@chenchaolong/plugin-vllm 0.0.5 → 0.0.7

package/dist/llm/llm.d.ts

@@ -10,6 +10,7 @@ export declare class VLLMLargeLanguageModel extends LargeLanguageModel {
     /**
      * Generate model schema from credentials for customizable models
      * This method dynamically generates parameter rules including thinking mode
+     * Merges parent class parameter rules (streaming, temperature, etc.) with thinking mode
      */
     getCustomizableModelSchemaFromCredentials(model: string, credentials: Record<string, any>): AIModelEntity | null;
 }

package/dist/llm/llm.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"llm.d.ts","sourceRoot":"","sources":["../../src/llm/llm.ts"],"names":[],"mappings":"AACA,OAAO,EACL,aAAa,EAGb,aAAa,EAId,MAAM,kBAAkB,CAAA;AAEzB,OAAO,EACL,2BAA2B,EAG3B,kBAAkB,EAClB,iBAAiB,EAClB,MAAM,sBAAsB,CAAA;AAE7B,OAAO,EAAE,oBAAoB,EAAE,MAAM,yBAAyB,CAAA;AAC9D,OAAO,EAAsB,oBAAoB,EAAE,MAAM,aAAa,CAAA;AAGtE,qBACa,sBAAuB,SAAQ,kBAAkB;;gBAGhD,aAAa,EAAE,oBAAoB;IAIzC,mBAAmB,CAAC,KAAK,EAAE,MAAM,EAAE,WAAW,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,CAAC;IAkBjF,YAAY,CAAC,YAAY,EAAE,aAAa,EAAE,OAAO,CAAC,EAAE,iBAAiB;IA6B9E;;;OAGG;IACM,yCAAyC,CAChD,KAAK,EAAE,MAAM,EACb,WAAW,EAAE,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC,GAC/B,aAAa,GAAG,IAAI;CAwExB"}
+{"version":3,"file":"llm.d.ts","sourceRoot":"","sources":["../../src/llm/llm.ts"],"names":[],"mappings":"AACA,OAAO,EACL,aAAa,EAGb,aAAa,EAId,MAAM,kBAAkB,CAAA;AAEzB,OAAO,EACL,2BAA2B,EAG3B,kBAAkB,EAClB,iBAAiB,EAClB,MAAM,sBAAsB,CAAA;AAE7B,OAAO,EAAE,oBAAoB,EAAE,MAAM,yBAAyB,CAAA;AAC9D,OAAO,EAAsB,oBAAoB,EAAE,MAAM,aAAa,CAAA;AAGtE,qBACa,sBAAuB,SAAQ,kBAAkB;;gBAGhD,aAAa,EAAE,oBAAoB;IAIzC,mBAAmB,CAAC,KAAK,EAAE,MAAM,EAAE,WAAW,EAAE,oBAAoB,GAAG,OAAO,CAAC,IAAI,CAAC;IAkBjF,YAAY,CAAC,YAAY,EAAE,aAAa,EAAE,OAAO,CAAC,EAAE,iBAAiB;IA+C9E;;;;OAIG;IACM,yCAAyC,CAChD,KAAK,EAAE,MAAM,EACb,WAAW,EAAE,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC,GAC/B,aAAa,GAAG,IAAI;CA0FxB"}

package/dist/llm/llm.js

@@ -39,8 +39,24 @@ let VLLMLargeLanguageModel = VLLMLargeLanguageModel_1 = class VLLMLargeLanguageM
             throw new Error(translate('Error.ModelCredentialsMissing', { model: copilotModel.model }));
         }
         const params = toCredentialKwargs(modelProperties, copilotModel.model);
+        // Get thinking parameter from model options (runtime parameter)
+        // This takes priority over the default value in credentials
+        const modelOptions = copilotModel.options;
+        const thinking = modelOptions?.thinking ?? modelProperties?.thinking ?? false;
+        // Merge modelKwargs with thinking parameter
+        // Ensure chat_template_kwargs structure is correct for vLLM API
+        const existingModelKwargs = (params.modelKwargs || {});
+        const existingChatTemplateKwargs = existingModelKwargs.chat_template_kwargs || {};
+        const modelKwargs = {
+            ...existingModelKwargs,
+            chat_template_kwargs: {
+                ...existingChatTemplateKwargs,
+                enable_thinking: !!thinking
+            }
+        };
         const fields = omitBy({
             ...params,
+            modelKwargs,
             streaming: copilotModel.options?.['streaming'] ?? true,
             // include token usage in the stream. this will include an additional chunk at the end of the stream with the token usage.
             streamUsage: true
@@ -57,12 +73,16 @@ let VLLMLargeLanguageModel = VLLMLargeLanguageModel_1 = class VLLMLargeLanguageM
     /**
      * Generate model schema from credentials for customizable models
      * This method dynamically generates parameter rules including thinking mode
+     * Merges parent class parameter rules (streaming, temperature, etc.) with thinking mode
      */
     getCustomizableModelSchemaFromCredentials(model, credentials) {
-        const rules = [];
+        // Get parent class parameter rules (streaming and temperature)
+        // This ensures we include common parameters from the base class
+        const parentSchema = super.getCustomizableModelSchemaFromCredentials(model, credentials);
+        const parentRules = parentSchema?.parameter_rules || [];
         // Add thinking mode parameter
         // This parameter enables thinking mode for models deployed on vLLM and SGLang
-        rules.push({
+        const thinkingRule = {
             name: 'thinking',
             type: ParameterType.BOOLEAN,
             label: {
@@ -75,7 +95,13 @@ let VLLMLargeLanguageModel = VLLMLargeLanguageModel_1 = class VLLMLargeLanguageM
             },
             required: false,
             default: credentials['thinking'] ?? false
-        });
+        };
+        // Merge parent rules with thinking rule
+        // Filter out any duplicate rules by name to ensure thinking rule takes precedence
+        const rules = [
+            ...parentRules,
+            thinkingRule
+        ].filter((rule, index, self) => index === self.findIndex(r => r.name === rule.name));
         // Determine completion type from credentials
         let completionType = 'chat';
         if (credentials['mode']) {
@@ -120,7 +146,13 @@ let VLLMLargeLanguageModel = VLLMLargeLanguageModel_1 = class VLLMLargeLanguageM
                 [ModelPropertyKey.MODE]: completionType,
                 [ModelPropertyKey.CONTEXT_SIZE]: contextSize
             },
-            parameter_rules: rules
+            parameter_rules: rules,
+            pricing: parentSchema?.pricing || {
+                input: credentials['input_price'] ?? 0,
+                output: credentials['output_price'] ?? 0,
+                unit: credentials['unit'] ?? 0,
+                currency: credentials['currency'] ?? 'USD'
+            }
         };
     }
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@chenchaolong/plugin-vllm",
-  "version": "0.0.5",
+  "version": "0.0.7",
   "author": {
     "name": "XpertAI",
     "url": "https://xpertai.cn"