refactor(ai-proxy): redesign streaming thinking promotion to buffer-and-flush

Instead of promoting reasoning to content inline per-chunk (which would emit reasoning as content prematurely if real content arrives later), the streaming path now buffers reasoning content and strips it from chunks. On the last chunk, if no content was ever seen, the buffered reasoning is flushed as a single content chunk. Also moves tests into test/openai.go TestOpenAI suite and adds MockHttpContext for provider-level streaming tests. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
feat(ai-proxy): add promoteThinkingOnEmpty and hiclawMode config options
2026-03-21 19:47:52 +08:00 · 2026-03-20 00:05:17 +08:00 · 2026-03-19 23:28:17 +08:00
4 changed files with 3 additions and 5 deletions
--- a/go.mod
+++ b/go.mod
@@ -31,7 +31,7 @@ require (
 	github.com/hudl/fargo v1.4.0
 	github.com/mholt/acmez v1.2.0
 	github.com/nacos-group/nacos-sdk-go v1.0.8
-	github.com/nacos-group/nacos-sdk-go/v2 v2.3.5
+	github.com/nacos-group/nacos-sdk-go/v2 v2.3.2
 	github.com/spf13/cobra v1.9.1
 	github.com/spf13/pflag v1.0.7
 	github.com/stretchr/testify v1.11.1
--- a/go.sum
+++ b/go.sum
@@ -3688,8 +3688,8 @@ github.com/mxk/go-flowrate v0.0.0-20140419014527-cca7078d478f h1:y5//uYreIhSUg3J
 github.com/mxk/go-flowrate v0.0.0-20140419014527-cca7078d478f/go.mod h1:ZdcZmHo+o7JKHSa8/e818NopupXU1YMK5fe1lsApnBw=
 github.com/nacos-group/nacos-sdk-go v1.0.8 h1:8pEm05Cdav9sQgJSv5kyvlgfz0SzFUUGI3pWX6SiSnM=
 github.com/nacos-group/nacos-sdk-go v1.0.8/go.mod h1:hlAPn3UdzlxIlSILAyOXKxjFSvDJ9oLzTJ9hLAK1KzA=
-github.com/nacos-group/nacos-sdk-go/v2 v2.3.5 h1:Hux7C4N4rWhwBF5Zm4yyYskrs9VTgrRTA8DZjoEhQTs=
-github.com/nacos-group/nacos-sdk-go/v2 v2.3.5/go.mod h1:ygUBdt7eGeYBt6Lz2HO3wx7crKXk25Mp80568emGMWU=
+github.com/nacos-group/nacos-sdk-go/v2 v2.3.2 h1:9QB2nCJzT5wkTVlxNYl3XL/7+G6p2USMi2gQh/ouQQo=
+github.com/nacos-group/nacos-sdk-go/v2 v2.3.2/go.mod h1:9FKXl6FqOiVmm72i8kADtbeK71egyG9y3uRDBg41tpQ=
 github.com/nats-io/jwt v0.3.0/go.mod h1:fRYCDE99xlTsqUzISS1Bi75UBJ6ljOJQOAAu5VglpSg=
 github.com/nats-io/jwt v0.3.2/go.mod h1:/euKqTS1ZD+zzjYrY7pseZrTtWQSjujC7xjPc8wL6eU=
 github.com/nats-io/nats-server/v2 v2.1.2/go.mod h1:Afk+wRZqkMQs/p45uXdrVLuab3gwv3Z8C4HTBu8GD/k=
--- a/plugins/wasm-go/extensions/model-mapper/main.go
+++ b/plugins/wasm-go/extensions/model-mapper/main.go
@@ -112,7 +112,6 @@ func parseConfig(json gjson.Result, config *Config) error {
 			"/video-synthesis",
 			"/rerank",
 			"/messages",
-			"/responses",
 		}
 	}

--- a/plugins/wasm-go/extensions/model-router/main.go
+++ b/plugins/wasm-go/extensions/model-router/main.go
@@ -80,7 +80,6 @@ func parseConfig(json gjson.Result, config *ModelRouterConfig) error {
 			"/video-synthesis",
 			"/rerank",
 			"/messages",
-			"/responses",
 		}
 	}