precis-cli 0.1.3__py3-none-any.whl

app/cli/shell/commands/ai/migrate.py

@@ -0,0 +1,281 @@
+# SPDX-License-Identifier: Apache-2.0
+#
+# Copyright 2026 Precis Team
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# backend/app/cli/shell/commands/ai/migrate.py
+"""
+@fileoverview CLI AI 配置迁移命令模块
+
+功能概述:
+- 提供 `ai migrate` 子命令，从旧脚本迁移生成 Precis V2 配置
+- 支持 Python pandas / SQL / Excel 公式 / 自然语言
+- 默认仅输出 YAML 预览，加 --apply 后写盘
+
+架构设计:
+- AIMigrateCommand 继承 Command 基类
+- 直接调用 ConfigMigrationService，不依赖后端 HTTP
+- 写盘逻辑委托 shared_services.generation_ops.apply_generated_config（CLI/TUI 同源）
+
+输入示例:
+    precis> ai migrate scripts/legacy_rules.sql data/users.xlsx
+    precis> ai migrate rules.py data/*.xlsx --apply
+
+输出示例:
+    YAML 预览或写盘成功提示
+"""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+import os
+from pathlib import Path
+from typing import Any
+
+from app.cli.shared_services.generation_ops import (
+    SUPPORTED_EXTENSIONS,
+)
+from app.cli.shared_services.generation_ops import (
+    apply_generated_config as _apply_generated_config,
+)
+from app.cli.shell.commands.base import Command, CommandResult, ProjectContext
+from app.cli.shell.formatter import Formatter
+from app.shared.services.ai.migrate_service import ConfigMigrationService
+from app.shared.services.llm.generation import (
+    GenerationOptions,
+    ProfilingOptions,
+)
+
+logger = logging.getLogger(__name__)
+
+
+def _infer_language(file_path: str) -> str:
+    """根据文件扩展名推断脚本语言。"""
+    lower = file_path.lower()
+    if lower.endswith(".py"):
+        return "python"
+    if lower.endswith(".sql"):
+        return "sql"
+    if lower.endswith(".xlsx") or lower.endswith(".xls") or lower.endswith(".csv"):
+        return "excel_formula"
+    return "natural_language"
+
+
+class AIMigrateCommand(Command):
+    """AI 配置迁移命令。
+
+    从旧脚本或业务描述迁移生成 Schema、Constraint、Regex 配置。
+    默认输出 YAML 预览，使用 --apply 写入项目目录。
+    """
+
+    def __init__(self) -> None:
+        super().__init__("migrate", aliases=["mig"])
+
+    @property
+    def description(self) -> str:
+        return "从旧脚本迁移生成 Precis V2 配置"
+
+    @property
+    def usage(self) -> str:
+        return "ai migrate <script_file> [data_files...] [--apply] [--language python|sql|excel_formula|natural_language] [--max-iterations N]"
+
+    def execute(self, args: list[str], context: ProjectContext) -> CommandResult:
+        """执行配置迁移命令。
+
+        Args:
+            args: 命令参数列表
+            context: 项目上下文
+
+        Returns:
+            命令执行结果
+        """
+        if not context.is_project_open:
+            return CommandResult.error("未打开项目，请先使用 'open <path>' 命令打开项目")
+
+        project_path = context.project_path
+        if not project_path:
+            return CommandResult.error("项目路径为空")
+
+        if not args:
+            return CommandResult.error(f"请提供脚本文件路径\n用法: {self.usage}")
+
+        # 解析参数
+        script_path = args[0]
+        file_patterns: list[str] = []
+        apply = False
+        language: str | None = None
+        max_iterations = 2
+        sample_rows = 100
+        sample_values = 100
+
+        i = 1
+        while i < len(args):
+            arg = args[i]
+            if arg == "--apply":
+                apply = True
+            elif arg == "--language":
+                i += 1
+                if i < len(args):
+                    language = args[i]
+                else:
+                    return CommandResult.error("--language 需要参数")
+            elif arg == "--max-iterations":
+                i += 1
+                if i < len(args):
+                    try:
+                        max_iterations = max(1, min(5, int(args[i])))
+                    except ValueError:
+                        return CommandResult.error(f"--max-iterations 需要整数， got: {args[i]}")
+                else:
+                    return CommandResult.error("--max-iterations 需要参数")
+            elif arg == "--sample-rows":
+                i += 1
+                if i < len(args):
+                    try:
+                        sample_rows = max(1, int(args[i]))
+                    except ValueError:
+                        return CommandResult.error(f"--sample-rows 需要整数， got: {args[i]}")
+                else:
+                    return CommandResult.error("--sample-rows 需要参数")
+            elif arg == "--sample-values":
+                i += 1
+                if i < len(args):
+                    try:
+                        sample_values = max(1, int(args[i]))
+                    except ValueError:
+                        return CommandResult.error(f"--sample-values 需要整数， got: {args[i]}")
+                else:
+                    return CommandResult.error("--sample-values 需要参数")
+            elif arg.startswith("--"):
+                return CommandResult.error(f"未知参数: {arg}\n用法: {self.usage}")
+            else:
+                file_patterns.append(arg)
+            i += 1
+
+        # 解析脚本路径
+        if not os.path.isabs(script_path):
+            script_path = os.path.join(project_path, script_path)
+        if not os.path.exists(script_path):
+            return CommandResult.error(f"脚本文件不存在: {script_path}")
+
+        if language is None:
+            language = _infer_language(script_path)
+
+        try:
+            with open(script_path, encoding="utf-8") as f:
+                script_content = f.read()
+        except Exception as e:
+            return CommandResult.error(f"读取脚本文件失败: {e}")
+
+        # 展开数据文件路径
+        file_paths: list[str] = []
+        for pattern in file_patterns:
+            if not os.path.isabs(pattern):
+                pattern = os.path.join(project_path, pattern)
+            import glob as _glob
+
+            matched = _glob.glob(pattern)
+            if matched:
+                file_paths.extend(matched)
+            elif os.path.exists(pattern):
+                file_paths.append(pattern)
+
+        file_paths = [p for p in file_paths if p.lower().endswith(SUPPORTED_EXTENSIONS)]
+        if not file_paths:
+            return CommandResult.error("请提供至少一个数据文件（.xlsx/.xls/.csv/.json/.jsonl）")
+
+        # 获取项目信息
+        config = context.project_config or {}
+        project_name = config.get("project", {}).get("name", Path(project_path).name)
+        project_id = config.get("project", {}).get("id", project_name)
+
+        print(Formatter.header("\nAI 配置迁移"))
+        print(Formatter.info(f"项目: {project_name}"))
+        print(Formatter.info(f"脚本: {os.path.relpath(script_path, project_path)} ({language})"))
+        print(Formatter.info("数据文件:"))
+        for p in file_paths:
+            print(f"  - {os.path.relpath(p, project_path)}")
+
+        def progress_callback(stage: str, progress: float, extra: dict[str, Any] | None = None) -> None:
+            """终端进度回调。"""
+            msg = extra.get("message") if extra else None
+            prefix = f"[{stage}]" if not msg else f"[{stage}] {msg}"
+            print(Formatter.info(f"  {prefix} {progress:.0f}%"))
+
+        service = ConfigMigrationService()
+        profiling_opts = ProfilingOptions(
+            sample_rows=sample_rows,
+            sample_values_per_column=sample_values,
+        )
+        gen_opts = GenerationOptions(
+            generate_schemas=True,
+            generate_constraints=True,
+            generate_regex_nodes=False,
+            keep_existing=True,
+        )
+
+        # CLI 是同步环境：asyncio.run 创建一次性事件循环并在结束时确保关闭
+        # （不再 get_running_loop/new_event_loop 泄漏循环）
+        try:
+            result = asyncio.run(
+                service.migrate_from_script(
+                    script_content=script_content,
+                    language=language,
+                    file_paths=file_paths,
+                    project_name=project_name,
+                    project_id=project_id,
+                    config_path=project_path,
+                    profiling_options=profiling_opts,
+                    generation_options=gen_opts,
+                    max_iterations=max_iterations,
+                    progress_callback=progress_callback,
+                )
+            )
+        except Exception as e:
+            logger.error(f"配置迁移失败: {e}", exc_info=True)
+            return CommandResult.error(f"配置迁移失败: {e}")
+
+        if not result.get("success"):
+            error = result.get("error") or "配置迁移失败"
+            return CommandResult.error(error)
+
+        yaml_preview = result.get("yaml_preview", "")
+        warnings = result.get("warnings", [])
+
+        if warnings:
+            print(Formatter.warning("\n迁移警告:"))
+            for w in warnings:
+                print(f"  - {w}")
+
+        if apply:
+            try:
+                _apply_generated_config(result, project_path)
+            except Exception as e:
+                logger.error(f"写盘失败: {e}", exc_info=True)
+                return CommandResult.error(f"配置已生成，但写盘失败: {e}")
+
+            schema_count = len(result.get("schemas", {}))
+            constraint_count = len(result.get("constraints", {}))
+            regex_count = len(result.get("regex_nodes", {}))
+            return CommandResult.ok(
+                f"配置已迁移并写入项目：{schema_count} 个 schema，"
+                f"{constraint_count} 个 constraint，{regex_count} 个 regex。"
+            )
+
+        print(Formatter.header("\nYAML 预览（未写盘，加 --apply 可应用）"))
+        print(yaml_preview)
+        return CommandResult.ok("配置迁移完成，以上为 YAML 预览。")
+
+
+__all__ = ["AIMigrateCommand"]

app/cli/shell/commands/ai/resolver.py

@@ -0,0 +1,224 @@
+# SPDX-License-Identifier: Apache-2.0
+#
+# Copyright 2026 Precis Team
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# backend/app/cli/shell/commands/ai/resolver.py
+"""
+@fileoverview AI 动作歧义解析模块
+
+功能概述:
+- 解析 AI 返回动作中的表名歧义（支持模糊匹配）
+- 当表名匹配到多个结果时，弹出交互式菜单让用户选择正确的表
+- 处理 VALIDATE_PROJECT 与约束操作（ADD/UPDATE/DELETE_CONSTRAINT_NODE）中的表名解析
+
+架构设计:
+- resolve_table_name(): 使用 find_matching_schemas 对表名进行模糊匹配
+- resolve_ambiguities(): 遍历 AI 返回的动作列表，解析表名歧义并标记跳过项
+- 静默模式下自动选择第一个匹配项，避免阻塞非交互流程
+
+输入示例:
+    resolve_table_name("users", "/path/to/project")
+    resolve_ambiguities(actions, "/path/to/project")
+
+输出示例:
+    "users_table_id" 或 None（未找到）
+    True/False（是否还有有效动作需要继续执行）
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any
+
+from app.cli.shell.formatter import Formatter
+from app.cli.shell.interactive_menu import InteractiveMenu
+from app.shared.services.llm.schema_resolver import find_matching_schemas
+
+logger = logging.getLogger(__name__)
+
+
+def resolve_table_name(table_name: str, project_path: str, silent: bool = False) -> str | None:
+    """将表名解析为表ID。
+
+    支持模糊匹配，如果有多个匹配则让用户选择。
+
+    Args:
+        table_name: 用户输入的表名
+        project_path: 项目路径
+        silent: 是否静默模式（不显示选择菜单）
+
+    Returns:
+        表ID，如果未找到则返回 None
+    """
+    matches = find_matching_schemas(project_path, table_name)
+
+    if not matches:
+        if not silent:
+            print(Formatter.warning(f"[!] 未找到匹配的表: '{table_name}'"))
+        return None
+
+    if len(matches) == 1:
+        # 只有一个匹配，直接返回
+        selected = matches[0]
+        if not silent:
+            print(Formatter.info(f"[i] 表 '{table_name}' 匹配到: {selected['name']} ({selected['id'][:10]}...)"))
+        return selected["id"]
+
+    # 多个匹配，让用户选择
+    if not silent:
+        print(f"\n{Formatter.warning('[!]')} 表名 '{table_name}' 匹配到多个表:")
+
+        menu = InteractiveMenu("请选择要校验的表:")
+        for m in matches:
+            display_id = f"{m['id'][:10]}..." if len(m["id"]) > 15 else m["id"]
+            menu.add_item(m["id"], f"{m['name']} ({display_id})", f"ID: {m['id']}")
+
+        menu.add_item("_cancel_", "取消", "不执行此操作")
+
+        choice = menu.show()
+        if choice and choice != "_cancel_":
+            selected = next(m for m in matches if m["id"] == choice)
+            return selected["id"]
+        # 用户取消，返回 None
+        return None
+
+    # 静默模式下有多个匹配，记录警告并返回第一个（最安全的选择）
+    logger.warning(f"表名 '{table_name}' 匹配到多个表，静默模式下默认选择第一个: {matches[0]['name']}")
+    return matches[0]["id"]
+
+
+def resolve_ambiguities(actions: list[dict[str, Any]], project_path: str) -> bool:
+    """解决动作中的歧义。返回是否继续执行。
+
+    遍历 AI 返回的动作列表，解析表名和列名的歧义。
+    对于模糊匹配的表名，弹出交互式菜单让用户选择。
+    标记为跳过的动作会在最后过滤掉。
+
+    Args:
+        actions: AI 返回的动作列表
+        project_path: 项目路径
+
+    Returns:
+        是否还有有效动作需要继续执行
+    """
+    if not actions or not project_path:
+        return True
+
+    for action in actions:
+        action_type = action.get("actionType", "")
+        spec = action.get("constraintSpec", {})
+
+        # 处理 VALIDATE_PROJECT 的表名解析
+        if action_type == "VALIDATE_PROJECT":
+            table_name = spec.get("tableName") or spec.get("targetNodeId")
+            tables = spec.get("tables", [])
+
+            # 解析单个表名
+            if table_name:
+                table_id = resolve_table_name(table_name, project_path)
+                if table_id:
+                    spec["targetNodeId"] = table_id
+                    spec["tableName"] = table_name
+                else:
+                    action["_skip"] = True
+                    continue
+
+            # 解析表名列表
+            elif tables:
+                resolved_ids = []
+                for name in tables:
+                    table_id = resolve_table_name(name, project_path, silent=True)
+                    if table_id:
+                        resolved_ids.append(table_id)
+                if resolved_ids:
+                    spec["tableIds"] = resolved_ids
+                else:
+                    action["_skip"] = True
+                    continue
+            continue
+
+        # Schema/Regex/Transform/Settings 操作不需要表名歧义解析
+        non_table_types = {
+            "ADD_SCHEMA",
+            "UPDATE_SCHEMA",
+            "DELETE_SCHEMA",
+            "ADD_REGEX",
+            "UPDATE_REGEX",
+            "DELETE_REGEX",
+            "ADD_TRANSFORM",
+            "UPDATE_TRANSFORM",
+            "DELETE_TRANSFORM",
+            "UPDATE_SETTINGS",
+        }
+        if action_type in non_table_types:
+            continue
+
+        # 处理约束操作的表名解析
+        table_id = spec.get("targetNodeId")
+        table_name = spec.get("tableName")
+
+        # 尝试匹配
+        identifier = table_id or table_name
+        if not identifier:
+            continue
+
+        matches = find_matching_schemas(project_path, identifier)
+
+        # 如果有多个模糊匹配，且没有精确匹配 ID 的情况
+        if len(matches) > 1:
+            print(f"\n{Formatter.warning('[!]')} 发现多个匹配的表 '{identifier}':")
+
+            menu = InteractiveMenu("请选择正确的表:")
+            for m in matches:
+                # 检查是否为加密 ID 格式，若是则显示一部分，否则显示全名
+                display_id = f"{m['id'][:10]}..." if len(m["id"]) > 15 else m["id"]
+                menu.add_item(m["id"], f"{m['name']} ({display_id})", f"ID: {m['id']}")
+
+            menu.add_item("_skip_", "跳过此操作", "不执行此约束更新")
+
+            choice = menu.show()
+            if choice is None or choice == "_skip_":
+                action["_skip"] = True
+                continue
+
+            # 找到选中的匹配项
+            selected = next(m for m in matches if m["id"] == choice)
+
+            # 更新 Action 中的关键字段
+            spec["targetNodeId"] = selected["id"]
+            spec["tableName"] = selected["name"]
+
+            # 同样更新列 (如果是根据名称匹配，可能也存在列名冲突)
+            # 这里简化处理，假设列在特定表内唯一
+            print(f"{Formatter.success('[OK]')} 已选择表: {selected['name']}")
+        elif len(matches) == 1:
+            # 只有一个匹配，直接更新为表ID
+            selected = matches[0]
+            spec["targetNodeId"] = selected["id"]
+            spec["tableName"] = selected["name"]
+        else:
+            # 4.5: 0 匹配跳过并提示——原实现静默继续（拼错的表名让约束作用于
+            # 空表集后"静默成功"），与 VALIDATE 分支的 _skip 行为对齐
+            print(f"{Formatter.warning('[!]')} 未找到表 '{identifier}'，已跳过该约束操作")
+            action["_skip"] = True
+            continue
+
+    # 统计并提示被跳过的动作
+    skipped_actions = [a for a in actions if a.get("_skip")]
+    if skipped_actions:
+        print(Formatter.info(f"\n[i] 已跳过 {len(skipped_actions)} 个操作（用户取消或表名未找到）"))
+
+    # 过滤掉标记为跳过的动作
+    actions[:] = [a for a in actions if not a.get("_skip")]
+    return len(actions) > 0

app/cli/shell/commands/ai/status.py

@@ -0,0 +1,101 @@
+# SPDX-License-Identifier: Apache-2.0
+#
+# Copyright 2026 Precis Team
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# backend/app/cli/shell/commands/ai/status.py
+"""
+@fileoverview AI 配置状态命令模块
+
+功能概述:
+- 提供 ai status 子命令显示当前 AI Provider 配置状态
+- 列出所有已配置 Provider，标记当前默认项与密钥状态
+- 显示 Provider 的模型、端点等信息
+
+架构设计:
+- AIStatusCommand 继承 Command 基类
+- 通过 CLI 配置管理器获取 Provider 列表和默认项
+- 使用 Formatter 进行颜色高亮输出
+
+输入示例:
+    ai status
+
+输出示例:
+    带颜色高亮的 Provider 配置状态文本
+"""
+
+from app.cli.shell.commands.base import Command, CommandResult, ProjectContext
+from app.cli.shell.config_storage import get_cli_config
+from app.cli.shell.formatter import Formatter
+
+
+class AIStatusCommand(Command):
+    """AI 配置状态命令。
+
+    显示所有已配置的 AI Provider 及其状态信息。
+    """
+
+    def __init__(self) -> None:
+        super().__init__("status")
+        self._cli_config = get_cli_config()
+
+    @property
+    def description(self) -> str:
+        return "显示 AI 配置状态"
+
+    @property
+    def usage(self) -> str:
+        return "ai status"
+
+    def execute(self, args: list[str], context: ProjectContext) -> CommandResult:
+        """执行状态查看命令。
+
+        Args:
+            args: 命令参数列表（此命令不需要参数）
+            context: 项目上下文
+
+        Returns:
+            格式化的配置状态文本
+        """
+        providers = self._cli_config.list_providers()
+
+        lines = [Formatter.header("\nCLI AI 配置状态")]
+
+        if not providers:
+            lines.append(Formatter.warning("\n未配置任何 AI Provider"))
+            lines.append(Formatter.info("请编辑 ~/.precis/ai_providers.yaml 配置文件"))
+        else:
+            lines.append(f"\n已配置的 Providers ({len(providers)}):")
+            active = self._cli_config.get_active_provider()
+
+            for p in providers:
+                is_active = active and active.id == p.id
+                status_icon = "*" if is_active else " "
+                has_key = "[有密钥]" if p.api_key else "[无密钥]"
+
+                lines.append(f"\n  [{status_icon}] {p.name} ({p.id})")
+                lines.append(f"     模型: {p.model}")
+                lines.append(f"     API Key: {has_key}")
+                if p.base_url:
+                    lines.append(f"     端点: {p.base_url}")
+
+            if active:
+                lines.append("\n" + Formatter.success("[*] 当前默认: " + active.name))
+            else:
+                lines.append("\n" + Formatter.warning("[!] 没有可用的默认 Provider"))
+
+        lines.append(Formatter.info("\n配置文件: ~/.precis/ai_providers.yaml"))
+        lines.append(Formatter.info("API Key 设置方式: 环境变量"))
+        lines.append(Formatter.info("  例如: OPENAI_API_KEY=sk-xxx"))
+
+        return CommandResult.ok("\n".join(lines))