Merge pull request #72 from lanej/feat/thinking-blocks-support

fix: handle Anthropic thinking blocks to prevent 400 errors

Author: ericc-ch

src/routes/messages/non-stream-translation.ts

@@ -15,6 +15,7 @@ import {
   type AnthropicMessagesPayload,
   type AnthropicResponse,
   type AnthropicTextBlock,
+  type AnthropicThinkingBlock,
   type AnthropicTool,
   type AnthropicToolResultBlock,
   type AnthropicToolUseBlock,
@@ -131,11 +132,21 @@ function handleAssistantMessage(
     (block): block is AnthropicTextBlock => block.type === "text",
   )
 
+  const thinkingBlocks = message.content.filter(
+    (block): block is AnthropicThinkingBlock => block.type === "thinking",
+  )
+
+  // Combine text and thinking blocks, as OpenAI doesn't have separate thinking blocks
+  const allTextContent = [
+    ...textBlocks.map((b) => b.text),
+    ...thinkingBlocks.map((b) => b.thinking),
+  ].join("\n\n")
+
   return toolUseBlocks.length > 0 ?
       [
         {
           role: "assistant",
-          content: textBlocks.map((b) => b.text).join("\n\n") || null,
+          content: allTextContent || null,
           tool_calls: toolUseBlocks.map((toolUse) => ({
             id: toolUse.id,
             type: "function",
@@ -169,22 +180,38 @@ function mapContent(
   const hasImage = content.some((block) => block.type === "image")
   if (!hasImage) {
     return content
-      .filter((block): block is AnthropicTextBlock => block.type === "text")
-      .map((block) => block.text)
+      .filter(
+        (block): block is AnthropicTextBlock | AnthropicThinkingBlock =>
+          block.type === "text" || block.type === "thinking",
+      )
+      .map((block) => (block.type === "text" ? block.text : block.thinking))
       .join("\n\n")
   }
 
   const contentParts: Array<ContentPart> = []
   for (const block of content) {
-    if (block.type === "text") {
-      contentParts.push({ type: "text", text: block.text })
-    } else if (block.type === "image") {
-      contentParts.push({
-        type: "image_url",
-        image_url: {
-          url: `data:${block.source.media_type};base64,${block.source.data}`,
-        },
-      })
+    switch (block.type) {
+      case "text": {
+        contentParts.push({ type: "text", text: block.text })
+
+        break
+      }
+      case "thinking": {
+        contentParts.push({ type: "text", text: block.thinking })
+
+        break
+      }
+      case "image": {
+        contentParts.push({
+          type: "image_url",
+          image_url: {
+            url: `data:${block.source.media_type};base64,${block.source.data}`,
+          },
+        })
+
+        break
+      }
+      // No default
     }
   }
   return contentParts
@@ -246,6 +273,7 @@ export function translateToAnthropic(
   const choice = response.choices[0]
   const textBlocks = getAnthropicTextBlocks(choice.message.content)
   const toolUseBlocks = getAnthropicToolUseBlocks(choice.message.tool_calls)
+  // Note: GitHub Copilot doesn't generate thinking blocks, so we don't include them in responses
 
   return {
     id: response.id,

tests/anthropic-request.test.ts

@@ -124,6 +124,79 @@ describe("Anthropic to OpenAI translation logic", () => {
     // Should fail validation
     expect(isValidChatCompletionRequest(openAIPayload)).toBe(false)
   })
+
+  test("should handle thinking blocks in assistant messages", () => {
+    const anthropicPayload: AnthropicMessagesPayload = {
+      model: "claude-3-5-sonnet-20241022",
+      messages: [
+        { role: "user", content: "What is 2+2?" },
+        {
+          role: "assistant",
+          content: [
+            {
+              type: "thinking",
+              thinking: "Let me think about this simple math problem...",
+            },
+            { type: "text", text: "2+2 equals 4." },
+          ],
+        },
+      ],
+      max_tokens: 100,
+    }
+    const openAIPayload = translateToOpenAI(anthropicPayload)
+    expect(isValidChatCompletionRequest(openAIPayload)).toBe(true)
+
+    // Check that thinking content is combined with text content
+    const assistantMessage = openAIPayload.messages.find(
+      (m) => m.role === "assistant",
+    )
+    expect(assistantMessage?.content).toContain(
+      "Let me think about this simple math problem...",
+    )
+    expect(assistantMessage?.content).toContain("2+2 equals 4.")
+  })
+
+  test("should handle thinking blocks with tool calls", () => {
+    const anthropicPayload: AnthropicMessagesPayload = {
+      model: "claude-3-5-sonnet-20241022",
+      messages: [
+        { role: "user", content: "What's the weather?" },
+        {
+          role: "assistant",
+          content: [
+            {
+              type: "thinking",
+              thinking:
+                "I need to call the weather API to get current weather information.",
+            },
+            { type: "text", text: "I'll check the weather for you." },
+            {
+              type: "tool_use",
+              id: "call_123",
+              name: "get_weather",
+              input: { location: "New York" },
+            },
+          ],
+        },
+      ],
+      max_tokens: 100,
+    }
+    const openAIPayload = translateToOpenAI(anthropicPayload)
+    expect(isValidChatCompletionRequest(openAIPayload)).toBe(true)
+
+    // Check that thinking content is included in the message content
+    const assistantMessage = openAIPayload.messages.find(
+      (m) => m.role === "assistant",
+    )
+    expect(assistantMessage?.content).toContain(
+      "I need to call the weather API",
+    )
+    expect(assistantMessage?.content).toContain(
+      "I'll check the weather for you.",
+    )
+    expect(assistantMessage?.tool_calls).toHaveLength(1)
+    expect(assistantMessage?.tool_calls?.[0].function.name).toBe("get_weather")
+  })
 })
 
 describe("OpenAI Chat Completion v1 Request Payload Validation with Zod", () => {