ZacharyZhang-NY/Kigi-CLI
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

feat(providers): add Qwen Token Plan (global + China)

Browse Source 
Providers 14-15 (17th & 18th registry variants), Alibaba DashScope
compatible-mode. Global: token-plan.ap-southeast-1.maas.aliyuncs.com,
QWEN_TOKEN_PLAN_API_KEY, models.dev alibaba-token-plan. China:
token-plan.cn-beijing.maas.aliyuncs.com, QWEN_TOKEN_PLAN_CN_API_KEY,
alibaba-token-plan-cn. Both Bearer + OpenAI listing + ChatCompletions +
Passthrough (stream_options.include_usage is documented-supported).

/models is auth-gated (401 without a key) so it doubles as the validator;
metadata from models.dev enrichment. The token plan is a multi-vendor
catalog (deepseek/kimi/minimax/glm/qwen); restrict_to_enriched keeps the 15
tool-calling chat models and drops the 4 qwen-image/wan image generators.

Review downgraded two flagged concerns: the enable_thinking non-streaming
400 cannot occur (Kigi never issues non-streaming ChatCompletions in
production — all inference streams), and Qwen thinking is NOT invisible
(reasoning_content is parsed regardless of the Passthrough dialect). No
defects. Residual (logged): restrict does an exact id-match of live /models
ids vs the models.dev keys; a mismatch fails safe (0 models) — verify with a
real key. Snapshot already bundles both providers (regenerated in ebf1105).

Tests: global e2e proves enrichment-supplied context (wire carries none),
non-vacuous tool_call restriction (qwen-image dropped), bare-id round-trip,
Passthrough; both variants' validation tests hit /models (401 reject) and
assert the correct per-variant console host. Registry at 18; picker 19 rows.
This commit is contained in:
ZacharyZhang-NY
2026-07-21 16:54:35 -04:00
parent ebf11057f8
commit 8245deb373
4 changed files with 241 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/codegen/kigi-shell/src/agent/auth_method.rs
+64 -4
View File
@@ -620,7 +620,9 @@ mod tests {
"cerebras",
"nvidia",
"vercel-ai-gateway",
"xai"
"xai",
"qwen-token-plan",
"qwen-token-plan-cn"
]
);
assert_eq!(default_id(&built), Some(XAI_API_KEY_METHOD_ID));
@@ -659,7 +661,9 @@ mod tests {
"cerebras",
"nvidia",
"vercel-ai-gateway",
"xai"
"xai",
"qwen-token-plan",
"qwen-token-plan-cn"
]
);
assert_eq!(default_id(&built), Some(CACHED_TOKEN_AUTH_METHOD_ID));
@@ -691,7 +695,9 @@ mod tests {
"cerebras",
"nvidia",
"vercel-ai-gateway",
"xai"
"xai",
"qwen-token-plan",
"qwen-token-plan-cn"
]
);
assert_eq!(default_id(&built), Some(CACHED_TOKEN_AUTH_METHOD_ID));
@@ -726,7 +732,9 @@ mod tests {
"cerebras",
"nvidia",
"vercel-ai-gateway",
"xai"
"xai",
"qwen-token-plan",
"qwen-token-plan-cn"
]
);
assert_eq!(default_id(&built), None);
@@ -1006,4 +1014,56 @@ mod tests {
.await
.expect("200 from /models must validate the key");
}
/// Qwen Token Plan (global): DashScope /models is auth-gated (401 for a bad
/// key), so it validates the key; the error names the Model Studio console.
#[tokio::test]
#[serial]
async fn qwen_token_plan_validates_against_models_and_rejects_bad_key() {
use wiremock::matchers::{method, path};
let server = wiremock::MockServer::start().await;
wiremock::Mock::given(method("GET"))
.and(path("/models"))
.respond_with(wiremock::ResponseTemplate::new(401))
.expect(1)
.mount(&server)
.await;
let _base = EnvGuard::set(kigi_models::QWEN_TOKEN_PLAN_BASE_URL_ENV, &server.uri());
let err =
authenticate_platform_api_key(kigi_models::PlatformId::QwenTokenPlan, Some("qtp-bad"))
.await
.expect_err("a 401 from /models must reject the key");
assert_eq!(
err.message,
"Invalid API key for qwen-token-plan \u{2014} check your key on \
modelstudio.console.alibabacloud.com"
);
}
/// Qwen Token Plan (China): distinct base URL + console host; same
/// auth-gated /models validation.
#[tokio::test]
#[serial]
async fn qwen_token_plan_cn_validates_against_models_and_rejects_bad_key() {
use wiremock::matchers::{method, path};
let server = wiremock::MockServer::start().await;
wiremock::Mock::given(method("GET"))
.and(path("/models"))
.respond_with(wiremock::ResponseTemplate::new(401))
.expect(1)
.mount(&server)
.await;
let _base = EnvGuard::set(kigi_models::QWEN_TOKEN_PLAN_CN_BASE_URL_ENV, &server.uri());
let err = authenticate_platform_api_key(
kigi_models::PlatformId::QwenTokenPlanCn,
Some("qtp-cn-bad"),
)
.await
.expect_err("a 401 from /models must reject the key");
assert_eq!(
err.message,
"Invalid API key for qwen-token-plan-cn \u{2014} check your key on \
bailian.console.aliyun.com"
);
}
}
crates/codegen/kigi-shell/src/agent/models_fetch.rs
+91
View File
@@ -2084,6 +2084,97 @@ mod tests {
);
}
/// Qwen-Token-Plan e2e: DashScope compatible-mode /models is auth-gated and
/// minimal (ids only), so enrichment supplies context; the restriction drops
/// the non-tool-calling qwen-image / wan generators the token plan lists;
/// bare id round-trips under the platform key; Passthrough dialect.
#[tokio::test(flavor = "multi_thread")]
#[serial_test::serial]
async fn qwen_token_plan_enriches_restricts_and_maps_passthrough() {
let platform_server = wiremock::MockServer::start().await;
wiremock::Mock::given(wiremock::matchers::method("GET"))
.and(wiremock::matchers::path("/models"))
.respond_with(wiremock::ResponseTemplate::new(200).set_body_json(
// DashScope /models is minimal: id/object only, NO context.
serde_json::json!({ "data": [
{ "id": "qwen3.7-max", "object": "model" },
{ "id": "qwen-image-2.0", "object": "model" }
]}),
))
.expect(1)
.mount(&platform_server)
.await;
let modelsdev_server = wiremock::MockServer::start().await;
wiremock::Mock::given(wiremock::matchers::method("GET"))
.and(wiremock::matchers::path("/api.json"))
.respond_with(wiremock::ResponseTemplate::new(200).set_body_json(
serde_json::json!({ "alibaba-token-plan": { "models": {
"qwen3.7-max": {
"limit": {"context": 1000000, "output": 32768},
"tool_call": true
},
// present in enrichment too, but not tool-calling → dropped.
"qwen-image-2.0": { "limit": {"context": 8192} }
}}}),
))
.expect(1)
.mount(&modelsdev_server)
.await;
let cache_dir = tempfile::tempdir().unwrap();
let _base = kigi_test_support::EnvGuard::set(
kigi_models::QWEN_TOKEN_PLAN_BASE_URL_ENV,
platform_server.uri(),
);
let _mdev = kigi_test_support::EnvGuard::set(
crate::agent::enrichment_fetch::MODELS_DEV_URL_ENV,
format!("{}/api.json", modelsdev_server.uri()),
);
let _mdev_cache = kigi_test_support::EnvGuard::set(
crate::agent::enrichment_fetch::MODELS_DEV_CACHE_DIR_ENV,
cache_dir.path(),
);
let endpoints = crate::agent::config::EndpointsConfig::default();
let keys = crate::agent::models::PlatformApiKeys::test_single(
kigi_models::PlatformId::QwenTokenPlan,
"qtp-1",
);
let result = tokio::task::spawn_blocking(move || {
fetch_platform_models_blocking(&endpoints, None, &keys)
})
.await
.unwrap()
.expect("fetch must succeed");
assert_eq!(
result
.models
.iter()
.map(|m| m.id.as_deref().unwrap_or_default())
.collect::<Vec<_>>(),
vec!["qwen-token-plan/qwen3.7-max"],
"qwen-image (not tool-calling) dropped; bare id kept under the platform key"
);
let entry = &result.models[0];
assert_eq!(
entry.context_window.get(),
1_000_000,
"context comes from enrichment (the wire listing carries none)"
);
assert_eq!(entry.max_completion_tokens, Some(32_768));
assert_eq!(entry.model, "qwen3.7-max", "the bare id rides the wire");
let model_entry = crate::agent::config::ModelEntry::from_config_entry(entry);
let creds = crate::agent::config::ResolvedCredentials {
api_key: Some("qtp-1".into()),
base_url: entry.base_url.clone(),
auth_type: kigi_chat_state::AuthType::ApiKey,
auth_scheme: Default::default(),
};
let cfg = crate::agent::config::sampling_config_for_model(&model_entry, creds, None);
assert_eq!(
cfg.chat_compat,
kigi_sampling_types::ChatCompat::Passthrough
);
}
#[test]
fn get_env_keys_parses_strings_and_rejects_non_strings() {
use crate::agent::config::EnvKeys;