Merge pull request #4846 from continuedev/nate/llm-info-fix

llm-info fix

Author: sestinj

core/llm/index.ts

@@ -177,7 +177,11 @@ export abstract class BaseLLM implements ILLM {
 
     this.model = options.model;
     // Use @continuedev/llm-info package to autodetect certain parameters
-    const llmInfo = findLlmInfo(this.model);
+    const modelSearchString =
+      this.providerName === "continue-proxy"
+        ? this.model?.split("/").pop() || this.model
+        : this.model;
+    const llmInfo = findLlmInfo(modelSearchString);
 
     const templateType =
       options.template ?? autodetectTemplateType(options.model);
@@ -195,11 +199,11 @@ export abstract class BaseLLM implements ILLM {
         options.completionOptions?.maxTokens ??
         (llmInfo?.maxCompletionTokens
           ? Math.min(
-            llmInfo.maxCompletionTokens,
-            // Even if the model has a large maxTokens, we don't want to use that every time,
-            // because it takes away from the context length
-            this.contextLength / 4,
-          )
+              llmInfo.maxCompletionTokens,
+              // Even if the model has a large maxTokens, we don't want to use that every time,
+              // because it takes away from the context length
+              this.contextLength / 4,
+            )
           : DEFAULT_MAX_TOKENS),
     };
     this.requestOptions = options.requestOptions;
@@ -840,7 +844,6 @@ export abstract class BaseLLM implements ILLM {
             signal,
             completionOptions,
           )) {
-
             if (chunk.role === "assistant") {
               completion += chunk.content;
               yield chunk;
@@ -948,15 +951,15 @@ export abstract class BaseLLM implements ILLM {
     );
   }
 
-  protected async * _streamComplete(
+  protected async *_streamComplete(
     prompt: string,
     signal: AbortSignal,
     options: CompletionOptions,
   ): AsyncGenerator<string> {
     throw new Error("Not implemented");
   }
 
-  protected async * _streamChat(
+  protected async *_streamChat(
     messages: ChatMessage[],
     signal: AbortSignal,
     options: CompletionOptions,

core/package-lock.json

@@ -49,7 +49,7 @@
     "@continuedev/config-types": "^1.0.13",
     "@continuedev/config-yaml": "^1.0.67",
     "@continuedev/fetch": "^1.0.4",
-    "@continuedev/llm-info": "^1.0.2",
+    "@continuedev/llm-info": "^1.0.8",
     "@continuedev/openai-adapters": "^1.0.18",
     "@modelcontextprotocol/sdk": "^1.5.0",
     "@mozilla/readability": "^0.5.0",

core/package.json

@@ -3,7 +3,7 @@ pluginGroup=com.github.continuedev.continueintellijextension
 pluginName=continue-intellij-extension
 pluginRepositoryUrl=https://github.com/continuedev/continue
 # SemVer format -> https://semver.org
-pluginVersion=1.0.7
+pluginVersion=1.0.8
 # Supported build number ranges and IntelliJ Platform versions -> https://plugins.jetbrains.com/docs/intellij/build-number-ranges.html
 pluginSinceBuild=223
 # IntelliJ Platform Properties -> https://plugins.jetbrains.com/docs/intellij/tools-gradle-intellij-plugin.html#configuration-intellij-extension

extensions/intellij/gradle.properties

@@ -2,7 +2,7 @@
   "name": "continue",
   "icon": "media/icon.png",
   "author": "Continue Dev, Inc",
-  "version": "1.1.16",
+  "version": "1.1.17",
   "repository": {
     "type": "git",
     "url": "https://github.com/continuedev/continue"

extensions/vscode/package-lock.json

@@ -1,6 +1,6 @@
 {
   "name": "@continuedev/llm-info",
-  "version": "1.0.7",
+  "version": "1.0.8",
   "description": "",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",

extensions/vscode/package.json

@@ -11,7 +11,17 @@ export const Anthropic: ModelProvider = {
       maxCompletionTokens: 8192,
       description:
         "Most intelligent model with the highest level of intelligence and capability.",
-      regex: /claude-3\.5-sonnet/i,
+      regex: /claude-3[.-]5-sonnet.*/i,
+      recommendedFor: ["chat"],
+    },
+    {
+      model: "claude-3-7-sonnet-latest",
+      displayName: "Claude 3.7 Sonnet",
+      contextLength: 200000,
+      maxCompletionTokens: 8192,
+      description:
+        "Most intelligent model with the highest level of intelligence and capability.",
+      regex: /claude-3[.-]7-sonnet.*/i,
       recommendedFor: ["chat"],
     },
     {