequestCandidateStatus::Cancelled);
let has_failed = candidates
.iter()
.any(|candidate| candidate.status == RequestCandidateStatus::Failed);
if has_cancelled && !has_failed {
return RequestCandidateFinalStatus::Cancelled;
}
RequestCandidateFinalStatus::Failed
}
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct DecisionTraceCandidate {
#[serde(flatten)]
pub candidate: StoredRequestCandidate,
pub provider_name: Option<String>,
pub provider_website: Option<String>,
pub provider_type: Option<String>,
pub endpoint_api_format: Option<String>,
pub endpoint_api_family: Option<String>,
pub endpoint_kind: Option<String>,
pub provider_key_name: Option<String>,
pub provider_key_auth_type: Option<String>,
pub provider_key_capabilities: Option<serde_json::Value>,
pub provider_key_is_active: Option<bool>,
}
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct DecisionTrace {
pub request_id: String,
pub total_candidates: usize,
pub final_status: RequestCandidateFinalStatus,
pub total_latency_ms: u64,
pub candidates: Vec<DecisionTraceCandidate>,
}
pub fn build_decision_trace(
trace: RequestCandidateTrace,
providers: Vec<StoredProviderCatalogProvider>,
endpoints: Vec<StoredProviderCatalogEndpoint>,
keys: Vec<StoredProviderCatalogKey>,
) -> DecisionTrace {
let provider_map = providers
.into_iter()
.map(|item| (item.id.clone(), item))
.collect::<BTreeMap<_, _>>();
let endpoint_map = endpoints
.into_iter()
.map(|item| (item.id.clone(), item))
.collect::<BTreeMap<_, _>>();
let key_map = keys
.into_iter()
.map(|item| (item.id.clone(), item))
.collect::<BTreeMap<_, _>>();
DecisionTrace {
request_id: trace.request_id,
total_candidates: trace.total_candidates,
final_status: trace.final_status,
total_latency_ms: trace.total_latency_ms,
candidates: trace
.candidates
.into_iter()
.map(|candidate| {
enrich_decision_trace_candidate(candidate, &provider_map, &endpoint_map, &key_map)
})
.collect(),
}
}
fn enrich_decision_trace_candidate(
candidate: StoredRequestCandidate,
provider_map: &BTreeMap<String, StoredProviderCatalogProvider>,
endpoint_map: &BTreeMap<String, StoredProviderCatalogEndpoint>,
key_map: &BTreeMap<String, StoredProviderCatalogKey>,
) -> DecisionTraceCandidate {
let provider = candidate
.provider_id
.as_ref()
.and_then(|provider_id| provider_map.get(provider_id));
let endpoint = candidate
.endpoint_id
.as_ref()
.and_then(|endpoint_id| endpoint_map.get(endpoint_id));
let provider_key = candidate
.key_id
.as_ref()
.and_then(|key_id| key_map.get(key_id));
DecisionTraceCandidate {
provider_name: provider.map(|item| item.name.clone()),
provider_website: provider.and_then(|item| item.website.clone()),
provider_type: provider.map(|item| item.provider_type.clone()),
endpoint_api_format: endpoint.map(|item| item.api_format.clone()),
endpoint_api_family: endpoint.and_then(|item| item.api_family.clone()),
endpoint_kind: endpoint.and_then(|item| item.endpoint_kind.clone()),
provider_key_name: provider_key
.map(|item| item.name.clone())
.or_else(|| candidate.api_key_name.clone()),
provider_key_auth_type: provider_key.map(|item| item.auth_type.clone()),
provider_key_capabilities: provider_key.and_then(|item| item.capabilities.clone()),
provider_key_is_active: provider_key.map(|item| item.is_active),
candidate,
}
}
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct PublicHealthStatusCount {
pub endpoint_id: String,
pub status: RequestCandidateStatus,
pub count: u64,
}
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct PublicHealthTimelineBucket {
pub endpoint_id: String,
pub segment_idx: u32,
pub total_count: u64,
pub success_count: u64,
pub failed_count: u64,
pub min_created_at_unix_secs: Option<u64>,
pub max_created_at_unix_secs: Option<u64>,
}
#[async_trait]
pub trait RequestCandidateReadRepository: Send + Sync {
async fn list_by_request_id(
&self,
request_id: &str,
) -> Result<Vec<StoredRequestCandidate>, crate::DataLayerError>;
async fn list_recent(
&self,
limit: usize,
) -> Result<Vec<StoredRequestCandidate>, crate::DataLayerError>;
async fn list_by_provider_id(
&self,
provider_id: &str,
limit: usize,
) -> Result<Vec<StoredRequestCandidate>, crate::DataLayerError>;
async fn list_finalized_by_endpoint_ids_since(
&self,
endpoint_ids: &[String],
since_unix_secs: u64,
limit: usize,
) -> Result<Vec<StoredRequestCandidate>, crate::DataLayerError>;
async fn count_finalized_statuses_by_endpoint_ids_since(
&self,
endpoint_ids: &[String],
since_unix_secs: u64,
) -> Result<Vec<PublicHealthStatusCount>, crate::DataLayerError>;
async fn aggregate_finalized_timeline_by_endpoint_ids_since(
&self,
endpoint_ids: &[String],
since_unix_secs: u64,
until_unix_secs: u64,
segments: u32,
) -> Result<Vec<PublicHealthTimelineBucket>, crate::DataLayerError>;
}
#[derive(Debug, Clone, PartialEq, serde::Serialize, serde::Deserialize)]
pub struct UpsertRequestCandidateRecord {
pub id: String,
pub request_id: String,
pub user_id: Option<String>,
pub api_key_id: Option<String>,
pub username: Option<String>,
pub api_key_name: Option<String>,
pub candidate_index: u32,
pub retry_index: u32,
pub provider_id: Option<String>,
pub endpoint_id: Option<String>,
pub key_id: Option<String>,
pub status: RequestCandidateStatus,
pub skip_reason: Option<String>,
pub is_cached: Option<bool>,
pub status_code: Option<u16>,
pub error_type: Option<String>,
pub error_message: Option<String>,
pub latency_ms: Option<u64>,
pub concurrent_requests: Option<u32>,
pub extra_data: Option<serde_json::Value>,
pub required_capabilities: Option<serde_json::Value>,
pub created_at_unix_secs: Option<u64>,
pub started_at_unix_secs: Option<u64>,
pub finished_at_unix_secs: Option<u64>,
}
impl UpsertRequestCandidateRecord {
pub fn validate(&self) -> Result<(), crate::DataLayerError> {
if self.id.trim().is_empty() {
return Err(crate::DataLayerError::InvalidInput(
"request candidate upsert id cannot be empty".to_string(),
));
}
if self.request_id.trim().is_empty() {
return Err(crate::DataLayerError::InvalidInput(
"request candidate upsert request_id cannot be empty".to_string(),
));
}
Ok(())
}
}
#[async_trait]
pub trait RequestCandidateWriteRepository: Send + Sync {
async fn upsert(
&self,
candidate: UpsertRequestCandidateRecord,
) -> Result<StoredRequestCandidate, crate::DataLayerError>;
async fn delete_created_before(
&self,
created_before_unix_secs: u64,
limit: usize,
) -> Result<usize, crate::DataLayerError>;
}
pub trait RequestCandidateRepository:
RequestCandidateReadRepository + RequestCandidateWriteRepository + Send + Sync
{
}
impl<T> RequestCandidateRepository for T where
T: RequestCandidateReadRepository + RequestCandidateWriteRepository + Send + Sync
{
}