@classicicn/codex-transfer 0.3.1 → 0.3.3

package/README.en.md

@@ -48,7 +48,7 @@ Options:
   -m, --model MODEL      Force override model name (highest priority)
   -c, --config PATH      Path to config file (JSON)
   -k, --insecure         Skip TLS certificate verification
-      --no-reasoning-effort  Don't send reasoning_effort to upstream
+      --reasoning-effort   Send reasoning effort to upstream (default: off)
   -d, --daemon           Run in background, logs to logs/ directory
   -h, --help             Show this help
 ```
@@ -91,7 +91,7 @@ Create a JSON config file at one of these locations (searched in order):
   "upstream": "https://api.deepseek.com/v1",
   "apiKey": "sk-your-key-here",
   "insecure": false,
-  "reasoningEffort": true,
+  "reasoningEffort": false,
   "modelMap": {
     "*": "deepseek-v4-pro",
     "codex-auto-review": "deepseek-v4-pro"
@@ -108,7 +108,7 @@ Create a JSON config file at one of these locations (searched in order):
 | `CODEX_TRANSFER_API_KEY` | _(empty)_ | API key forwarded to upstream |
 | `CODEX_TRANSFER_CONFIG` | _(auto)_ | Path to config file |
 | `CODEX_TRANSFER_INSECURE` | `false` | Set to `"1"` or `"true"` to skip TLS verification |
-| `CODEX_TRANSFER_REASONING_EFFORT` | `true` | Set to `"0"` or `"false"` to disable sending reasoning_effort |
+| `CODEX_TRANSFER_REASONING_EFFORT` | `false` | Set to `"1"` or `"true"` to enable sending reasoning_effort |
 
 ### Model Name Mapping
 
@@ -211,7 +211,7 @@ Codex CLI controls model reasoning intensity via `reasoning.effort` (none/low/me
 | High | `high` | `thinking: {type: "enabled"}, reasoning_effort: "high"` | `thinking: {type: "enabled"}` |
 | Ultra | `xhigh` | `thinking: {type: "enabled"}, reasoning_effort: "max"` | `thinking: {type: "enabled"}` |
 
-**Compatibility strategy**: The `thinking` toggle is always sent (supported by all providers). `reasoning_effort` is sent by default (natively supported by DeepSeek); if the upstream doesn't support this field, it can be disabled via the `--no-reasoning-effort` CLI flag or `"reasoningEffort": false` in the config file.
+**Compatibility strategy**: The `thinking` toggle is always sent (supported by all providers). `reasoning_effort` is not sent by default (to avoid 400 errors from incompatible providers); to enable it, use the `--reasoning-effort` CLI flag or `"reasoningEffort": true` in the config file.
 
 ### Session Management
 
@@ -375,12 +375,17 @@ const { app, port } = createTransfer({
 
 ## Changelog
 
+### v0.3.3 (2026-05-09)
+
+- **Non-streaming tool call support**: `fromChatResponse()` now correctly handles `tool_calls`, generating `function_call` output items (previously the non-streaming path ignored tool calls entirely)
+- **Streaming arguments incremental delivery**: `response.output_item.added` and `response.function_call_arguments.delta` events are now emitted in real time during streaming, instead of being sent in batch after stream completion
+
 ### v0.3.0 (2026-05-08)
 
 - **Token usage**: Extract usage from upstream streaming responses — Codex now correctly displays context utilization (fixes 0% display issue)
 - **Usage details**: Auto-map `cached_tokens` and `reasoning_tokens`, compatible with both OpenAI and DeepSeek upstream formats
 - **Reasoning effort**: Map `reasoning.effort` to DeepSeek `thinking`/`reasoning_effort` and MiMo/Kimi/GLM `thinking` toggle
-- **Config-driven control**: New `--no-reasoning-effort` / `reasoningEffort` option to strip reasoning effort parameters on demand
+- **Config-driven control**: `reasoning_effort` is off by default; enable via `--reasoning-effort` or `reasoningEffort` config option
 
 ### v0.2.0 (2026-05-07)
 

package/README.md

@@ -48,7 +48,7 @@ codex-transfer [options]
   -m, --model MODEL      强制覆盖模型名称（最高优先级）
   -c, --config PATH      配置文件路径（JSON 格式）
   -k, --insecure         跳过 TLS 证书验证（企业代理/自签证书场景）
-      --no-reasoning-effort  不向上传递 reasoning_effort 参数
+      --reasoning-effort   向上游传递 reasoning_effort 参数（默认关闭）
   -d, --daemon           后台运行，日志写入 logs/ 目录
   -h, --help             显示帮助信息
 ```
@@ -91,7 +91,7 @@ CLI 参数 > 环境变量 > 配置文件 > 默认值
   "upstream": "https://api.deepseek.com/v1",
   "apiKey": "sk-your-key-here",
   "insecure": false,
-  "reasoningEffort": true,
+  "reasoningEffort": false,
   "modelMap": {
     "*": "deepseek-v4-pro",
     "codex-auto-review": "deepseek-v4-pro"
@@ -108,7 +108,7 @@ CLI 参数 > 环境变量 > 配置文件 > 默认值
 | `CODEX_TRANSFER_API_KEY` | _(空)_ | 转发给上游的 API Key |
 | `CODEX_TRANSFER_CONFIG` | _(自动)_ | 配置文件路径 |
 | `CODEX_TRANSFER_INSECURE` | `false` | 设为 `"1"` 或 `"true"` 跳过 TLS 验证 |
-| `CODEX_TRANSFER_REASONING_EFFORT` | `true` | 设为 `"0"` 或 `"false"` 关闭 reasoning_effort 传递 |
+| `CODEX_TRANSFER_REASONING_EFFORT` | `false` | 设为 `"1"` 或 `"true"` 开启 reasoning_effort 传递 |
 
 ### 模型名称映射
 
@@ -211,7 +211,7 @@ Codex CLI 通过 `reasoning.effort` 控制模型推理强度（极低/低/中/
 | 高 | `high` | `thinking: {type: "enabled"}, reasoning_effort: "high"` | `thinking: {type: "enabled"}` |
 | 超高 | `xhigh` | `thinking: {type: "enabled"}, reasoning_effort: "max"` | `thinking: {type: "enabled"}` |
 
-**兼容策略**：`thinking` 参数始终发送（所有厂商支持）。`reasoning_effort` 默认发送（DeepSeek 原生支持）；如果上游不支持该字段，可通过 `--no-reasoning-effort` CLI 参数或配置文件 `"reasoningEffort": false` 关闭。
+**兼容策略**：`thinking` 参数始终发送（所有厂商支持）。`reasoning_effort` 默认不发送（避免不兼容的上游返回 400 错误）；如需启用，可通过 `--reasoning-effort` CLI 参数或配置文件 `"reasoningEffort": true` 开启。
 
 ### 会话管理
 
@@ -375,12 +375,17 @@ const { app, port } = createTransfer({
 
 ## 更新日志
 
+### v0.3.3 (2026-05-09)
+
+- **非流式工具调用支持**：`fromChatResponse()` 现在正确处理 `tool_calls`，生成 `function_call` 输出项（此前非流式路径完全忽略工具调用）
+- **流式 arguments 增量推送**：函数调用的 `response.output_item.added` 和 `response.function_call_arguments.delta` 事件在流式过程中实时发出，而非流结束后一次性发送
+
 ### v0.3.0 (2026-05-08)
 
 - **Token 用量**：从上游流式响应中提取 usage，Codex 可正确显示上下文占用率（修复 0% 问题）
 - **用量详情**：自动映射 `cached_tokens` 和 `reasoning_tokens`，兼容 OpenAI 和 DeepSeek 两种上游格式
 - **推理强度**：映射 `reasoning.effort` 到 DeepSeek `thinking`/`reasoning_effort`、MiMo/Kimi/GLM `thinking` 开关
-- **配置化控制**：新增 `--no-reasoning-effort` / `reasoningEffort` 配置项，按需剥离推理强度参数
+- **配置化控制**：`reasoning_effort` 默认关闭，可通过 `--reasoning-effort` 或 `reasoningEffort` 配置项按需开启
 
 ### v0.2.0 (2026-05-07)
 

package/dist/codex-transfer.mjs

@@ -2903,6 +2903,7 @@ var SessionStore = class {
 };
 
 // src/translate.ts
+import { randomUUID as randomUUID2 } from "node:crypto";
 function toChatRequest(req, history, sessions) {
   const messages = [...history];
   const systemText = req.instructions ?? req.system;
@@ -3019,13 +3020,33 @@ function fromChatResponse(id, model, chat) {
     completion_tokens: 0,
     total_tokens: 0
   };
-  const output = [
-    {
+  const output = [];
+  if (text) {
+    output.push({
       type: "message",
       role: "assistant",
       content: [{ type: "output_text", text }]
-    }
-  ];
+    });
+  }
+  for (const tc of choice.message.tool_calls ?? []) {
+    const tcRecord = tc;
+    const func = tcRecord.function;
+    output.push({
+      type: "function_call",
+      id: `fc_${randomUUID2().replace(/-/g, "")}`,
+      call_id: tcRecord.id ?? "",
+      name: func?.name ?? "",
+      arguments: func?.arguments ?? "{}",
+      status: "completed"
+    });
+  }
+  if (output.length === 0) {
+    output.push({
+      type: "message",
+      role: "assistant",
+      content: [{ type: "output_text", text: "" }]
+    });
+  }
   const respUsage = mapUsage(usage);
   const response = {
     id,
@@ -3107,7 +3128,7 @@ function reorderForToolCalls(messages) {
 }
 
 // src/stream.ts
-import { randomUUID as randomUUID2 } from "node:crypto";
+import { randomUUID as randomUUID3 } from "node:crypto";
 async function* translateStream(args2, signal) {
   const {
     url,
@@ -3120,7 +3141,7 @@ async function* translateStream(args2, signal) {
     model
   } = args2;
   try {
-    const msgItemId = `msg_${randomUUID2().replace(/-/g, "")}`;
+    const msgItemId = `msg_${randomUUID3().replace(/-/g, "")}`;
     const createdAt = Math.floor(Date.now() / 1e3);
     yield formatSSE("response.created", {
       type: "response.created",
@@ -3251,13 +3272,45 @@ async function* translateStream(args2, signal) {
                   for (const dc of deltaCalls) {
                     let entry = toolCalls.get(dc.index);
                     if (!entry) {
-                      entry = { id: "", name: "", arguments: "" };
+                      entry = {
+                        id: "",
+                        name: "",
+                        arguments: "",
+                        fcItemId: `fc_${randomUUID3().replace(/-/g, "")}`,
+                        emittedAdded: false
+                      };
                       toolCalls.set(dc.index, entry);
                     }
                     if (dc.id) entry.id = dc.id;
                     if (dc.function?.name) entry.name += dc.function.name;
-                    if (dc.function?.arguments)
+                    if (!entry.emittedAdded && entry.id && entry.name) {
+                      const fcOutputIndex = (emittedMessageItem ? 1 : 0) + dc.index;
+                      yield formatSSE("response.output_item.added", {
+                        type: "response.output_item.added",
+                        output_index: fcOutputIndex,
+                        item: {
+                          type: "function_call",
+                          id: entry.fcItemId,
+                          call_id: entry.id,
+                          name: entry.name,
+                          arguments: "",
+                          status: "in_progress"
+                        }
+                      });
+                      entry.emittedAdded = true;
+                    }
+                    if (dc.function?.arguments) {
                       entry.arguments += dc.function.arguments;
+                      if (entry.emittedAdded) {
+                        const fcOutputIndex = (emittedMessageItem ? 1 : 0) + dc.index;
+                        yield formatSSE("response.function_call_arguments.delta", {
+                          type: "response.function_call_arguments.delta",
+                          item_id: entry.fcItemId,
+                          output_index: fcOutputIndex,
+                          delta: dc.function.arguments
+                        });
+                      }
+                    }
                   }
                 }
                 if (choice.finish_reason) {
@@ -3294,43 +3347,39 @@ async function* translateStream(args2, signal) {
     const fcItems = [];
     let relIdx = 0;
     for (const [, tc] of toolCalls) {
-      const fcItemId = `fc_${randomUUID2().replace(/-/g, "")}`;
       const outputIndex = baseIndex + relIdx;
-      yield formatSSE("response.output_item.added", {
-        type: "response.output_item.added",
-        output_index: outputIndex,
-        item: {
-          type: "function_call",
-          id: fcItemId,
-          call_id: tc.id,
-          name: tc.name,
-          arguments: "",
-          status: "in_progress"
-        }
-      });
-      if (tc.arguments) {
-        yield formatSSE("response.function_call_arguments.delta", {
-          type: "response.function_call_arguments.delta",
-          item_id: fcItemId,
+      if (!tc.emittedAdded && tc.id && tc.name) {
+        yield formatSSE("response.output_item.added", {
+          type: "response.output_item.added",
           output_index: outputIndex,
-          delta: tc.arguments
+          item: {
+            type: "function_call",
+            id: tc.fcItemId,
+            call_id: tc.id,
+            name: tc.name,
+            arguments: "",
+            status: "in_progress"
+          }
+        });
+        tc.emittedAdded = true;
+      }
+      if (tc.emittedAdded) {
+        yield formatSSE("response.output_item.done", {
+          type: "response.output_item.done",
+          output_index: outputIndex,
+          item: {
+            type: "function_call",
+            id: tc.fcItemId,
+            call_id: tc.id,
+            name: tc.name,
+            arguments: tc.arguments,
+            status: "completed"
+          }
         });
       }
-      yield formatSSE("response.output_item.done", {
-        type: "response.output_item.done",
-        output_index: outputIndex,
-        item: {
-          type: "function_call",
-          id: fcItemId,
-          call_id: tc.id,
-          name: tc.name,
-          arguments: tc.arguments,
-          status: "completed"
-        }
-      });
       fcItems.push({
         type: "function_call",
-        id: fcItemId,
+        id: tc.fcItemId,
         call_id: tc.id,
         name: tc.name,
         arguments: tc.arguments,
@@ -3417,7 +3466,7 @@ var DEFAULT_CONFIG = {
   apiKey: "",
   insecure: false,
   modelMap: {},
-  reasoningEffort: true
+  reasoningEffort: false
 };
 function loadConfig(configPath) {
   const fileConfig = loadConfigFile(configPath);
@@ -3430,7 +3479,7 @@ function loadConfig(configPath) {
     ),
     modelMap: fileConfig.modelMap ?? DEFAULT_CONFIG.modelMap,
     reasoningEffort: parseBool(
-      process.env.CODEX_TRANSFER_REASONING_EFFORT ?? fileConfig.reasoningEffort ?? true
+      process.env.CODEX_TRANSFER_REASONING_EFFORT ?? fileConfig.reasoningEffort ?? false
     )
   };
 }
@@ -3674,8 +3723,8 @@ for (let i = 0; i < args.length; i++) {
     overrides.configPath = args[++i];
   } else if ((a === "--model" || a === "-m") && args[i + 1]) {
     overrides.model = args[++i];
-  } else if (a === "--no-reasoning-effort") {
-    overrides.reasoningEffort = "false";
+  } else if (a === "--reasoning-effort") {
+    overrides.reasoningEffort = "true";
   } else if (a === "--help" || a === "-h") {
     console.log(`
 codex-transfer \u2014 Responses API \u2194 Chat Completions bridge
@@ -3690,7 +3739,7 @@ Options:
   -m, --model MODEL      Override model name (highest priority model mapping)
   -c, --config PATH      Path to config file (JSON)
   -k, --insecure         Skip TLS certificate verification
-      --no-reasoning-effort  Don't send reasoning_effort to upstream
+      --reasoning-effort   Send reasoning effort to upstream (default: off)
   -d, --daemon           Run in background, logs to logs/ directory
   -h, --help             Show this help
 
@@ -3700,7 +3749,7 @@ Environment variables:
   CODEX_TRANSFER_API_KEY      Same as --api-key
   CODEX_TRANSFER_CONFIG       Same as --config
   CODEX_TRANSFER_INSECURE     Set to "1" to skip TLS verification
-  CODEX_TRANSFER_REASONING_EFFORT  Set to "0" to disable reasoning_effort
+  CODEX_TRANSFER_REASONING_EFFORT  Set to "1" to enable reasoning_effort
 
 Config file options:
   modelMap               Model name mapping, e.g. {"*": "deepseek-v4-pro"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@classicicn/codex-transfer",
-  "version": "0.3.1",
+  "version": "0.3.3",
   "description": "Responses API ↔ Chat Completions translation bridge for Codex — use DeepSeek, Kimi, Qwen, and other providers with Codex",
   "type": "module",
   "main": "dist/codex-transfer.mjs",