feat: add hunyuan client (#980)

Argcfile.sh

@@ -88,6 +88,7 @@ OPENAI_COMPATIBLE_PLATFORMS=( \
   github,gpt-4o-mini,https://models.inference.ai.azure.com \
   groq,llama3-8b-8192,https://api.groq.com/openai/v1 \
   huggingface,meta-llama/Meta-Llama-3-8B-Instruct,https://api-inference.huggingface.co/v1 \
+  hunyuan,hunyuan-large,https://api.hunyuan.cloud.tencent.com/v1 \
   lingyiwanwu,yi-large,https://api.lingyiwanwu.com/v1 \
   mistral,open-mistral-nemo,https://api.mistral.ai/v1 \
   moonshot,moonshot-v1-8k,https://api.moonshot.cn/v1 \

config.example.yaml

@@ -255,12 +255,18 @@ clients:
     api_key: xxx
     secret_key: xxx
 
-  # See https://help.aliyun.com/zh/dashscope/
+  # See https://dashscope.aliyun.com/
   - type: openai-compatible
     name: qianwen
     api_base: https://dashscope.aliyuncs.com/compatible-mode/v1
     api_key: xxx
 
+  # See https://cloud.tencent.com/product/hunyuan
+  - type: openai-compatible
+    name: hunyuan
+    api_base: https://api.hunyuan.cloud.tencent.com/v1
+    api_key: xxx
+
   # See https://platform.moonshot.cn/docs/intro
   - type: openai-compatible
     name: moonshot

models.yaml

@@ -112,7 +112,6 @@
     - name: text-embedding-004
       type: embedding
       input_price: 0
-      output_price: 0
       max_tokens_per_chunk: 2048
       default_chunk_size: 1500
       max_batch_size: 100
@@ -811,6 +810,66 @@
       default_chunk_size: 2000
       max_batch_size: 25
 
+# links:
+#  - https://cloud.tencent.com/document/product/1729/104753
+#  - https://cloud.tencent.com/document/product/1729/97731
+#  - https://cloud.tencent.com/document/product/1729/111007
+- platform: hunyuan
+  models:
+    - name: hunyuan-turbo-latest
+      max_input_tokens: 28000
+      max_output_tokens: 4096
+      input_price: 2.1
+      output_price: 7.0
+      supports_function_calling: true
+    - name: hunyuan-pro
+      max_input_tokens: 28000
+      max_output_tokens: 4096
+      input_price: 4.2
+      output_price: 14.0
+      supports_function_calling: true
+    - name: hunyuan-large
+      max_input_tokens: 28000
+      max_output_tokens: 4096
+      supports_function_calling: true
+    - name: hunyuan-standard
+      max_input_tokens: 30000
+      max_output_tokens: 2048
+      input_price: 0.112
+      output_price: 0.28
+      supports_function_calling: true
+    - name: hunyuan-standard-256K
+      max_input_tokens: 250000
+      max_output_tokens: 6144
+      input_price: 0.07
+      output_price: 0.28
+      supports_function_calling: true
+    - name: hunyuan-lite
+      max_input_tokens: 250000
+      max_output_tokens: 6144
+      input_price: 0
+      output_price: 0
+      supports_function_calling: true
+    - name: hunyuan-functioncall
+      max_input_tokens: 28000
+      max_output_tokens: 4096
+      input_price: 0.56
+      output_price: 1.12
+      supports_function_calling: true
+    - name: hunyuan-vision
+      max_input_tokens: 6144
+      max_output_tokens: 2048
+      input_price: 2.52
+      output_price: 2.52
+      supports_function_calling: true
+      supports_vision: true
+    - name: hunyuan-embedding
+      type: embedding
+      input_price: 0.01
+      max_tokens_per_chunk: 1024
+      default_chunk_size: 1000
+      max_batch_size: 100
+
 # Links:
 #  - https://platform.moonshot.cn/docs/intro
 #  - https://platform.moonshot.cn/docs/pricing/chat

src/client/mod.rs

@@ -35,7 +35,7 @@ register_client!(
     (ernie, "ernie", ErnieConfig, ErnieClient),
 );
 
-pub const OPENAI_COMPATIBLE_PLATFORMS: [(&str, &str); 21] = [
+pub const OPENAI_COMPATIBLE_PLATFORMS: [(&str, &str); 22] = [
     ("ai21", "https://api.ai21.com/studio/v1"),
     ("cloudflare", ""),
     ("deepinfra", "https://api.deepinfra.com/v1/openai"),
@@ -44,6 +44,7 @@ pub const OPENAI_COMPATIBLE_PLATFORMS: [(&str, &str); 21] = [
     ("github", "https://models.inference.ai.azure.com"),
     ("groq", "https://api.groq.com/openai/v1"),
     ("huggingface", "https://api-inference.huggingface.co/v1"),
+    ("hunyuan", "https://api.hunyuan.cloud.tencent.com/v1"),
     ("lingyiwanwu", "https://api.lingyiwanwu.com/v1"),
     ("mistral", "https://api.mistral.ai/v1"),
     ("moonshot", "https://api.moonshot.cn/v1"),