chcode 0.1.0__py3-none-any.whl

chcode/prompts.py

@@ -0,0 +1,640 @@
+"""
+统一交互层 — 所有用户交互都通过此模块
+
+用 questionary 实现下拉列表、确认框、文本输入等。
+在 async 上下文中用 asyncio.to_thread 包装同步的 questionary 调用。
+"""
+
+from __future__ import annotations
+
+import asyncio
+import os
+from typing import Any
+
+import questionary
+from rich.console import Console
+
+console = Console()
+
+
+async def select(
+    message: str,
+    choices: list[str],
+    default: str | None = None,
+) -> str | None:
+    """下拉单选"""
+
+    def _ask():
+        return questionary.select(
+            message=message,
+            choices=choices,
+            default=default,
+        ).ask()
+
+    return await asyncio.to_thread(_ask)
+
+
+async def confirm(message: str, default: bool = True) -> bool:
+    """确认框"""
+
+    def _ask():
+        return questionary.confirm(
+            message=message,
+            default=default,
+        ).ask()
+
+    return await asyncio.to_thread(_ask)
+
+
+async def checkbox(message: str, choices: list[str]) -> list[str]:
+    """多选框"""
+
+    def _ask():
+        return questionary.checkbox(message=message, choices=choices).ask()
+
+    return await asyncio.to_thread(_ask) or []
+
+
+async def text(message: str, default: str = "") -> str:
+    """文本输入"""
+
+    def _ask():
+        return questionary.text(
+            message=message,
+            default=default,
+        ).ask()
+
+    return await asyncio.to_thread(_ask)
+
+
+async def password(message: str) -> str:
+    """密码输入（隐藏回显）"""
+
+    def _ask():
+        return questionary.password(
+            message=message,
+        ).ask()
+
+    return await asyncio.to_thread(_ask)
+
+
+async def select_or_custom(
+    message: str,
+    preset_choices: list[str],
+    custom_label: str = "自定义输入...",
+    custom_prompt: str = "请输入: ",
+    default: str | None = None,
+) -> str:
+    """下拉选择 + 自定义输入。末尾有「自定义输入...」选项。"""
+    choices = list(preset_choices) + [custom_label]
+    result = await select(message, choices, default=default)
+    if result == custom_label:
+        return await text(custom_prompt)
+    return result
+
+
+# ─── 模型配置表单专用 ──────────────────────────────────────────
+
+BASE_URL_PRESETS = [
+    "https://api.openai.com/v1",
+    "https://api-inference.modelscope.cn/v1",
+    "https://open.bigmodel.cn/api/paas/v4",
+    "https://api.deepseek.com/v1",
+    "https://dashscope.aliyuncs.com/compatible-mode/v1",
+    "https://api.longcat.chat/openai/v1",
+]
+
+MODELSCOPE_BASE_URL = "https://api-inference.modelscope.cn/v1"
+
+MODELSCOPE_PRESETS = [
+    {
+        "model": "ZhipuAI/GLM-5",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "Qwen/Qwen3-235B-A22B-Thinking-2507",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 0.6,
+        "top_p": 0.95,
+        "stream_usage": True,
+        "extra_body": {"top_k": 20},
+    },
+    {
+        "model": "Qwen/Qwen3-235B-A22B-Instruct-2507",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 0.7,
+        "top_p": 0.8,
+        "stream_usage": True,
+        "extra_body": {"top_k": 20},
+    },
+    {
+        "model": "Qwen/Qwen3.5-397B-A17B",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 0.6,
+        "top_p": 0.95,
+        "stream_usage": True,
+        "extra_body": {"top_k": 20, "repetition_penalty": 1.0},
+    },
+    {
+        "model": "deepseek-ai/DeepSeek-V3.2",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "MiniMax/MiniMax-M2.5",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+        "extra_body": {"top_k": 40},
+    },
+    {
+        "model": "moonshotai/Kimi-K2.5",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "ZhipuAI/GLM-5.1",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "Qwen/Qwen3-Coder-480B-A35B-Instruct",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 0.7,
+        "top_p": 0.8,
+        "stream_usage": True,
+        "extra_body": {"top_k": 20, "repetition_penalty": 1.05},
+    },
+    {
+        "model": "XiaomiMiMo/MiMo-V2-Flash",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 0.3,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "deepseek-ai/DeepSeek-R1-0528",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 0.6,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "Qwen/Qwen3-Next-80B-A3B-Thinking",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 0.6,
+        "top_p": 0.95,
+        "stream_usage": True,
+        "extra_body": {"top_k": 20},
+    },
+    # {
+    #     "model": "MiniMax/MiniMax-M2.7",
+    #     "base_url": MODELSCOPE_BASE_URL,
+    #     "temperature": 1.0,
+    #     "top_p": 0.95,
+    #     "stream_usage": True,
+    #     "extra_body": {"top_k": 40},
+    # },
+    {
+        "model": "deepseek-ai/DeepSeek-V4-Pro",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 1.0,
+        "stream_usage": True,
+    },
+    {
+        "model": "deepseek-ai/DeepSeek-V4-Flash",
+        "base_url": MODELSCOPE_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 1.0,
+        "stream_usage": True,
+    },
+]
+
+API_KEY_ENV_VARS = [
+    ("BIGMODEL_API_KEY", "智谱 GLM"),
+    ("ModelScopeToken", "ModelScope"),
+    ("OPENAI_API_KEY", "OpenAI"),
+    ("DEEPSEEK_API_KEY", "DeepSeek"),
+    ("DASHSCOPE_API_KEY", "通义千问"),
+    ("ANTHROPIC_API_KEY", "Anthropic Claude"),
+    ("LONGCAT_API_KEY", "LongCat"),
+]
+
+TEMPERATURE_PRESETS = ["0", "0.3", "0.5", "0.6", "0.7", "1.0", "1.5", "2.0"]
+TOP_P_PRESETS = ["0.5", "0.7", "0.9", "0.95", "1.0"]
+TOP_K_PRESETS = ["1", "5", "10", "20", "40", "50"]
+MAX_COMPLETION_TOKENS_PRESETS = ["32768", "65536", "122880", "204800"]
+FREQ_PENALTY_PRESETS = ["0", "0.2", "0.5", "1.0", "1.5", "2.0"]
+PRESENCE_PENALTY_PRESETS = ["0", "0.2", "0.5", "1.0", "1.5", "2.0"]
+
+SKIP_LABEL = "跳过 (不设置)"
+
+
+class _SkipSentinel:
+    """哨兵对象，区分「跳过此字段」和「用户取消整个表单」。"""
+
+    _instance = None
+
+    def __new__(cls):
+        if cls._instance is None:
+            cls._instance = super().__new__(cls)
+        return cls._instance
+
+    def __repr__(self):
+        return "SKIP"
+
+
+_SKIP = _SkipSentinel()
+
+
+async def _ask_hyperparam(
+    message: str,
+    preset_choices: list[str],
+    existing_value: str | None = None,
+    custom_prompt: str = "请输入: ",
+) -> Any:
+    """单个超参输入，支持「跳过」。返回值 / _SKIP / None(取消)。"""
+    choices = [SKIP_LABEL] + list(preset_choices) + ["自定义输入..."]
+
+    default = None
+    if existing_value is not None and existing_value in preset_choices:
+        default = existing_value
+
+    result = await select(message, choices, default=default)
+    if result is None:
+        return None
+    if result == SKIP_LABEL:
+        return _SKIP
+    if result == "自定义输入...":
+        raw = await text(custom_prompt)
+        if raw is None or raw.strip() == "":
+            return _SKIP
+        return raw.strip()
+    return result
+
+
+async def model_config_form(
+    existing_config: dict | None = None,
+) -> dict | None:
+    """
+    模型配置表单 — 全部用下拉列表 + 文本输入
+
+    Args:
+        existing_config: 现有配置（编辑模式）
+
+    Returns:
+        配置字典，用户取消返回 None
+    """
+    import os
+
+    cfg = dict(existing_config) if existing_config else {}
+
+    # ─── 必填字段 ───
+    is_editing = bool(cfg)
+    KEEP_LABEL = "保持当前值"
+
+    # ── 模型名称 ──
+    model_name = cfg.get("model", "")
+    if not model_name:
+        model_name = await text("输入模型名称: ")
+        if not model_name or not model_name.strip():
+            return None
+        model_name = model_name.strip()
+
+    # ── Base URL ──
+    base_url = cfg.get("base_url", "")
+    if is_editing and base_url:
+        _keep_url = f"{KEEP_LABEL} ({base_url})"
+        _url_choices = [_keep_url] + list(BASE_URL_PRESETS) + ["自定义输入..."]
+        result = await select("选择 API Base URL:", _url_choices, default=_keep_url)
+        if result is None:
+            return None
+        if result == _keep_url:
+            pass  # base_url unchanged
+        elif result == "自定义输入...":
+            base_url = await text("输入 Base URL: ")
+            if not base_url or not base_url.strip():
+                return None
+        else:
+            base_url = result
+    else:
+        _url_choices = ["魔搭 (ModelScope)"] + list(BASE_URL_PRESETS) + ["自定义输入..."]
+        result = await select("选择 API Base URL:", _url_choices)
+        if result is None:
+            return None
+        if result == "魔搭 (ModelScope)":
+            base_url = MODELSCOPE_BASE_URL
+        elif result == "自定义输入...":
+            base_url = await text("输入 Base URL: ")
+            if not base_url or not base_url.strip():
+                return None
+        else:
+            base_url = result
+
+    # API Key — 先展示环境变量快捷选择
+    existing_api_key = cfg.get("api_key", "")
+
+    env_choices = [
+        f"{var} ({desc})" for var, desc in API_KEY_ENV_VARS if os.getenv(var)
+    ]
+
+    if is_editing:
+        _masked = (
+            existing_api_key[:6] + "****" + existing_api_key[-4:]
+            if len(existing_api_key) > 10
+            else "****"
+        )
+        env_choices.insert(0, f"保持当前 Key ({_masked})")
+
+    env_choices.append("手动输入 API Key...")
+    if env_choices:
+        result = await select("选择 API Key 来源:", env_choices)
+        if result is None:
+            return None
+        if result.startswith("保持当前 Key"):
+            api_key = existing_api_key
+        elif result == "手动输入 API Key...":
+            api_key = await password("输入 API Key: ")
+        else:
+            var_name = result.split(" (")[0]
+            api_key = os.getenv(var_name, "")
+    else:
+        api_key = await password("输入 API Key: ")
+
+    if not api_key:
+        console.print("[red]API Key 不能为空[/red]")
+        return None
+
+    config: dict[str, Any] = {
+        "model": model_name,
+        "base_url": base_url,
+        "api_key": api_key,
+        "stream_usage": True,
+    }
+
+    # 编辑旧配置时清理已移除的 max_tokens 字段
+    cfg.pop("max_tokens", None)
+
+    # ─── 超参（可选） ───
+    want_hyperparams = await confirm("配置超参数？", default=False)
+    if want_hyperparams:
+        # temperature
+        t_val = str(cfg["temperature"]) if "temperature" in cfg else None
+        result = await _ask_hyperparam(
+            "Temperature:",
+            TEMPERATURE_PRESETS,
+            existing_value=t_val,
+            custom_prompt="输入 temperature: ",
+        )
+        if result is None:
+            return None
+        if result is not _SKIP:
+            config["temperature"] = float(result)
+        else:
+            config.pop("temperature", None)
+
+        # top_p
+        tp_val = str(cfg["top_p"]) if "top_p" in cfg else None
+        result = await _ask_hyperparam(
+            "Top P:",
+            TOP_P_PRESETS,
+            existing_value=tp_val,
+            custom_prompt="输入 top_p: ",
+        )
+        if result is None:
+            return None
+        if result is not _SKIP:
+            config["top_p"] = float(result)
+        else:
+            config.pop("top_p", None)
+
+        # top_k → extra_body
+        existing_extra = cfg.get("extra_body", {})
+        tk_val = (
+            str(existing_extra["top_k"])
+            if isinstance(existing_extra, dict) and "top_k" in existing_extra
+            else None
+        )
+        result = await _ask_hyperparam(
+            "Top K:",
+            TOP_K_PRESETS,
+            existing_value=tk_val,
+            custom_prompt="输入 top_k: ",
+        )
+        if result is None:
+            return None
+        if result is not _SKIP:
+            # 合并到已有的 extra_body（可能已有 max_completion_tokens）
+            _eb = dict(existing_extra) if isinstance(existing_extra, dict) else {}
+            _eb["top_k"] = int(result)
+            config["extra_body"] = _eb
+        else:
+            # 跳过 top_k，但仍保留 extra_body 中的其他字段（如 max_completion_tokens）
+            if isinstance(existing_extra, dict):
+                _eb = {k: v for k, v in existing_extra.items() if k != "top_k"}
+                if _eb:
+                    config["extra_body"] = _eb
+
+        # max_completion_tokens → extra_body
+        _eb = config.get("extra_body", {})
+        mct_val = (
+            str(_eb["max_completion_tokens"])
+            if isinstance(_eb, dict) and "max_completion_tokens" in _eb
+            else None
+        )
+        result = await _ask_hyperparam(
+            "Max Completion Tokens:",
+            MAX_COMPLETION_TOKENS_PRESETS,
+            existing_value=mct_val,
+            custom_prompt="输入 max_completion_tokens: ",
+        )
+        if result is None:
+            return None
+        if result is not _SKIP:
+            _eb = dict(_eb) if isinstance(_eb, dict) else {}
+            _eb["max_completion_tokens"] = int(result)
+            config["extra_body"] = _eb
+        else:
+            if isinstance(_eb, dict):
+                _eb = {k: v for k, v in _eb.items() if k != "max_completion_tokens"}
+                if _eb:
+                    config["extra_body"] = _eb
+                else:
+                    config.pop("extra_body", None)
+
+        # stop_sequences
+        ss_val = None
+        if "stop_sequences" in cfg:
+            v = cfg["stop_sequences"]
+            ss_val = ", ".join(str(x) for x in v) if isinstance(v, list) else str(v)
+        result = await _ask_hyperparam(
+            "Stop Sequences:",
+            [],  # 无预设，只有自定义
+            existing_value=ss_val,
+            custom_prompt="输入停止序列 (逗号分隔): ",
+        )
+        if result is None:
+            return None
+        if result is not _SKIP:
+            config["stop_sequences"] = [
+                s.strip() for s in str(result).split(",") if s.strip()
+            ]
+        else:
+            config.pop("stop_sequences", None)
+
+        # frequency_penalty
+        fp_val = str(cfg["frequency_penalty"]) if "frequency_penalty" in cfg else None
+        result = await _ask_hyperparam(
+            "Frequency Penalty:",
+            FREQ_PENALTY_PRESETS,
+            existing_value=fp_val,
+            custom_prompt="输入 frequency_penalty: ",
+        )
+        if result is None:
+            return None
+        if result is not _SKIP:
+            config["frequency_penalty"] = float(result)
+        else:
+            config.pop("frequency_penalty", None)
+
+        # presence_penalty
+        pp_val = str(cfg["presence_penalty"]) if "presence_penalty" in cfg else None
+        result = await _ask_hyperparam(
+            "Presence Penalty:",
+            PRESENCE_PENALTY_PRESETS,
+            existing_value=pp_val,
+            custom_prompt="输入 presence_penalty: ",
+        )
+        if result is None:
+            return None
+        if result is not _SKIP:
+            config["presence_penalty"] = float(result)
+        else:
+            config.pop("presence_penalty", None)
+
+        # max_retries - 固定为 4（失败 5 次后自动切换备用模型），不可配置
+        config["max_retries"] = 4
+
+    return config
+
+
+async def configure_modelscope() -> dict | None:
+    """魔搭快捷配置 — 只需 API Key，自动生成 12 个预定义模型。"""
+    # 收集 API Key
+    env_choices = [
+        f"{var} ({desc})"
+        for var, desc in API_KEY_ENV_VARS
+        if var == "ModelScopeToken" and os.getenv(var)
+    ]
+    if env_choices:
+        result = await select(
+            "检测到 ModelScope Token，是否使用？", env_choices + ["手动输入..."]
+        )
+        if result is None:
+            return None
+        if result == "手动输入...":
+            api_key = await password("输入 ModelScope API Key: ")
+        else:
+            # 提取 env var 名
+            var_name = result.split(" (")[0]
+            api_key = os.getenv(var_name, "")
+    else:
+        api_key = await password("输入 ModelScope API Key: ")
+
+    if not api_key or not api_key.strip():
+        return None
+    api_key = api_key.strip()
+
+    # 用预设模型 + api_key 构建配置
+    default_cfg = dict(MODELSCOPE_PRESETS[0])
+    default_cfg["api_key"] = api_key
+
+    fallback = {}
+    for preset in MODELSCOPE_PRESETS[1:]:
+        cfg = dict(preset)
+        cfg["api_key"] = api_key
+        fallback[cfg["model"]] = cfg
+
+    return {"default": default_cfg, "fallback": fallback}
+
+
+LONGCAT_BASE_URL = "https://api.longcat.chat/openai/v1"
+
+LONGCAT_PRESETS = [
+    {
+        "model": "LongCat-2.0-Preview",
+        "base_url": LONGCAT_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "LongCat-Flash-Chat",
+        "base_url": LONGCAT_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "LongCat-Flash-Thinking",
+        "base_url": LONGCAT_BASE_URL,
+        "temperature": 0.6,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+    {
+        "model": "LongCat-Flash-Lite",
+        "base_url": LONGCAT_BASE_URL,
+        "temperature": 1.0,
+        "top_p": 0.95,
+        "stream_usage": True,
+    },
+]
+
+
+async def configure_longcat() -> dict | None:
+    """LongCat 快捷配置 — 只需 API Key，自动生成 4 个预定义模型。"""
+    env_choices = [
+        f"{var} ({desc})"
+        for var, desc in API_KEY_ENV_VARS
+        if var == "LONGCAT_API_KEY" and os.getenv(var)
+    ]
+    if env_choices:
+        result = await select(
+            "检测到 LongCat API Key，是否使用？", env_choices + ["手动输入..."]
+        )
+        if result is None:
+            return None
+        if result == "手动输入...":
+            api_key = await password("输入 LongCat API Key: ")
+        else:
+            var_name = result.split(" (")[0]
+            api_key = os.getenv(var_name, "")
+    else:
+        api_key = await password("输入 LongCat API Key: ")
+
+    if not api_key or not api_key.strip():
+        return None
+    api_key = api_key.strip()
+
+    default_cfg = dict(LONGCAT_PRESETS[0])
+    default_cfg["api_key"] = api_key
+
+    fallback = {}
+    for preset in LONGCAT_PRESETS[1:]:
+        cfg = dict(preset)
+        cfg["api_key"] = api_key
+        fallback[cfg["model"]] = cfg
+
+    return {"default": default_cfg, "fallback": fallback}