adapters: 支持多模态（对照 llmsproxy）

- ollama: content 数组拆分文本/images(base64)，映射到 Ollama messages.images；
  流式透传 reasoning_content 与增量 tool_calls
- anthropic: OpenAI content(字符串/数组) 转 Anthropic blocks(文本 + image{source})
  + thinking/空 arguments tool_use/input_json_delta 流
- openai 透传 adapter 已保留多模态 content 数组（验收确认）
- 新增多模态单测：ollama images / anthropic image.source / openai passthrough

验证: go test ./... 27 包 0 失败；Windows 交叉编译通过；部署后服务健康

internal/lua/adapters/anthropic.lua

@@ -11,13 +11,42 @@ function adapter.transform_request(raw_body)
     local ok, req = pcall(json.decode, raw_body)
     if not ok then return raw_body end
 
+    -- 将 OpenAI 风格 content（字符串或 [{type:*}] 数组）拆成文本/图片块
+    local function collect_blocks(content)
+        if type(content) == "string" then
+            return { { type = "text", text = content } }
+        end
+        local blocks = {}
+        for _, p in ipairs(content or {}) do
+            if p.type == "text" then
+                table.insert(blocks, { type = "text", text = p.text })
+            elseif p.type == "image_url" and type(p.image_url) == "table" and p.image_url.url then
+                local mt, b64 = string.match(p.image_url.url, "^data:([^,]+);base64,(.+)$")
+                if b64 then
+                    table.insert(blocks, { type = "image", source = { type = "base64", media_type = mt or "image/png", data = b64 } })
+                else
+                    table.insert(blocks, { type = "image", source = { type = "url", url = p.image_url.url } })
+                end
+            end
+        end
+        return blocks
+    end
+    local function text_of(content)
+        if type(content) == "string" then return content end
+        local t = ""
+        for _, p in ipairs(content or {}) do
+            if p.type == "text" and p.text then t = t .. p.text end
+        end
+        return t
+    end
+
     local msgs = {}
     local system = ""
     for _, m in ipairs(req.messages or {}) do
         if m.role == "system" then
-            system = system .. m.content .. "\n"
+            system = system .. text_of(m.content) .. "\n"
         else
-            table.insert(msgs, { role = m.role, content = m.content })
+            table.insert(msgs, { role = m.role, content = collect_blocks(m.content) })
         end
     end
 
@@ -74,12 +103,41 @@ function adapter.transform_stream_chunk(raw_chunk)
     if chunk.type == "message_delta" then
         return json.encode({ content = "", done = (chunk.delta and chunk.delta.stop_reason ~= nil) })
     end
+    if chunk.type == "content_block_start" and chunk.content_block
+        and chunk.content_block.type == "tool_use" then
+        -- first fragment of a tool call: emit index + id + name, empty args
+        return json.encode({
+            content = "", done = false,
+            tool_calls = { {
+                index = chunk.index or 0,
+                id = chunk.content_block.id or "",
+                type = "function",
+                ["function"] = { name = chunk.content_block.name or "", arguments = "" }
+            } }
+        })
+    end
     if chunk.type == "content_block_delta" and chunk.delta then
+        if chunk.delta.type == "input_json_delta" then
+            -- incremental JSON fragment; clients accumulate across chunks
+            local unified = { content = "", done = false, tool_calls = { {
+                index = chunk.index or 0,
+                id = "",
+                type = "function",
+                ["function"] = { name = "", arguments = chunk.delta.partial_json or "" }
+            } } }
+            return json.encode(unified)
+        end
+        if chunk.delta.type == "thinking_delta" and chunk.delta.thinking then
+            return json.encode({ content = "", done = false, reasoning_content = chunk.delta.thinking })
+        end
         return json.encode({ content = chunk.delta.text or "", done = false })
     end
     if chunk.type == "message_stop" then
         return json.encode({ content = "", done = true })
     end
+    if chunk.type == "content_block_stop" then
+        return json.encode({ content = "", done = false })
+    end
     return ""
 end
 

internal/lua/adapters/ollama.lua

@@ -19,11 +19,29 @@ function adapter.transform_request(raw_body)
         }
     }
 
-    -- 转换 messages 格式（Ollama 兼容 OpenAI 的 messages 格式）
+    -- 转换 messages 格式（Ollama messages 支持 images base64 数组）
     if req.messages then
         local msgs = {}
         for _, m in ipairs(req.messages) do
-            table.insert(msgs, { role = m.role, content = m.content })
+            local text, images
+            if type(m.content) == "string" then
+                text, images = m.content, nil
+            else
+                text = ""
+                images = {}
+                for _, p in ipairs(m.content or {}) do
+                    if p.type == "text" then
+                        text = text .. (p.text or "")
+                    elseif p.type == "image_url" and type(p.image_url) == "table" and p.image_url.url then
+                        local b64 = string.match(p.image_url.url, "^data:[^,]+;base64,(.+)$")
+                        if b64 then table.insert(images, b64) end
+                    end
+                end
+                if #images == 0 then images = nil end
+            end
+            local msg = { role = m.role, content = text }
+            if images then msg.images = images end
+            table.insert(msgs, msg)
         end
         ollama_req.messages = msgs
     end
@@ -54,10 +72,29 @@ function adapter.transform_stream_chunk(raw_chunk)
     if not ok then return "" end
     if not chunk.message then return "" end
 
-    return json.encode({
+    local unified = {
         content = chunk.message.content or "",
         done = chunk.done or false
-    })
+    }
+    if chunk.message.reasoning_content then
+        unified.reasoning_content = chunk.message.reasoning_content
+    end
+    if chunk.message.tool_calls then
+        local tools = {}
+        for _, tc in ipairs(chunk.message.tool_calls) do
+            table.insert(tools, {
+                index = #tools,
+                id = tc.id or ("call_" .. #tools),
+                type = "function",
+                ["function"] = {
+                    name = tc["function"] and tc["function"].name or "",
+                    arguments = tc["function"] and (tc["function"].arguments or "{}") or "{}"
+                }
+            })
+        end
+        unified.tool_calls = tools
+    end
+    return json.encode(unified)
 end
 
 return adapter

internal/lua/vm_test.go

@@ -135,3 +135,58 @@ func TestVMConcurrentCalls(t *testing.T) {
 		t.Fatalf("concurrent call: %v", err)
 	}
 }
+func TestOllamaMultimodal(t *testing.T) {
+	vm := NewVM(t.TempDir())
+	if err := vm.Start(); err != nil {
+		t.Fatalf("start vm: %v", err)
+	}
+	defer vm.Stop()
+
+	raw := `{"model":"llama3","messages":[{"role":"user","content":"hi"},{"role":"user","content":[{"type":"text","text":"look"},{"type":"image_url","image_url":{"url":"data:image/png;base64,AAAA"}}]}]}`
+	out, err := vm.CallTransformRequest("ollama", raw)
+	if err != nil {
+		t.Fatalf("transform_request: %v", err)
+	}
+	if !strings.Contains(out, `"images":["AAAA"]`) {
+		t.Fatalf("expected base64 images array, got: %s", out)
+	}
+	if !strings.Contains(out, `"content":"look"`) {
+		t.Fatalf("text not preserved: %s", out)
+	}
+}
+
+func TestAnthropicMultimodal(t *testing.T) {
+	vm := NewVM(t.TempDir())
+	if err := vm.Start(); err != nil {
+		t.Fatalf("start vm: %v", err)
+	}
+	defer vm.Stop()
+
+	raw := `{"model":"claude-sonnet-4-20250514","messages":[{"role":"user","content":[{"type":"text","text":"what is this"},{"type":"image_url","image_url":{"url":"data:image/png;base64,AAEC"}}]}]}`
+	out, err := vm.CallTransformRequest("anthropic", raw)
+	if err != nil {
+		t.Fatalf("transform_request: %v", err)
+	}
+	if !strings.Contains(out, `"type":"image"`) || !strings.Contains(out, `"media_type":"image/png"`) || !strings.Contains(out, `"data":"AAEC"`) {
+		t.Fatalf("expected anthropic image source block, got: %s", out)
+	}
+	if !strings.Contains(out, `"type":"text"`) {
+		t.Fatalf("expected text block preserved: %s", out)
+	}
+}
+
+func TestOpenAIPassthroughKeepsMultimodal(t *testing.T) {
+	vm := NewVM(t.TempDir())
+	if err := vm.Start(); err != nil {
+		t.Fatalf("start vm: %v", err)
+	}
+	defer vm.Stop()
+	raw := `{"model":"auto","messages":[{"role":"user","content":[{"type":"text","text":"hi"},{"type":"image_url","image_url":{"url":"data:image/png;base64,QUJD"}}]}]}`
+	out, err := vm.CallTransformRequest("openai", raw)
+	if err != nil {
+		t.Fatalf("transform_request: %v", err)
+	}
+	if !strings.Contains(out, `"image_url"`) || !strings.Contains(out, "QUJD") {
+		t.Fatalf("openai passthrough dropped multimodal: %s", out)
+	}
+}