Merge pull request #371 from Kayphoon/feature/embedding-model-support

feat: add embedding and rerank support
This commit is contained in:
fawney19
2026-05-04 12:56:11 +08:00
committed by GitHub
87 changed files with 5517 additions and 184 deletions

View File

@@ -8,6 +8,38 @@ use aether_data_contracts::repository::global_models::AdminGlobalModelListQuery;
use serde_json::json;
use std::collections::BTreeMap;
const EMBEDDING_API_FORMATS: &[&str] = &[
"openai:embedding",
"jina:embedding",
"gemini:embedding",
"doubao:embedding",
];
fn json_value_contains_string(value: &serde_json::Value, expected: &str) -> bool {
match value {
serde_json::Value::String(value) => value.trim().eq_ignore_ascii_case(expected),
serde_json::Value::Array(values) => values
.iter()
.any(|value| json_value_contains_string(value, expected)),
serde_json::Value::Object(object) => object
.values()
.any(|value| json_value_contains_string(value, expected)),
_ => false,
}
}
fn json_value_contains_embedding_metadata(value: &serde_json::Value) -> bool {
value
.as_object()
.and_then(|object| object.get("embedding"))
.and_then(serde_json::Value::as_bool)
.unwrap_or(false)
|| json_value_contains_string(value, "embedding")
|| EMBEDDING_API_FORMATS
.iter()
.any(|api_format| json_value_contains_string(value, api_format))
}
pub(crate) async fn build_admin_global_model_providers_payload(
state: &AdminAppState<'_>,
global_model_id: &str,
@@ -53,6 +85,7 @@ pub(crate) async fn build_admin_global_model_providers_payload(
"supports_vision": admin_provider_model_effective_capability(&model, "vision"),
"supports_function_calling": admin_provider_model_effective_capability(&model, "function_calling"),
"supports_streaming": admin_provider_model_effective_capability(&model, "streaming"),
"supports_embedding": admin_provider_model_effective_capability(&model, "embedding"),
"is_active": model.is_active,
}))
})
@@ -123,6 +156,14 @@ pub(crate) async fn build_admin_model_catalog_payload(
.and_then(|value| value.get("streaming"))
.and_then(serde_json::Value::as_bool)
.unwrap_or(false);
let mut supports_embedding = global_model
.supported_capabilities
.as_ref()
.is_some_and(json_value_contains_embedding_metadata)
|| global_model
.config
.as_ref()
.is_some_and(json_value_contains_embedding_metadata);
for model in provider_models {
let Some(provider) = provider_ids.get(&model.provider_id) else {
@@ -143,9 +184,12 @@ pub(crate) async fn build_admin_model_catalog_payload(
admin_provider_model_effective_capability(&model, "function_calling");
let model_supports_streaming =
admin_provider_model_effective_capability(&model, "streaming");
let model_supports_embedding =
admin_provider_model_effective_capability(&model, "embedding");
supports_vision |= model_supports_vision;
supports_function_calling |= model_supports_function_calling;
supports_streaming |= model_supports_streaming;
supports_embedding |= model_supports_embedding;
providers.push(json!({
"provider_id": provider.id,
"provider_name": provider.name,
@@ -162,6 +206,7 @@ pub(crate) async fn build_admin_model_catalog_payload(
"supports_vision": model_supports_vision,
"supports_function_calling": model_supports_function_calling,
"supports_streaming": model_supports_streaming,
"supports_embedding": model_supports_embedding,
"is_active": model.is_active,
}));
}
@@ -189,6 +234,7 @@ pub(crate) async fn build_admin_model_catalog_payload(
"supports_vision": supports_vision,
"supports_function_calling": supports_function_calling,
"supports_streaming": supports_streaming,
"supports_embedding": supports_embedding,
}),
}));
}

View File

@@ -1,6 +1,13 @@
use crate::handlers::admin::shared::unix_secs_to_rfc3339;
use aether_data_contracts::repository::global_models::StoredAdminProviderModel;
const EMBEDDING_API_FORMATS: &[&str] = &[
"openai:embedding",
"jina:embedding",
"gemini:embedding",
"doubao:embedding",
];
pub(crate) fn model_tiered_pricing_first_tier_value(
tiered_pricing: Option<&serde_json::Value>,
field_name: &str,
@@ -26,6 +33,50 @@ fn model_effective_capability(
})
}
fn value_contains_string(value: &serde_json::Value, expected: &str) -> bool {
match value {
serde_json::Value::String(value) => value.trim().eq_ignore_ascii_case(expected),
serde_json::Value::Array(values) => values
.iter()
.any(|value| value_contains_string(value, expected)),
serde_json::Value::Object(object) => object
.values()
.any(|value| value_contains_string(value, expected)),
_ => false,
}
}
fn value_has_true_key(value: &serde_json::Value, key: &str) -> bool {
value
.as_object()
.and_then(|object| object.get(key))
.and_then(serde_json::Value::as_bool)
.unwrap_or(false)
}
fn value_contains_embedding_metadata(value: &serde_json::Value) -> bool {
value_has_true_key(value, "embedding")
|| value_contains_string(value, "embedding")
|| EMBEDDING_API_FORMATS
.iter()
.any(|api_format| value_contains_string(value, api_format))
}
fn model_effective_embedding_capability(model: &StoredAdminProviderModel) -> bool {
model
.config
.as_ref()
.is_some_and(value_contains_embedding_metadata)
|| model
.global_model_supported_capabilities
.as_ref()
.is_some_and(value_contains_embedding_metadata)
|| model
.global_model_config
.as_ref()
.is_some_and(value_contains_embedding_metadata)
}
pub(crate) fn timestamp_or_now(value: Option<u64>, now_unix_secs: u64) -> serde_json::Value {
unix_secs_to_rfc3339(value.unwrap_or(now_unix_secs))
.map(serde_json::Value::String)
@@ -110,6 +161,7 @@ pub(crate) fn admin_provider_model_effective_capability(
model.global_model_config.as_ref(),
"image_generation",
),
"embedding" => model_effective_embedding_capability(model),
_ => false,
}
}

View File

@@ -288,6 +288,7 @@ mod tests {
" OPENAI:RESPONSES ".to_string(),
"claude:messages".to_string(),
"gemini:generate_content".to_string(),
"jina:rerank".to_string(),
"openai:responses".to_string(),
]))
.expect("formats should normalize"),
@@ -295,6 +296,7 @@ mod tests {
"openai:responses".to_string(),
"claude:messages".to_string(),
"gemini:generate_content".to_string(),
"jina:rerank".to_string(),
])
);
}