liusongtao
/
AIDeveloper-PC
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

支持功能模块

This commit is contained in:
liusongtao 2026-03-05 13:38:26 +08:00
parent 3fc095652e
commit 29636b9b94
12 changed files with 1142 additions and 1608 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

209
requirements_generator/config.py
View File

@ -1,146 +1,123 @@
# config.py - 全局配置管理
# config.py - 全局配置
import os
from dotenv import load_dotenv
load_dotenv()
# ── LLM 配置 ──────────────────────────────────────────
LLM_API_KEY = os.getenv("OPENAI_API_KEY", "")
LLM_BASE_URL = os.getenv("OPENAI_BASE_URL", "https://api.openai.com/v1")
LLM_MODEL = os.getenv("LLM_MODEL", "gpt-4o")
LLM_TEMPERATURE = float(os.getenv("LLM_TEMPERATURE", "0.3"))
# ── LLM ──────────────────────────────────────────────
LLM_API_KEY = os.getenv("OPENAI_API_KEY", "")
LLM_API_BASE = os.getenv("OPENAI_BASE_URL", "https://api.openai.com/v1")
LLM_MODEL = os.getenv("LLM_MODEL", "gpt-4o")
LLM_TIMEOUT = int(os.getenv("LLM_TIMEOUT", "60"))
LLM_MAX_RETRY = int(os.getenv("LLM_MAX_RETRY", "3"))
# ── 数据库配置 ─────────────────────────────────────────
DB_PATH = os.getenv("DB_PATH", "data/requirement_analyzer.db")
# ── 数据库 ────────────────────────────────────────────
DB_PATH = os.getenv("DB_PATH", "data/requirement_analyzer.db")
# ── 输出配置 ───────────────────────────────────────────
# ── 输出目录 ──────────────────────────────────────────
OUTPUT_BASE_DIR = os.getenv("OUTPUT_BASE_DIR", "output")
DEFAULT_LANGUAGE = os.getenv("DEFAULT_LANGUAGE", "python")
DEFAULT_MODULE = os.getenv("DEFAULT_MODULE", "default")
# ══════════════════════════════════════════════════════
# Prompt 模板
# ══════════════════════════════════════════════════════
# ── Prompt 模板 ───────────────────────────────────────
DECOMPOSE_PROMPT_TEMPLATE = """
你是一位资深软件架构师和产品经理。请根据以下信息，将原始需求分解为若干个可独立实现的功能需求。
DECOMPOSE_PROMPT_TEMPLATE = """\
你是一名资深软件架构师，请将以下原始需求分解为独立的功能需求列表。
{knowledge_section}
## 原始需求
【原始需求】
{raw_requirement}
## 输出要求
请严格按照以下 JSON 格式输出，不要包含任何额外说明：
{{
"functional_requirements": [
{{
"index": 1,
"title": "功能需求标题（简洁，10字以内）",
"description": "功能需求详细描述（包含输入、处理逻辑、输出）",
"function_name": "snake_case函数名",
"priority": "high|medium|low"
}}
]
}}
要求：
1. 每个功能需求必须是独立可实现的最小单元
2. function_name 使用 snake_case 命名，清晰表达函数用途
3. 分解粒度适中，通常 5-15 个功能需求
4. 优先级根据业务重要性判断
"""
# ── 函数签名 JSON 生成 Prompt ──────────────────────────
FUNC_SIGNATURE_PROMPT_TEMPLATE = """
你是一位资深软件架构师。请根据以下功能需求描述，设计该函数的完整接口签名，并以 JSON 格式输出。
{knowledge_section}
## 功能需求
需求编号：{requirement_id}
标题：{title}
函数名：{function_name}
详细描述：{description}
【输出要求】
以 JSON 数组格式输出，每个元素包含以下字段：
- title: 功能标题（简短，10字以内）
- description: 功能描述（详细说明该功能的职责与边界，50字以内）
- function_name: 对应的函数名（snake_case，动词开头）
- priority: 优先级（high / medium / low）
- module: 所属功能模块名称（snake_case，如 user_auth / order_service）
## 输出格式
请严格按照以下 JSON 结构输出，不要包含任何额外说明或 markdown 标记：
{{
"name": "{function_name}",
"requirement_id": "{requirement_id}",
"description": "简洁的一句话功能描述（英文）",
"type": "function",
"parameters": {{
"<param_name>": {{
"type": "integer|string|boolean|float|list|dict|object",
"inout": "in|out|inout",
"description": "参数说明（英文）",
"required": true
}}
}},
"return": {{
"type": "integer|string|boolean|float|list|dict|object|void",
"description": "整体返回值说明（英文，一句话概括）",
"on_success": {{
"value": "具体成功返回值或范围，如 0、true、user object、list of items 等",
"description": "成功时的返回值含义（英文）"
}},
"on_failure": {{
"value": "具体失败返回值或范围，如 nonzero、false、null、empty list、raises Exception 等",
"description": "失败时的返回值含义，或抛出的异常类型（英文）"
}}
【示例输出】
[
{{
"title": "用户注册",
"description": "接收用户名、密码、邮箱，校验合法性后创建用户账号并返回用户ID",
"function_name": "register_user",
"priority": "high",
"module": "user_auth"
}}
}}
]
## 设计规范
1. 参数名使用 snake_case，类型使用通用类型（不绑定具体语言）
2. inout 字段含义：
- in = 仅输入参数
- out = 仅输出参数（通过参数传出结果，如指针/引用）
- inout = 既作输入又作输出
3. 所有描述字段使用英文
4. return 字段规则：
- 若函数无返回值（void），type 填 "void"，on_success/on_failure 均填 null
- 若返回值只有成功场景（如纯查询），on_failure 可描述为 "null or empty"
- on_success.value / on_failure.value 填写具体值或值域描述，不要填写空字符串
5. 若函数无参数，parameters 填 {{}}
6. required 字段为布尔值 true 或 false
只输出 JSON 数组，不要有任何额外说明。
"""
# ── 代码生成 Prompt（含签名约束）─────────────────────────
CODE_GEN_PROMPT_TEMPLATE = """
你是一位资深 {language} 工程师。请根据以下功能需求和【函数签名规范】，生成完整的 {language} 函数代码。
FUNC_SIGNATURE_PROMPT_TEMPLATE = """\
你是一名资深软件工程师，请根据以下功能需求生成标准函数签名信息。
【功能需求】
- 需求编号: {requirement_id}
- 标题: {title}
- 描述: {description}
- 函数名: {function_name}
- 所属模块: {module}
{knowledge_section}
## 功能需求
标题：{title}
描述：{description}
【输出要求】
以 JSON 对象格式输出，包含以下字段：
- name: 函数名（与上方一致）
- requirement_id: 需求编号
- description: 函数功能描述（英文，一句话）
- type: 固定为 "function"
- module: 所属模块名称
- parameters: 参数字典，key 为参数名，value 包含：
- type: 数据类型（integer/string/boolean/float/list/dict/object/void/any）
- inout: in / out / inout
- required: true / false
- description: 参数说明
- return:
- type: 返回类型
- on_success: {{ "value": "...", "description": "..." }} 或 null（void）
- on_failure: {{ "value": "...", "description": "..." }} 或 null（void）
## 【必须严格遵守】函数签名规范
以下 JSON 定义了函数的精确接口，生成的代码必须与之完全一致，不得擅自增减或改名参数：
只输出 JSON 对象，不要有任何额外说明。
"""
```json
CODE_GEN_PROMPT_TEMPLATE = """\
你是一名资深 {language} 工程师，请根据以下函数签名和功能描述生成完整的函数实现代码。
【函数签名】
{signature_json}
```
### 签名字段说明
- `name`：函数名，必须完全一致
- `parameters`：每个 key 即为参数名，`type` 为数据类型，`inout` 含义：
- `in` = 普通输入参数
- `out` = 输出参数（Python 中通过返回值或可变容器传出）
- `inout` = 既作输入又作输出
- `return.type`：返回值类型
- `return.on_success`：成功时的返回值，代码实现必须与此一致
- `return.on_failure`：失败时的返回值或异常，代码实现必须与此一致
【功能描述】
{description}
## 输出要求
1. 只输出纯代码，不要包含 markdown 代码块标记
2. 函数签名（名称、参数列表、返回类型）必须与上方 JSON 规范完全一致
3. 成功/失败的返回值必须严格遵守 return.on_success / return.on_failure 的定义
4. 包含完整的类型注解（Python 使用 type hints）
5. 包含详细的 docstring，其中 Returns 段须注明成功值与失败值
6. 包含必要的异常处理
7. 代码风格遵循 PEP8（Python）或对应语言规范
8. 在文件顶部用注释注明：需求编号、功能标题、函数签名摘要
9. 如需导入第三方库，请在顶部统一导入
{knowledge_section}
【输出要求】
1. 只输出 {language} 代码，不要有任何 Markdown 标记（不要 ```）
2. 包含完整的函数实现（含必要的 import）
3. 包含函数文档注释（docstring / JSDoc 等）
4. 包含基本的参数校验与错误处理
5. 代码风格遵循 {language} 最佳实践
"""
MODULE_CLASSIFY_PROMPT_TEMPLATE = """\
你是一名资深软件架构师，请将以下功能需求列表分类到合适的功能模块中。
【功能需求列表】
{requirements_json}
【输出要求】
以 JSON 数组格式输出，每个元素包含：
- function_name: 函数名（与输入一致）
- module: 所属模块名称（snake_case，如 user_auth / order_service / payment）
模块划分原则：
1. 功能相近的需求归入同一模块
2. 模块名使用英文 snake_case
3. 模块数量控制在 2~8 个之间
4. 若某需求确实无法归类，使用 "default" 模块
只输出 JSON 数组，不要有任何额外说明。
"""

241
requirements_generator/core/code_generator.py
View File

@ -1,99 +1,89 @@
# core/code_generator.py - 代码生成核心逻辑（签名约束版）
# core/code_generator.py - 代码生成（按模块路由到子目录）
import os
import json
from typing import Optional, List, Callable
from pathlib import Path
from typing import List, Optional, Callable
import config
from core.llm_client import LLMClient
from database.models import FunctionalRequirement, CodeFile
from utils.output_writer import write_code_file, get_file_extension
class CodeGenerator:
"""
根据功能需求 + 函数签名约束，使用 LLM 生成代码函数文件。
签名由 RequirementAnalyzer.build_function_signature() 预先生成，
注入 Prompt 后可确保代码参数列表与签名 JSON 完全一致。
"""
"""根据函数签名约束，调用 LLM 生成代码文件，并按模块写入子目录"""
def __init__(self, llm_client: Optional[LLMClient] = None):
"""
初始化代码生成器
Args:
llm_client: LLM 客户端实例，为 None 时自动创建
"""
self.llm = llm_client or LLMClient()
def __init__(self, llm: LLMClient):
self.llm = llm
# ══════════════════════════════════════════════════
# 单个生成
# 单个代码文件生成
# ══════════════════════════════════════════════════
def generate(
self,
func_req: FunctionalRequirement,
func_req: FunctionalRequirement,
output_dir: str,
language: str = config.DEFAULT_LANGUAGE,
knowledge: str = "",
signature: Optional[dict] = None,
language: str = None,
knowledge: str = "",
signature: dict = None,
) -> CodeFile:
"""
为单个功能需求生成代码文件
为单个功能需求生成代码文件，写入 output_dir/<module>/ 子目录。
Args:
func_req: 功能需求对象（必须含有效 id）
output_dir: 代码输出目录
language: 目标编程语言
knowledge: 知识库文本（可选）
signature: 函数签名 dict（由 RequirementAnalyzer 生成）。
传入后将作为强约束注入 Prompt，确保代码参数
与签名 JSON 完全一致；为 None 时退化为无约束模式。
func_req: 功能需求对象
output_dir: 项目根输出目录
language: 目标语言
knowledge: 知识库文本
signature: 函数签名 dict（可选，有则作为约束）
Returns:
CodeFile 对象（含生成的代码内容和文件路径，未持久化）
CodeFile 对象（未持久化）
Raises:
ValueError: func_req.id 为 None
Exception: LLM 调用失败或文件写入失败
RuntimeError: LLM 调用失败
"""
if func_req.id is None:
raise ValueError("FunctionalRequirement 必须先持久化（id 不能为 None）")
language = language or config.DEFAULT_LANGUAGE
module = (func_req.module or config.DEFAULT_MODULE).strip()
knowledge_section = self._build_knowledge_section(knowledge)
signature_json = self._build_signature_json(signature, func_req)
# 按模块创建子目录
module_dir = os.path.join(output_dir, module)
os.makedirs(module_dir, exist_ok=True)
self._ensure_init_py(module_dir)
# 构建 Prompt
sig_json = json.dumps(signature, ensure_ascii=False, indent=2) if signature else "{}"
knowledge_section = f"【参考知识库】\n{knowledge}\n" if knowledge else ""
prompt = config.CODE_GEN_PROMPT_TEMPLATE.format(
language=language,
knowledge_section=knowledge_section,
title=func_req.title,
description=func_req.description,
signature_json=signature_json,
language = language,
signature_json = sig_json,
description = func_req.description,
knowledge_section = knowledge_section,
)
code_content = self.llm.chat(
system_prompt=(
f"你是一位资深 {language} 工程师，只输出纯代码，"
"不添加任何 markdown 标记。函数签名必须与提供的 JSON 规范完全一致。"
),
user_prompt=prompt,
)
try:
code_content = self.llm.chat(
prompt,
system = f"You are an expert {language} developer. Output only code.",
temperature = 0.2,
max_tokens = 4096,
)
except Exception as e:
raise RuntimeError(f"代码生成失败 [{func_req.function_name}]: {e}")
file_path = write_code_file(
output_dir=output_dir,
function_name=func_req.function_name,
language=language,
content=code_content,
)
ext = get_file_extension(language)
# 写入文件
ext = self._get_extension(language)
file_name = f"{func_req.function_name}{ext}"
file_path = os.path.join(module_dir, file_name)
Path(file_path).write_text(code_content, encoding="utf-8")
return CodeFile(
project_id=func_req.project_id,
func_req_id=func_req.id,
file_name=file_name,
file_path=file_path,
language=language,
content=code_content,
func_req_id = func_req.id,
file_name = file_name,
file_path = file_path,
module = module,
language = language,
content = code_content,
)
# ══════════════════════════════════════════════════
@ -102,115 +92,82 @@ class CodeGenerator:
def generate_batch(
self,
func_reqs: List[FunctionalRequirement],
output_dir: str,
language: str = config.DEFAULT_LANGUAGE,
knowledge: str = "",
signatures: Optional[List[dict]] = None,
on_progress: Optional[Callable] = None,
func_reqs: List[FunctionalRequirement],
output_dir: str,
language: str = None,
knowledge: str = "",
signatures: Optional[List[dict]] = None,
on_progress: Optional[Callable] = None,
) -> List[CodeFile]:
"""
批量生成代码文件
批量生成代码文件。
Args:
func_reqs: 功能需求列表
output_dir: 输出目录
output_dir: 项目根输出目录
language: 目标语言
knowledge: 知识库文本
signatures: 与 func_reqs 等长的签名列表（索引对应）。
为 None 时所有条目均以无约束模式生成。
on_progress: 进度回调 fn(index, total, func_req, code_file, error)
signatures: 与 func_reqs 等长的签名列表（索引对应）
on_progress: 进度回调 fn(index, total, req, code_file, error)
Returns:
成功生成的 CodeFile 列表
"""
results = []
total = len(func_reqs)
language = language or config.DEFAULT_LANGUAGE
total = len(func_reqs)
results = []
sig_map = self._build_signature_map(func_reqs, signatures)
# 构建 func_req.id → signature 的快速查找表
sig_map = self._build_signature_map(func_reqs, signatures)
for i, req in enumerate(func_reqs):
for i, req in enumerate(func_reqs, 1):
sig = sig_map.get(req.id)
try:
code_file = self.generate(
func_req=req,
output_dir=output_dir,
language=language,
knowledge=knowledge,
signature=sig,
func_req = req,
output_dir = output_dir,
language = language,
knowledge = knowledge,
signature = sig,
)
results.append(code_file)
if on_progress:
on_progress(i + 1, total, req, code_file, None)
on_progress(i, total, req, code_file, None)
except Exception as e:
if on_progress:
on_progress(i + 1, total, req, None, e)
on_progress(i, total, req, None, e)
return results
# ══════════════════════════════════════════════════
# 私有工具方法
# 工具方法
# ══════════════════════════════════════════════════
@staticmethod
def _build_knowledge_section(knowledge: str) -> str:
"""构建知识库 Prompt 段落"""
if not knowledge or not knowledge.strip():
return ""
return (
"## 参考知识库（实现时请遵循以下规范）\n"
f"{knowledge}\n\n---\n"
)
@staticmethod
def _build_signature_json(
signature: Optional[dict],
func_req: FunctionalRequirement,
) -> str:
"""
将签名 dict 序列化为格式化 JSON 字符串；
若签名为 None，则构造最小占位签名，保持 Prompt 结构完整。
Args:
signature: 签名 dict 或 None
func_req: 对应的功能需求（用于占位签名）
Returns:
JSON 字符串
"""
if signature:
return json.dumps(signature, ensure_ascii=False, indent=2)
# 无签名时的最小占位，提示 LLM 自行设计但保持格式
fallback = {
"name": func_req.function_name,
"requirement_id": f"REQ.{func_req.index_no:02d}",
"description": func_req.description,
"type": "function",
"parameters": "<<请根据功能描述自行设计参数>>",
"return": "<<请根据功能描述自行设计返回值>>",
}
return json.dumps(fallback, ensure_ascii=False, indent=2)
@staticmethod
def _build_signature_map(
func_reqs: List[FunctionalRequirement],
func_reqs: List[FunctionalRequirement],
signatures: Optional[List[dict]],
) -> dict:
"""
构建 func_req.id → signature 映射表
Args:
func_reqs: 功能需求列表
signatures: 与 func_reqs 等长的签名列表，或 None
Returns:
{req_id: signature_dict} 字典
"""
"""构建 func_req.id → signature 的快速查找表"""
if not signatures:
return {}
sig_map = {}
for req, sig in zip(func_reqs, signatures):
if req.id is not None and sig:
sig_map[req.id] = sig
return sig_map
return {
req.id: sig
for req, sig in zip(func_reqs, signatures)
if req.id is not None
}
@staticmethod
def _get_extension(language: str) -> str:
ext_map = {
"python": ".py", "javascript": ".js", "typescript": ".ts",
"java": ".java", "go": ".go", "rust": ".rs",
"cpp": ".cpp", "c": ".c", "csharp": ".cs",
"ruby": ".rb", "php": ".php", "swift": ".swift", "kotlin": ".kt",
}
return ext_map.get(language.lower(), ".txt")
@staticmethod
def _ensure_init_py(directory: str) -> None:
"""在目录中创建 __init__.py（Python 包标识）"""
init = os.path.join(directory, "__init__.py")
if not os.path.exists(init):
Path(init).write_text("# Auto-generated module package\n", encoding="utf-8")

160
requirements_generator/core/llm_client.py
View File

@ -1,90 +1,114 @@
# core/llm_client.py - LLM 客户端封装
# core/llm_client.py - LLM API 调用封装
import time
import json
from typing import Optional
from openai import OpenAI, APIError, APITimeoutError, RateLimitError
import config
class LLMClient:
"""
OpenAI 兼容 LLM 客户端封装。
支持任何兼容 OpenAI API 格式的服务（OpenAI / Azure / 本地模型等）。
"""
"""封装 OpenAI 兼容接口，提供统一的调用入口与重试机制"""
def __init__(
self,
api_key: str = config.LLM_API_KEY,
base_url: str = config.LLM_BASE_URL,
model: str = config.LLM_MODEL,
temperature: float = config.LLM_TEMPERATURE,
api_key: str = None,
api_base: str = None,
model: str = None,
):
"""
初始化 LLM 客户端
Args:
api_key: API 密钥
base_url: API 基础 URL
model: 模型名称
temperature: 生成温度（0~1，越低越确定）
Raises:
ImportError: 未安装 openai 库
ValueError: api_key 为空
"""
try:
from openai import OpenAI
except ImportError:
raise ImportError("请安装 openai: pip install openai")
if not api_key:
raise ValueError("LLM_API_KEY 未配置，请在 .env 文件中设置")
self.model = model
self.temperature = temperature
self._client = OpenAI(api_key=api_key, base_url=base_url)
def chat(self, system_prompt: str, user_prompt: str) -> str:
"""
发送对话请求，返回模型回复文本
Args:
system_prompt: 系统提示词
user_prompt: 用户输入
Returns:
模型回复的文本内容
Raises:
Exception: API 调用失败
"""
response = self._client.chat.completions.create(
model=self.model,
temperature=self.temperature,
messages=[
{"role": "system", "content": system_prompt},
{"role": "user", "content": user_prompt},
],
self.model = model or config.LLM_MODEL
self.client = OpenAI(
api_key = api_key or config.LLM_API_KEY,
base_url = api_base or config.LLM_API_BASE,
timeout = config.LLM_TIMEOUT,
)
return response.choices[0].message.content.strip()
def chat_json(self, system_prompt: str, user_prompt: str) -> dict:
def chat(
self,
prompt: str,
system: str = "You are a helpful assistant.",
temperature: float = 0.2,
max_tokens: int = 4096,
) -> str:
"""
发送对话请求，解析并返回 JSON 结果
发送单轮对话请求，返回模型回复文本。
Args:
system_prompt: 系统提示词
user_prompt: 用户输入
prompt: 用户消息
system: 系统提示词
temperature: 采样温度
max_tokens: 最大输出 token 数
Returns:
解析后的 dict 对象
模型回复的纯文本字符串
Raises:
json.JSONDecodeError: 模型返回非合法 JSON
RuntimeError: 超过最大重试次数后仍失败
"""
raw = self.chat(system_prompt, user_prompt)
last_error = None
for attempt in range(1, config.LLM_MAX_RETRY + 1):
try:
resp = self.client.chat.completions.create(
model = self.model,
messages = [
{"role": "system", "content": system},
{"role": "user", "content": prompt},
],
temperature = temperature,
max_tokens = max_tokens,
)
return resp.choices[0].message.content.strip()
except RateLimitError as e:
wait = 2 ** attempt
last_error = e
time.sleep(wait)
except APITimeoutError as e:
last_error = e
time.sleep(2)
except APIError as e:
last_error = e
if attempt >= config.LLM_MAX_RETRY:
break
time.sleep(1)
raise RuntimeError(
f"LLM 调用失败（已重试 {config.LLM_MAX_RETRY} 次）: {last_error}"
)
def chat_json(
self,
prompt: str,
system: str = "You are a helpful assistant. Always respond with valid JSON.",
temperature: float = 0.1,
max_tokens: int = 4096,
) -> any:
"""
发送请求并将回复解析为 JSON 对象。
Returns:
解析后的 Python 对象（dict 或 list）
Raises:
ValueError: JSON 解析失败
RuntimeError: LLM 调用失败
"""
raw = self.chat(prompt, system=system, temperature=temperature, max_tokens=max_tokens)
# 去除可能的 markdown 代码块包裹
raw = raw.strip()
if raw.startswith("```"):
lines = raw.split("\n")
raw = "\n".join(lines[1:-1])
return json.loads(raw)
cleaned = self._strip_markdown_code_block(raw)
try:
return json.loads(cleaned)
except json.JSONDecodeError as e:
raise ValueError(f"LLM 返回内容无法解析为 JSON: {e}\n原始内容:\n{raw}")
@staticmethod
def _strip_markdown_code_block(text: str) -> str:
"""去除 ```json ... ``` 或 ``` ... ``` 包裹"""
text = text.strip()
if text.startswith("```"):
lines = text.splitlines()
# 去掉首行（```json 或 ```）和末行（```）
inner = lines[1:] if lines[-1].strip() == "```" else lines[1:]
if inner and inner[-1].strip() == "```":
inner = inner[:-1]
text = "\n".join(inner).strip()
return text

298
requirements_generator/core/requirement_analyzer.py
View File

@ -1,6 +1,6 @@
# core/requirement_analyzer.py - 需求分解 & 函数签名生成
import re
from typing import List, Optional
import json
from typing import List, Optional, Callable
import config
from core.llm_client import LLMClient
@ -8,19 +8,10 @@ from database.models import FunctionalRequirement
class RequirementAnalyzer:
"""
使用 LLM 将原始需求分解为功能需求列表，并生成函数接口签名。
支持注入知识库上下文以提升分解质量。
"""
"""负责需求分解、模块分类、函数签名生成"""
def __init__(self, llm_client: Optional[LLMClient] = None):
"""
初始化需求分析器
Args:
llm_client: LLM 客户端实例，为 None 时自动创建
"""
self.llm = llm_client or LLMClient()
def __init__(self, llm: LLMClient):
self.llm = llm
# ══════════════════════════════════════════════════
# 需求分解
@ -29,12 +20,12 @@ class RequirementAnalyzer:
def decompose(
self,
raw_requirement: str,
project_id: int,
raw_req_id: int,
knowledge: str = "",
project_id: int,
raw_req_id: int,
knowledge: str = "",
) -> List[FunctionalRequirement]:
"""
将原始需求分解为功能需求列表
将原始需求文本分解为功能需求列表（含模块分类）。
Args:
raw_requirement: 原始需求文本
@ -44,196 +35,175 @@ class RequirementAnalyzer:
Returns:
FunctionalRequirement 对象列表（未持久化，id=None）
Raises:
ValueError: LLM 返回格式不合法
json.JSONDecodeError: JSON 解析失败
"""
knowledge_section = self._build_knowledge_section(knowledge)
knowledge_section = (
f"【参考知识库】\n{knowledge}\n" if knowledge else ""
)
prompt = config.DECOMPOSE_PROMPT_TEMPLATE.format(
knowledge_section=knowledge_section,
raw_requirement=raw_requirement,
raw_requirement = raw_requirement,
knowledge_section = knowledge_section,
)
result = self.llm.chat_json(
system_prompt="你是一位资深软件架构师，擅长需求分析与系统设计。",
user_prompt=prompt,
)
try:
items = self.llm.chat_json(prompt)
if not isinstance(items, list):
raise ValueError("LLM 返回结果不是数组")
except Exception as e:
raise RuntimeError(f"需求分解失败: {e}")
items = result.get("functional_requirements", [])
if not items:
raise ValueError("LLM 未返回任何功能需求，请检查原始需求描述")
requirements = []
for item in items:
reqs = []
for i, item in enumerate(items, 1):
req = FunctionalRequirement(
project_id=project_id,
raw_req_id=raw_req_id,
index_no=int(item.get("index", len(requirements) + 1)),
title=item.get("title", "未命名功能"),
description=item.get("description", ""),
function_name=self._sanitize_function_name(
item.get("function_name", f"func_{len(requirements)+1}")
),
priority=item.get("priority", "medium"),
project_id = project_id,
raw_req_id = raw_req_id,
index_no = i,
title = item.get("title", f"功能{i}"),
description = item.get("description", ""),
function_name = item.get("function_name", f"function_{i}"),
priority = item.get("priority", "medium"),
module = item.get("module", config.DEFAULT_MODULE),
status = "pending",
is_custom = False,
)
requirements.append(req)
return requirements
reqs.append(req)
return reqs
# ══════════════════════════════════════════════════
# 函数签名生成（新增）
# 模块分类（独立步骤，可对已有需求列表重新分类）
# ══════════════════════════════════════════════════
def classify_modules(
self,
func_reqs: List[FunctionalRequirement],
knowledge: str = "",
) -> List[dict]:
"""
对功能需求列表进行模块分类，返回 {function_name: module} 映射列表。
Args:
func_reqs: 功能需求列表
knowledge: 知识库文本（可选）
Returns:
[{"function_name": "...", "module": "..."}, ...]
"""
req_list = [
{
"index_no": r.index_no,
"title": r.title,
"description": r.description,
"function_name": r.function_name,
}
for r in func_reqs
]
knowledge_section = f"【参考知识库】\n{knowledge}\n" if knowledge else ""
prompt = config.MODULE_CLASSIFY_PROMPT_TEMPLATE.format(
requirements_json = json.dumps(req_list, ensure_ascii=False, indent=2),
knowledge_section = knowledge_section,
)
try:
result = self.llm.chat_json(prompt)
if not isinstance(result, list):
raise ValueError("LLM 返回结果不是数组")
return result
except Exception as e:
raise RuntimeError(f"模块分类失败: {e}")
# ══════════════════════════════════════════════════
# 函数签名生成
# ══════════════════════════════════════════════════
def build_function_signature(
self,
func_req: FunctionalRequirement,
knowledge: str = "",
func_req: FunctionalRequirement,
requirement_id: str = "",
knowledge: str = "",
) -> dict:
"""
为单个功能需求生成函数接口签名 JSON
为单个功能需求生成函数签名 dict。
Args:
func_req: 功能需求对象（需含有效 id）
knowledge: 知识库文本（可选）
func_req: 功能需求对象
requirement_id: 需求编号字符串（如 "REQ.01"）
knowledge: 知识库文本
Returns:
符合接口规范的 dict，包含 name/requirement_id/description/
type/parameters/return 字段
函数签名 dict
Raises:
json.JSONDecodeError: LLM 返回非合法 JSON
RuntimeError: LLM 调用或解析失败
"""
requirement_id = self._format_requirement_id(func_req.index_no)
knowledge_section = self._build_knowledge_section(knowledge)
knowledge_section = f"【参考知识库】\n{knowledge}\n" if knowledge else ""
prompt = config.FUNC_SIGNATURE_PROMPT_TEMPLATE.format(
knowledge_section=knowledge_section,
requirement_id=requirement_id,
title=func_req.title,
function_name=func_req.function_name,
description=func_req.description,
requirement_id = requirement_id or f"REQ.{func_req.index_no:02d}",
title = func_req.title,
description = func_req.description,
function_name = func_req.function_name,
module = func_req.module or config.DEFAULT_MODULE,
knowledge_section = knowledge_section,
)
signature = self.llm.chat_json(
system_prompt=(
"你是一位资深软件架构师，专注于 API 接口设计。"
"只输出合法 JSON，不添加任何说明文字。"
),
user_prompt=prompt,
)
# 确保关键字段存在，做兜底处理
signature.setdefault("name", func_req.function_name)
signature.setdefault("requirement_id", requirement_id)
signature.setdefault("description", func_req.description)
signature.setdefault("type", "function")
signature.setdefault("parameters", {})
signature.setdefault("return", {"type": "void", "description": ""})
return signature
try:
sig = self.llm.chat_json(prompt)
if not isinstance(sig, dict):
raise ValueError("LLM 返回结果不是 dict")
# 确保 module 字段存在
if "module" not in sig:
sig["module"] = func_req.module or config.DEFAULT_MODULE
return sig
except Exception as e:
raise RuntimeError(f"签名生成失败 [{func_req.function_name}]: {e}")
def build_function_signatures_batch(
self,
func_reqs: List[FunctionalRequirement],
knowledge: str = "",
on_progress=None,
func_reqs: List[FunctionalRequirement],
knowledge: str = "",
on_progress: Optional[Callable] = None,
) -> List[dict]:
"""
批量为功能需求列表生成函数接口签名
批量生成函数签名，失败时使用降级结构。
Args:
func_reqs: 功能需求列表
knowledge: 知识库文本（可选）
on_progress: 进度回调 fn(index, total, func_req, signature, error)
knowledge: 知识库文本
on_progress: 进度回调 fn(index, total, req, signature, error)
Returns:
签名 dict 列表，顺序与 func_reqs 一致；
生成失败的条目使用降级结构填充，不中断整体流程
与 func_reqs 等长的签名 dict 列表（索引一一对应）
"""
results = []
total = len(func_reqs)
signatures = []
total = len(func_reqs)
for i, req in enumerate(func_reqs):
for i, req in enumerate(func_reqs, 1):
req_id = f"REQ.{req.index_no:02d}"
try:
sig = self.build_function_signature(req, knowledge)
results.append(sig)
if on_progress:
on_progress(i + 1, total, req, sig, None)
sig = self.build_function_signature(req, req_id, knowledge)
error = None
except Exception as e:
# 降级：用基础信息填充，保证 JSON 完整性
fallback = self._build_fallback_signature(req)
results.append(fallback)
if on_progress:
on_progress(i + 1, total, req, fallback, e)
sig = self._fallback_signature(req, req_id)
error = e
return results
signatures.append(sig)
if on_progress:
on_progress(i, total, req, sig, error)
# ══════════════════════════════════════════════════
# 私有工具方法
# ══════════════════════════════════════════════════
return signatures
@staticmethod
def _build_knowledge_section(knowledge: str) -> str:
"""构建知识库 Prompt 段落"""
if not knowledge or not knowledge.strip():
return ""
return f"""## 参考知识库
{knowledge}
---
"""
@staticmethod
def _sanitize_function_name(name: str) -> str:
"""
清理函数名，确保符合 snake_case 规范
Args:
name: 原始函数名
Returns:
合法的 snake_case 函数名
"""
name = re.sub(r"[^a-zA-Z0-9_]", "_", name).lower()
name = re.sub(r"_+", "_", name).strip("_")
if name and name[0].isdigit():
name = "func_" + name
return name or "unnamed_function"
@staticmethod
def _format_requirement_id(index_no: int) -> str:
"""
将序号格式化为需求编号字符串
Args:
index_no: 功能需求序号（从 1 开始）
Returns:
格式化编号，如 'REQ.01'、'REQ.12'
"""
return f"REQ.{index_no:02d}"
@staticmethod
def _build_fallback_signature(func_req: FunctionalRequirement) -> dict:
"""
构建降级签名（LLM 调用失败时使用）
Args:
func_req: 功能需求对象
Returns:
包含基础信息的签名 dict
"""
def _fallback_signature(
req: FunctionalRequirement,
requirement_id: str,
) -> dict:
"""生成降级签名结构（LLM 失败时使用）"""
return {
"name": func_req.function_name,
"requirement_id": f"REQ.{func_req.index_no:02d}",
"description": func_req.description,
"name": req.function_name,
"requirement_id": requirement_id,
"description": req.description,
"type": "function",
"module": req.module or config.DEFAULT_MODULE,
"parameters": {},
"return": {
"type": "void",
"description": "TODO: define return value"
"return": {
"type": "any",
"on_success": {"value": "...", "description": "成功时返回值"},
"on_failure": {"value": "None", "description": "失败时返回 None"},
},
"_note": "Auto-generated fallback due to LLM error"
}

378
requirements_generator/database/db_manager.py
View File

@ -1,314 +1,152 @@
# database/db_manager.py - 数据库操作管理器
import sqlite3
# database/db_manager.py - 数据库 CRUD 操作封装
import os
from datetime import datetime
from typing import List, Optional
from contextlib import contextmanager
from database.models import (
CREATE_TABLES_SQL, Project, RawRequirement,
FunctionalRequirement, CodeFile
)
from sqlalchemy import create_engine
from sqlalchemy.orm import sessionmaker, Session
import config
from database.models import Base, Project, RawRequirement, FunctionalRequirement, CodeFile
class DBManager:
"""SQLite 数据库管理器，封装所有 CRUD 操作"""
def __init__(self, db_path: str = config.DB_PATH):
self.db_path = db_path
def __init__(self, db_path: str = None):
db_path = db_path or config.DB_PATH
os.makedirs(os.path.dirname(db_path), exist_ok=True)
self._init_db()
self.engine = create_engine(f"sqlite:///{db_path}", echo=False)
Base.metadata.create_all(self.engine)
self._Session = sessionmaker(bind=self.engine)
# ── 连接上下文管理器 ──────────────────────────────────
@contextmanager
def _get_conn(self):
"""获取数据库连接（自动提交/回滚）"""
conn = sqlite3.connect(self.db_path)
conn.row_factory = sqlite3.Row
conn.execute("PRAGMA foreign_keys = ON")
try:
yield conn
conn.commit()
except Exception:
conn.rollback()
raise
finally:
conn.close()
def _init_db(self):
"""初始化数据库，创建所有表"""
with self._get_conn() as conn:
conn.executescript(CREATE_TABLES_SQL)
def _session(self) -> Session:
return self._Session()
# ══════════════════════════════════════════════════
# Project CRUD
# Project
# ══════════════════════════════════════════════════
def create_project(self, project: Project) -> int:
"""创建项目，返回新项目 ID"""
sql = """
INSERT INTO projects (name, description, language, output_dir, created_at, updated_at)
VALUES (?, ?, ?, ?, ?, ?)
"""
with self._get_conn() as conn:
cur = conn.execute(sql, (
project.name, project.description, project.language,
project.output_dir, project.created_at, project.updated_at
))
return cur.lastrowid
def get_project_by_id(self, project_id: int) -> Optional[Project]:
"""根据 ID 查询项目"""
with self._get_conn() as conn:
row = conn.execute(
"SELECT * FROM projects WHERE id = ?", (project_id,)
).fetchone()
if row is None:
return None
return Project(
id=row["id"], name=row["name"], description=row["description"],
language=row["language"], output_dir=row["output_dir"],
created_at=row["created_at"], updated_at=row["updated_at"]
)
with self._session() as s:
s.add(project)
s.commit()
s.refresh(project)
return project.id
def get_project_by_name(self, name: str) -> Optional[Project]:
"""根据名称查询项目"""
with self._get_conn() as conn:
row = conn.execute(
"SELECT * FROM projects WHERE name = ?", (name,)
).fetchone()
if row is None:
return None
return Project(
id=row["id"], name=row["name"], description=row["description"],
language=row["language"], output_dir=row["output_dir"],
created_at=row["created_at"], updated_at=row["updated_at"]
)
with self._session() as s:
return s.query(Project).filter_by(name=name).first()
def list_projects(self) -> List[Project]:
"""列出所有项目"""
with self._get_conn() as conn:
rows = conn.execute(
"SELECT * FROM projects ORDER BY created_at DESC"
).fetchall()
return [
Project(
id=r["id"], name=r["name"], description=r["description"],
language=r["language"], output_dir=r["output_dir"],
created_at=r["created_at"], updated_at=r["updated_at"]
) for r in rows
]
def get_project_by_id(self, project_id: int) -> Optional[Project]:
with self._session() as s:
return s.get(Project, project_id)
def update_project(self, project: Project) -> None:
"""更新项目信息"""
project.updated_at = datetime.now().isoformat()
sql = """
UPDATE projects
SET name=?, description=?, language=?, output_dir=?, updated_at=?
WHERE id=?
"""
with self._get_conn() as conn:
conn.execute(sql, (
project.name, project.description, project.language,
project.output_dir, project.updated_at, project.id
))
with self._session() as s:
s.merge(project)
s.commit()
def delete_project(self, project_id: int) -> None:
"""删除项目（级联删除所有关联数据）"""
with self._get_conn() as conn:
conn.execute("DELETE FROM projects WHERE id = ?", (project_id,))
def list_projects(self) -> List[Project]:
with self._session() as s:
return s.query(Project).order_by(Project.created_at.desc()).all()
# ══════════════════════════════════════════════════
# RawRequirement CRUD
# RawRequirement
# ══════════════════════════════════════════════════
def create_raw_requirement(self, req: RawRequirement) -> int:
"""创建原始需求，返回新记录 ID"""
sql = """
INSERT INTO raw_requirements
(project_id, content, source_type, source_name, knowledge, created_at)
VALUES (?, ?, ?, ?, ?, ?)
"""
with self._get_conn() as conn:
cur = conn.execute(sql, (
req.project_id, req.content, req.source_type,
req.source_name, req.knowledge, req.created_at
))
return cur.lastrowid
def create_raw_requirement(self, raw_req: RawRequirement) -> int:
with self._session() as s:
s.add(raw_req)
s.commit()
s.refresh(raw_req)
return raw_req.id
def get_raw_requirement(self, req_id: int) -> Optional[RawRequirement]:
"""根据 ID 查询原始需求"""
with self._get_conn() as conn:
row = conn.execute(
"SELECT * FROM raw_requirements WHERE id = ?", (req_id,)
).fetchone()
if row is None:
return None
return RawRequirement(
id=row["id"], project_id=row["project_id"], content=row["content"],
source_type=row["source_type"], source_name=row["source_name"],
knowledge=row["knowledge"], created_at=row["created_at"]
)
def list_raw_requirements_by_project(self, project_id: int) -> List[RawRequirement]:
"""查询项目下所有原始需求"""
with self._get_conn() as conn:
rows = conn.execute(
"SELECT * FROM raw_requirements WHERE project_id = ? ORDER BY created_at",
(project_id,)
).fetchall()
return [
RawRequirement(
id=r["id"], project_id=r["project_id"], content=r["content"],
source_type=r["source_type"], source_name=r["source_name"],
knowledge=r["knowledge"], created_at=r["created_at"]
) for r in rows
]
def get_raw_requirement(self, raw_req_id: int) -> Optional[RawRequirement]:
with self._session() as s:
return s.get(RawRequirement, raw_req_id)
# ══════════════════════════════════════════════════
# FunctionalRequirement CRUD
# FunctionalRequirement
# ══════════════════════════════════════════════════
def create_functional_requirement(self, req: FunctionalRequirement) -> int:
"""创建功能需求，返回新记录 ID"""
sql = """
INSERT INTO functional_requirements
(project_id, raw_req_id, index_no, title, description,
function_name, priority, status, is_custom, created_at, updated_at)
VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
"""
with self._get_conn() as conn:
cur = conn.execute(sql, (
req.project_id, req.raw_req_id, req.index_no,
req.title, req.description, req.function_name,
req.priority, req.status, int(req.is_custom),
req.created_at, req.updated_at
))
return cur.lastrowid
with self._session() as s:
s.add(req)
s.commit()
s.refresh(req)
return req.id
def get_functional_requirement(self, req_id: int) -> Optional[FunctionalRequirement]:
"""根据 ID 查询功能需求"""
with self._get_conn() as conn:
row = conn.execute(
"SELECT * FROM functional_requirements WHERE id = ?", (req_id,)
).fetchone()
if row is None:
return None
return self._row_to_func_req(row)
with self._session() as s:
return s.get(FunctionalRequirement, req_id)
def list_functional_requirements(self, project_id: int) -> List[FunctionalRequirement]:
"""查询项目下所有功能需求（按序号排序）"""
with self._get_conn() as conn:
rows = conn.execute(
"""SELECT * FROM functional_requirements
WHERE project_id = ? ORDER BY index_no""",
(project_id,)
).fetchall()
return [self._row_to_func_req(r) for r in rows]
def update_functional_requirement(self, req: FunctionalRequirement) -> None:
"""更新功能需求"""
req.updated_at = datetime.now().isoformat()
sql = """
UPDATE functional_requirements
SET title=?, description=?, function_name=?, priority=?,
status=?, index_no=?, updated_at=?
WHERE id=?
"""
with self._get_conn() as conn:
conn.execute(sql, (
req.title, req.description, req.function_name,
req.priority, req.status, req.index_no,
req.updated_at, req.id
))
def delete_functional_requirement(self, req_id: int) -> None:
"""删除功能需求"""
with self._get_conn() as conn:
conn.execute(
"DELETE FROM functional_requirements WHERE id = ?", (req_id,)
with self._session() as s:
return (
s.query(FunctionalRequirement)
.filter_by(project_id=project_id)
.order_by(FunctionalRequirement.index_no)
.all()
)
def _row_to_func_req(self, row) -> FunctionalRequirement:
"""sqlite Row → FunctionalRequirement 对象"""
return FunctionalRequirement(
id=row["id"], project_id=row["project_id"],
raw_req_id=row["raw_req_id"], index_no=row["index_no"],
title=row["title"], description=row["description"],
function_name=row["function_name"], priority=row["priority"],
status=row["status"], is_custom=bool(row["is_custom"]),
created_at=row["created_at"], updated_at=row["updated_at"]
)
def update_functional_requirement(self, req: FunctionalRequirement) -> None:
with self._session() as s:
s.merge(req)
s.commit()
# ══════════════════════════════════════════════════
# CodeFile CRUD
# ══════════════════════════════════════════════════
def delete_functional_requirement(self, req_id: int) -> None:
with self._session() as s:
obj = s.get(FunctionalRequirement, req_id)
if obj:
s.delete(obj)
s.commit()
def create_code_file(self, code_file: CodeFile) -> int:
"""创建代码文件记录，返回新记录 ID"""
sql = """
INSERT INTO code_files
(project_id, func_req_id, file_name, file_path,
language, content, created_at, updated_at)
VALUES (?, ?, ?, ?, ?, ?, ?, ?)
def bulk_update_modules(self, updates: List[dict]) -> None:
"""
with self._get_conn() as conn:
cur = conn.execute(sql, (
code_file.project_id, code_file.func_req_id,
code_file.file_name, code_file.file_path,
code_file.language, code_file.content,
code_file.created_at, code_file.updated_at
))
return cur.lastrowid
批量更新功能需求的 module 字段。
Args:
updates: [{"function_name": "...", "module": "..."}, ...]
"""
with self._session() as s:
name_to_module = {u["function_name"]: u["module"] for u in updates}
reqs = s.query(FunctionalRequirement).filter(
FunctionalRequirement.function_name.in_(name_to_module.keys())
).all()
for req in reqs:
req.module = name_to_module.get(req.function_name, config.DEFAULT_MODULE)
s.commit()
# ══════════════════════════════════════════════════
# CodeFile
# ══════════════════════════════════════════════════
def upsert_code_file(self, code_file: CodeFile) -> int:
"""插入或更新代码文件（按 func_req_id 唯一键）"""
existing = self.get_code_file_by_func_req(code_file.func_req_id)
if existing:
code_file.id = existing.id
code_file.updated_at = datetime.now().isoformat()
sql = """
UPDATE code_files
SET file_name=?, file_path=?, language=?, content=?, updated_at=?
WHERE id=?
"""
with self._get_conn() as conn:
conn.execute(sql, (
code_file.file_name, code_file.file_path,
code_file.language, code_file.content,
code_file.updated_at, code_file.id
))
return code_file.id
else:
return self.create_code_file(code_file)
with self._session() as s:
existing = (
s.query(CodeFile)
.filter_by(func_req_id=code_file.func_req_id)
.first()
)
if existing:
existing.file_name = code_file.file_name
existing.file_path = code_file.file_path
existing.module = code_file.module
existing.language = code_file.language
existing.content = code_file.content
s.commit()
return existing.id
else:
s.add(code_file)
s.commit()
s.refresh(code_file)
return code_file.id
def get_code_file_by_func_req(self, func_req_id: int) -> Optional[CodeFile]:
"""根据功能需求 ID 查询代码文件"""
with self._get_conn() as conn:
row = conn.execute(
"SELECT * FROM code_files WHERE func_req_id = ?", (func_req_id,)
).fetchone()
if row is None:
return None
return self._row_to_code_file(row)
def list_code_files_by_project(self, project_id: int) -> List[CodeFile]:
"""查询项目下所有代码文件"""
with self._get_conn() as conn:
rows = conn.execute(
"SELECT * FROM code_files WHERE project_id = ? ORDER BY created_at",
(project_id,)
).fetchall()
return [self._row_to_code_file(r) for r in rows]
def _row_to_code_file(self, row) -> CodeFile:
"""sqlite Row → CodeFile 对象"""
return CodeFile(
id=row["id"], project_id=row["project_id"],
func_req_id=row["func_req_id"], file_name=row["file_name"],
file_path=row["file_path"], language=row["language"],
content=row["content"], created_at=row["created_at"],
updated_at=row["updated_at"]
)
def list_code_files(self, project_id: int) -> List[CodeFile]:
with self._session() as s:
return (
s.query(CodeFile)
.join(FunctionalRequirement)
.filter(FunctionalRequirement.project_id == project_id)
.all()
)

201
requirements_generator/database/models.py
View File

@ -1,122 +1,95 @@
# database/models.py - 数据模型定义（SQLite 建表 DDL）
from dataclasses import dataclass, field
# database/models.py - SQLAlchemy ORM 数据模型
from datetime import datetime
from typing import Optional
from sqlalchemy import (
Column, Integer, String, Text, Boolean,
DateTime, ForeignKey, Enum,
)
from sqlalchemy.orm import declarative_base, relationship
# ══════════════════════════════════════════════════════
# DDL 建表语句
# ══════════════════════════════════════════════════════
CREATE_TABLES_SQL = """
PRAGMA foreign_keys = ON;
-- 项目表
CREATE TABLE IF NOT EXISTS projects (
id INTEGER PRIMARY KEY AUTOINCREMENT,
name TEXT NOT NULL UNIQUE, -- 项目名称
description TEXT, -- 项目描述
language TEXT NOT NULL DEFAULT 'python', -- 目标代码语言
output_dir TEXT NOT NULL, -- 输出目录路径
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL
);
-- 原始需求表
CREATE TABLE IF NOT EXISTS raw_requirements (
id INTEGER PRIMARY KEY AUTOINCREMENT,
project_id INTEGER NOT NULL, -- 关联项目
content TEXT NOT NULL, -- 需求原文
source_type TEXT NOT NULL DEFAULT 'text', -- text | file
source_name TEXT, -- 文件名（文件输入时）
knowledge TEXT, -- 合并后的知识库内容
created_at TEXT NOT NULL,
FOREIGN KEY (project_id) REFERENCES projects(id) ON DELETE CASCADE
);
-- 功能需求表
CREATE TABLE IF NOT EXISTS functional_requirements (
id INTEGER PRIMARY KEY AUTOINCREMENT,
project_id INTEGER NOT NULL,
raw_req_id INTEGER NOT NULL, -- 关联原始需求
index_no INTEGER NOT NULL, -- 序号
title TEXT NOT NULL, -- 功能标题
description TEXT NOT NULL, -- 功能描述
function_name TEXT NOT NULL, -- 对应函数名
priority TEXT NOT NULL DEFAULT 'medium', -- high|medium|low
status TEXT NOT NULL DEFAULT 'pending', -- pending|generated|skipped
is_custom INTEGER NOT NULL DEFAULT 0, -- 是否用户自定义 (0/1)
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL,
FOREIGN KEY (project_id) REFERENCES projects(id) ON DELETE CASCADE,
FOREIGN KEY (raw_req_id) REFERENCES raw_requirements(id) ON DELETE CASCADE
);
-- 代码文件表
CREATE TABLE IF NOT EXISTS code_files (
id INTEGER PRIMARY KEY AUTOINCREMENT,
project_id INTEGER NOT NULL,
func_req_id INTEGER NOT NULL UNIQUE, -- 关联功能需求（1对1）
file_name TEXT NOT NULL, -- 文件名
file_path TEXT NOT NULL, -- 完整路径
language TEXT NOT NULL, -- 代码语言
content TEXT NOT NULL, -- 代码内容
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL,
FOREIGN KEY (project_id) REFERENCES projects(id) ON DELETE CASCADE,
FOREIGN KEY (func_req_id) REFERENCES functional_requirements(id) ON DELETE CASCADE
);
"""
# ══════════════════════════════════════════════════════
# 数据类（Python 对象映射）
# ══════════════════════════════════════════════════════
@dataclass
class Project:
name: str
output_dir: str
language: str = "python"
description: str = ""
id: Optional[int] = None
created_at: str = field(default_factory=lambda: datetime.now().isoformat())
updated_at: str = field(default_factory=lambda: datetime.now().isoformat())
Base = declarative_base()
@dataclass
class RawRequirement:
project_id: int
content: str
source_type: str = "text" # text | file
source_name: Optional[str] = None
knowledge: Optional[str] = None
id: Optional[int] = None
created_at: str = field(default_factory=lambda: datetime.now().isoformat())
class Project(Base):
"""项目表"""
__tablename__ = "projects"
id = Column(Integer, primary_key=True, autoincrement=True)
name = Column(String(200), nullable=False, unique=True)
language = Column(String(50), nullable=False, default="python")
description = Column(Text, nullable=True)
output_dir = Column(String(500), nullable=True)
created_at = Column(DateTime, default=datetime.utcnow)
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
raw_requirements = relationship("RawRequirement", back_populates="project", cascade="all, delete-orphan")
functional_requirements = relationship("FunctionalRequirement", back_populates="project", cascade="all, delete-orphan")
def __repr__(self):
return f"<Project(id={self.id}, name={self.name!r}, language={self.language!r})>"
@dataclass
class FunctionalRequirement:
project_id: int
raw_req_id: int
index_no: int
title: str
description: str
function_name: str
priority: str = "medium"
status: str = "pending"
is_custom: bool = False
id: Optional[int] = None
created_at: str = field(default_factory=lambda: datetime.now().isoformat())
updated_at: str = field(default_factory=lambda: datetime.now().isoformat())
class RawRequirement(Base):
"""原始需求表"""
__tablename__ = "raw_requirements"
id = Column(Integer, primary_key=True, autoincrement=True)
project_id = Column(Integer, ForeignKey("projects.id"), nullable=False)
content = Column(Text, nullable=False)
source_type = Column(String(20), nullable=False, default="text") # text / file
source_name = Column(String(200), nullable=True)
knowledge = Column(Text, nullable=True)
created_at = Column(DateTime, default=datetime.utcnow)
project = relationship("Project", back_populates="raw_requirements")
functional_requirements = relationship("FunctionalRequirement", back_populates="raw_requirement", cascade="all, delete-orphan")
def __repr__(self):
return f"<RawRequirement(id={self.id}, project_id={self.project_id})>"
@dataclass
class CodeFile:
project_id: int
func_req_id: int
file_name: str
file_path: str
language: str
content: str
id: Optional[int] = None
created_at: str = field(default_factory=lambda: datetime.now().isoformat())
updated_at: str = field(default_factory=lambda: datetime.now().isoformat())
class FunctionalRequirement(Base):
"""功能需求表"""
__tablename__ = "functional_requirements"
id = Column(Integer, primary_key=True, autoincrement=True)
project_id = Column(Integer, ForeignKey("projects.id"), nullable=False)
raw_req_id = Column(Integer, ForeignKey("raw_requirements.id"), nullable=False)
index_no = Column(Integer, nullable=False)
title = Column(String(200), nullable=False)
description = Column(Text, nullable=False)
function_name = Column(String(200), nullable=False)
priority = Column(Enum("high", "medium", "low"), default="medium")
module = Column(String(100), nullable=True, default="default") # 功能模块
status = Column(String(50), nullable=False, default="pending") # pending / generated / failed
is_custom = Column(Boolean, nullable=False, default=False)
created_at = Column(DateTime, default=datetime.utcnow)
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
project = relationship("Project", back_populates="functional_requirements")
raw_requirement = relationship("RawRequirement", back_populates="functional_requirements")
code_files = relationship("CodeFile", back_populates="functional_requirement", cascade="all, delete-orphan")
def __repr__(self):
return (
f"<FunctionalRequirement(id={self.id}, title={self.title!r}, "
f"module={self.module!r}, status={self.status!r})>"
)
class CodeFile(Base):
"""生成代码文件表"""
__tablename__ = "code_files"
id = Column(Integer, primary_key=True, autoincrement=True)
func_req_id = Column(Integer, ForeignKey("functional_requirements.id"), nullable=False)
file_name = Column(String(200), nullable=False)
file_path = Column(String(500), nullable=False)
module = Column(String(100), nullable=True) # 冗余存储，方便查询
language = Column(String(50), nullable=False)
content = Column(Text, nullable=True)
generated_at = Column(DateTime, default=datetime.utcnow)
functional_requirement = relationship("FunctionalRequirement", back_populates="code_files")
def __repr__(self):
return f"<CodeFile(id={self.id}, file_name={self.file_name!r}, module={self.module!r})>"

697
requirements_generator/main.py
View File

File diff suppressed because it is too large Load Diff

11
requirements_generator/requirements.txt
View File

@ -1,6 +1,7 @@
openai>=1.0.0
python-dotenv>=1.0.0
openai>=1.30.0
click>=8.1.0
rich>=13.0.0
python-docx>=0.8.11
PyPDF2>=3.0.0
click>=8.1.0
sqlalchemy>=2.0.0
python-dotenv>=1.0.0
pypdf>=4.0.0
python-docx>=1.1.0

11
requirements_generator/run.bat Normal file
View File

@ -0,0 +1,11 @@
#!/usr/bin/env bash
# 安装依赖
pip install -r requirements.txt
# 设置环境变量
set OPENAI_API_KEY="sk-AUmOuFI731Ty5Nob38jY26d8lydfDT-QkE2giqb0sCuPCAE2JH6zjLM4lZLpvL5WMYPOocaMe2FwVDmqM_9KimmKACjR"
set OPENAI_BASE_URL="https://openapi.monica.im/v1" # 或其他兼容接口
set LLM_MODEL="gpt-4o"
python main.py

0
requirements_generator/run.sh Executable file → Normal file
View File

173
requirements_generator/utils/file_handler.py
View File

@ -1,95 +1,13 @@
# utils/file_handler.py - 文件读取工具（支持 txt/md/pdf/docx）
# utils/file_handler.py - 文件读取工具（支持 txt / md / pdf / docx）
import os
from typing import List, Optional
from pathlib import Path
def read_text_file(file_path: str) -> str:
"""
读取纯文本文件内容（.txt / .md / .py 等）
Args:
file_path: 文件路径
Returns:
文件文本内容
Raises:
FileNotFoundError: 文件不存在
UnicodeDecodeError: 编码错误时尝试 latin-1 兜底
"""
path = Path(file_path)
if not path.exists():
raise FileNotFoundError(f"文件不存在: {file_path}")
try:
return path.read_text(encoding="utf-8")
except UnicodeDecodeError:
return path.read_text(encoding="latin-1")
def read_pdf_file(file_path: str) -> str:
"""
读取 PDF 文件内容
Args:
file_path: PDF 文件路径
Returns:
提取的文本内容
Raises:
ImportError: 未安装 PyPDF2
FileNotFoundError: 文件不存在
"""
try:
import PyPDF2
except ImportError:
raise ImportError("请安装 PyPDF2: pip install PyPDF2")
path = Path(file_path)
if not path.exists():
raise FileNotFoundError(f"文件不存在: {file_path}")
texts = []
with open(file_path, "rb") as f:
reader = PyPDF2.PdfReader(f)
for page in reader.pages:
text = page.extract_text()
if text:
texts.append(text)
return "\n".join(texts)
def read_docx_file(file_path: str) -> str:
"""
读取 Word (.docx) 文件内容
Args:
file_path: docx 文件路径
Returns:
提取的文本内容（段落合并）
Raises:
ImportError: 未安装 python-docx
FileNotFoundError: 文件不存在
"""
try:
from docx import Document
except ImportError:
raise ImportError("请安装 python-docx: pip install python-docx")
path = Path(file_path)
if not path.exists():
raise FileNotFoundError(f"文件不存在: {file_path}")
doc = Document(file_path)
return "\n".join(para.text for para in doc.paragraphs if para.text.strip())
from typing import List
def read_file_auto(file_path: str) -> str:
"""
根据文件扩展名自动选择读取方式
自动识别文件类型并读取文本内容。
支持格式：.txt / .md / .pdf / .docx / 其他（按 UTF-8 读取）
Args:
file_path: 文件路径
@ -98,45 +16,66 @@ def read_file_auto(file_path: str) -> str:
文件文本内容
Raises:
ValueError: 不支持的文件类型
FileNotFoundError: 文件不存在
RuntimeError: 读取失败
"""
ext = Path(file_path).suffix.lower()
readers = {
".txt": read_text_file,
".md": read_text_file,
".py": read_text_file,
".json": read_text_file,
".yaml": read_text_file,
".yml": read_text_file,
".pdf": read_pdf_file,
".docx": read_docx_file,
}
reader = readers.get(ext)
if reader is None:
raise ValueError(f"不支持的文件类型: {ext}，支持: {list(readers.keys())}")
return reader(file_path)
if not os.path.exists(file_path):
raise FileNotFoundError(f"文件不存在: {file_path}")
ext = os.path.splitext(file_path)[1].lower()
try:
if ext == ".pdf":
return _read_pdf(file_path)
elif ext in (".docx", ".doc"):
return _read_docx(file_path)
else:
# txt / md / 其他文本格式
with open(file_path, "r", encoding="utf-8", errors="replace") as f:
return f.read()
except Exception as e:
raise RuntimeError(f"读取文件失败 [{file_path}]: {e}")
def _read_pdf(file_path: str) -> str:
"""读取 PDF 文件文本"""
try:
from pypdf import PdfReader
except ImportError:
raise RuntimeError("读取 PDF 需要安装 pypdf：pip install pypdf")
reader = PdfReader(file_path)
pages = [page.extract_text() or "" for page in reader.pages]
return "\n".join(pages)
def _read_docx(file_path: str) -> str:
"""读取 Word 文档文本"""
try:
from docx import Document
except ImportError:
raise RuntimeError("读取 docx 需要安装 python-docx：pip install python-docx")
doc = Document(file_path)
paragraphs = [p.text for p in doc.paragraphs if p.text.strip()]
return "\n".join(paragraphs)
def merge_knowledge_files(file_paths: List[str]) -> str:
"""
合并多个知识库文件为单一文本
合并多个知识库文件为单一文本。
Args:
file_paths: 知识库文件路径列表
file_paths: 文件路径列表
Returns:
合并后的知识库文本（包含文件名分隔符）
合并后的文本（各文件以分隔线隔开）
"""
if not file_paths:
return ""
sections = []
for fp in file_paths:
parts = []
for path in file_paths:
try:
content = read_file_auto(fp)
file_name = Path(fp).name
sections.append(f"### 知识库文件: {file_name}\n{content}")
content = read_file_auto(path)
parts.append(f"--- {os.path.basename(path)} ---\n{content}")
except Exception as e:
sections.append(f"### 知识库文件: {fp}\n[读取失败: {e}]")
return "\n\n".join(sections)
parts.append(f"--- {os.path.basename(path)} [读取失败: {e}] ---")
return "\n\n".join(parts)

371
requirements_generator/utils/output_writer.py
View File

@ -6,133 +6,74 @@ from typing import Dict, List
import config
# 各语言文件扩展名映射
LANGUAGE_EXT_MAP: Dict[str, str] = {
"python": ".py",
"javascript": ".js",
"typescript": ".ts",
"java": ".java",
"go": ".go",
"rust": ".rs",
"cpp": ".cpp",
"c": ".c",
"csharp": ".cs",
"ruby": ".rb",
"php": ".php",
"swift": ".swift",
"kotlin": ".kt",
}
# 合法的通用类型集合
VALID_TYPES = {
"integer", "string", "boolean", "float",
"list", "dict", "object", "void", "any",
}
# 合法的 inout 值
VALID_INOUT = {"in", "out", "inout"}
def get_file_extension(language: str) -> str:
"""
获取指定语言的文件扩展名
Args:
language: 编程语言名称（小写）
Returns:
文件扩展名（含点号，如 '.py'）
"""
return LANGUAGE_EXT_MAP.get(language.lower(), ".txt")
# ══════════════════════════════════════════════════════
# 目录管理
# ══════════════════════════════════════════════════════
def build_project_output_dir(project_name: str) -> str:
"""
构建项目输出目录路径
Args:
project_name: 项目名称
Returns:
输出目录路径
"""
safe_name = "".join(c if c.isalnum() or c in "-_" else "_" for c in project_name)
return os.path.join(config.OUTPUT_BASE_DIR, safe_name)
safe = "".join(c if c.isalnum() or c in "-_" else "_" for c in project_name)
return os.path.join(config.OUTPUT_BASE_DIR, safe)
def ensure_project_dir(project_name: str) -> str:
"""
确保项目输出目录存在，不存在则创建
Args:
project_name: 项目名称
Returns:
创建好的目录路径
"""
"""确保项目根输出目录存在，并创建 __init__.py"""
output_dir = build_project_output_dir(project_name)
os.makedirs(output_dir, exist_ok=True)
init_file = os.path.join(output_dir, "__init__.py")
if not os.path.exists(init_file):
Path(init_file).write_text(
"# Auto-generated project package\n", encoding="utf-8"
)
Path(init_file).write_text("# Auto-generated project package\n", encoding="utf-8")
return output_dir
def write_code_file(
output_dir: str,
function_name: str,
language: str,
content: str,
) -> str:
"""
将代码内容写入指定目录的文件
def ensure_module_dir(output_dir: str, module: str) -> str:
"""确保模块子目录存在，并创建 __init__.py"""
module_dir = os.path.join(output_dir, module)
os.makedirs(module_dir, exist_ok=True)
init_file = os.path.join(module_dir, "__init__.py")
if not os.path.exists(init_file):
Path(init_file).write_text(
f"# Auto-generated module package: {module}\n", encoding="utf-8"
)
return module_dir
Args:
output_dir: 输出目录路径
function_name: 函数名（用于生成文件名）
language: 编程语言
content: 代码内容
Returns:
写入的文件完整路径
"""
ext = get_file_extension(language)
file_name = f"{function_name}{ext}"
file_path = os.path.join(output_dir, file_name)
Path(file_path).write_text(content, encoding="utf-8")
return file_path
# ══════════════════════════════════════════════════════
# README
# ══════════════════════════════════════════════════════
def write_project_readme(
output_dir: str,
project_name: str,
output_dir: str,
project_name: str,
project_description: str,
requirements_summary: str,
modules: List[str] = None,
) -> str:
"""
在项目目录生成 README.md 文件
"""生成项目 README.md"""
module_section = ""
if modules:
module_list = "\n".join(f"- `{m}/`" for m in sorted(set(modules)))
module_section = f"\n## 功能模块\n\n{module_list}\n"
Args:
output_dir: 项目输出目录
project_name: 项目名称
requirements_summary: 功能需求摘要文本
Returns:
README.md 文件路径
"""
readme_content = f"""# {project_name}
content = f"""# {project_name}
> Auto-generated by Requirement Analyzer
{project_description or ""}
{module_section}
## 功能需求列表
{requirements_summary}
"""
readme_path = os.path.join(output_dir, "README.md")
Path(readme_path).write_text(readme_content, encoding="utf-8")
return readme_path
path = os.path.join(output_dir, "README.md")
Path(path).write_text(content, encoding="utf-8")
return path
# ══════════════════════════════════════════════════════
@ -140,26 +81,18 @@ def write_project_readme(
# ══════════════════════════════════════════════════════
def build_signatures_document(
project_name: str,
project_name: str,
project_description: str,
signatures: List[dict],
signatures: List[dict],
) -> dict:
"""
将函数签名列表包装为带项目信息的顶层文档结构。
构建顶层签名文档结构::
Args:
project_name: 项目名称，写入 "project" 字段
project_description: 项目描述，写入 "description" 字段
signatures: 函数签名 dict 列表，写入 "functions" 字段
Returns:
顶层文档 dict，结构为::
{
"project": "<project_name>",
"description": "<project_description>",
"functions": [ ... ]
}
{
"project": "<name>",
"description": "<description>",
"functions": [ ... ]
}
"""
return {
"project": project_name,
@ -169,118 +102,45 @@ def build_signatures_document(
def patch_signatures_with_url(
signatures: List[dict],
signatures: List[dict],
func_name_to_url: Dict[str, str],
) -> List[dict]:
"""
将代码文件的路径（URL）回写到对应函数签名的 "url" 字段。
遍历签名列表，根据 signature["name"] 在 func_name_to_url 中查找
对应路径，找到则写入 "url" 字段；未找到则写入空字符串，不抛出异常。
"url" 字段插入位置紧跟在 "type" 字段之后，以保持字段顺序的可读性::
{
"name": "create_user",
"requirement_id": "REQ.01",
"description": "...",
"type": "function",
"url": "/abs/path/to/create_user.py", ← 新增
"parameters": { ... },
"return": { ... }
}
将代码文件路径回写到签名的 "url" 字段（紧跟 "type" 之后）。
Args:
signatures: 原始签名列表（in-place 修改）
func_name_to_url: {函数名: 代码文件绝对路径} 映射表，
由 CodeGenerator.generate_batch() 的进度回调收集
signatures: 签名列表（in-place 修改）
func_name_to_url: {函数名: 文件绝对路径}
Returns:
修改后的签名列表（与传入的同一对象，方便链式调用）
修改后的签名列表
"""
for sig in signatures:
func_name = sig.get("name", "")
url = func_name_to_url.get(func_name, "")
url = func_name_to_url.get(sig.get("name", ""), "")
_insert_field_after(sig, after_key="type", new_key="url", new_value=url)
return signatures
def _insert_field_after(
d: dict,
after_key: str,
new_key: str,
new_value,
) -> None:
"""
在有序 dict 中将 new_key 插入到 after_key 之后。
若 after_key 不存在，则追加到末尾。
若 new_key 已存在，则直接更新其值（不改变位置）。
Args:
d: 目标 dict（Python 3.7+ 保证插入顺序）
after_key: 参考键名
new_key: 要插入的键名
new_value: 要插入的值
"""
def _insert_field_after(d: dict, after_key: str, new_key: str, new_value) -> None:
"""在有序 dict 中将 new_key 插入到 after_key 之后"""
if new_key in d:
d[new_key] = new_value
return
items = list(d.items())
insert_pos = len(items)
for i, (k, _) in enumerate(items):
if k == after_key:
insert_pos = i + 1
break
insert_pos = next((i + 1 for i, (k, _) in enumerate(items) if k == after_key), len(items))
items.insert(insert_pos, (new_key, new_value))
d.clear()
d.update(items)
def write_function_signatures_json(
output_dir: str,
signatures: List[dict],
project_name: str,
output_dir: str,
signatures: List[dict],
project_name: str,
project_description: str,
file_name: str = "function_signatures.json",
file_name: str = "function_signatures.json",
) -> str:
"""
将函数签名列表连同项目信息一起导出为 JSON 文件。
输出的 JSON 顶层结构为::
{
"project": "<project_name>",
"description": "<project_description>",
"functions": [
{
"name": "...",
"requirement_id": "...",
"description": "...",
"type": "function",
"url": "/abs/path/to/xxx.py",
"parameters": { ... },
"return": { ... }
},
...
]
}
Args:
output_dir: JSON 文件写入目录
signatures: 函数签名 dict 列表（应已通过
patch_signatures_with_url() 写入 "url" 字段）
project_name: 项目名称
project_description: 项目描述
file_name: 输出文件名，默认 function_signatures.json
Returns:
写入的 JSON 文件完整路径
Raises:
OSError: 目录不可写
"""
"""将签名列表导出为 JSON 文件"""
os.makedirs(output_dir, exist_ok=True)
document = build_signatures_document(project_name, project_description, signatures)
file_path = os.path.join(output_dir, file_name)
@ -294,137 +154,70 @@ def write_function_signatures_json(
# ══════════════════════════════════════════════════════
def validate_signature_schema(signature: dict) -> List[str]:
"""
校验单个函数签名 dict 是否符合规范。
校验范围：
- 顶层必填字段：name / requirement_id / description / type / parameters
- 可选字段 "url"：若存在则必须为非空字符串
- parameters：每个参数的 type / inout / required 字段
- return：type 字段 + on_success / on_failure 子结构
- void 函数：on_success / on_failure 应为 null
- 非 void 函数：on_success / on_failure 必须存在，
且 value（非空）与 description（非空）均需填写
Args:
signature: 单个函数签名 dict
Returns:
错误信息字符串列表，列表为空表示校验通过
"""
"""校验单个函数签名结构，返回错误列表（空列表表示通过）"""
errors: List[str] = []
# ── 顶层必填字段 ──────────────────────────────────
for key in ("name", "requirement_id", "description", "type", "parameters"):
if key not in signature:
errors.append(f"缺少顶层字段: '{key}'")
# ── url 字段（可选，存在时校验非空）─────────────────
if "url" in signature:
if not isinstance(signature["url"], str):
errors.append("'url' 字段必须是字符串类型")
elif signature["url"] == "":
errors.append("'url' 字段不能为空字符串（代码文件路径未成功回写）")
errors.append("'url' 字段不能为空字符串")
# ── parameters ────────────────────────────────────
params = signature.get("parameters", {})
if not isinstance(params, dict):
errors.append("'parameters' 必须是 dict 类型")
else:
if isinstance(params, dict):
for pname, pdef in params.items():
if not isinstance(pdef, dict):
errors.append(f"参数 '{pname}' 定义必须是 dict")
continue
# type（支持联合类型，如 "string|integer"）
if "type" not in pdef:
errors.append(f"参数 '{pname}' 缺少 'type' 字段")
errors.append(f"参数 '{pname}' 缺少 'type'")
else:
parts = [p.strip() for p in pdef["type"].split("|")]
if not all(p in VALID_TYPES for p in parts):
errors.append(
f"参数 '{pname}' 的 type='{pdef['type']}' 含有不合法的类型"
)
# inout
errors.append(f"参数 '{pname}' type='{pdef['type']}' 含不合法类型")
if "inout" not in pdef:
errors.append(f"参数 '{pname}' 缺少 'inout' 字段")
errors.append(f"参数 '{pname}' 缺少 'inout'")
elif pdef["inout"] not in VALID_INOUT:
errors.append(
f"参数 '{pname}' 的 inout='{pdef['inout']}' 应为 in/out/inout"
)
# required
errors.append(f"参数 '{pname}' inout='{pdef['inout']}' 应为 in/out/inout")
if "required" not in pdef:
errors.append(f"参数 '{pname}' 缺少 'required' 字段")
errors.append(f"参数 '{pname}' 缺少 'required'")
elif not isinstance(pdef["required"], bool):
errors.append(
f"参数 '{pname}' 的 'required' 应为布尔值 true/false，"
f"当前为: {pdef['required']!r}"
)
errors.append(f"参数 '{pname}' 'required' 应为布尔值")
# ── return ────────────────────────────────────────
ret = signature.get("return")
if ret is None:
errors.append(
"缺少 'return' 字段（void 函数请填 "
"{\"type\": \"void\", \"on_success\": null, \"on_failure\": null}）"
)
elif not isinstance(ret, dict):
errors.append("'return' 必须是 dict 类型")
else:
errors.append("缺少 'return' 字段")
elif isinstance(ret, dict):
ret_type = ret.get("type")
if not ret_type:
errors.append("'return' 缺少 'type' 字段")
errors.append("'return' 缺少 'type'")
elif ret_type not in VALID_TYPES:
errors.append(f"'return.type'='{ret_type}' 不在合法类型列表中")
errors.append(f"'return.type'='{ret_type}' 不合法")
is_void = (ret_type == "void")
for sub_key in ("on_success", "on_failure"):
sub = ret.get(sub_key)
if is_void:
if sub is not None:
errors.append(
f"void 函数的 'return.{sub_key}' 应为 null，"
f"当前为: {sub!r}"
)
errors.append(f"void 函数 'return.{sub_key}' 应为 null")
else:
if sub is None:
errors.append(
f"非 void 函数缺少 'return.{sub_key}'，"
f"请描述{'成功' if sub_key == 'on_success' else '失败'}时的返回值"
)
elif not isinstance(sub, dict):
errors.append(f"'return.{sub_key}' 必须是 dict 类型")
else:
if "value" not in sub:
errors.append(f"'return.{sub_key}' 缺少 'value' 字段")
elif sub["value"] == "":
errors.append(
f"'return.{sub_key}.value' 不能为空字符串，"
f"请填写具体返回值、值域描述或结构示例"
)
if "description" not in sub or sub.get("description") in (None, ""):
errors.append(f"非 void 函数缺少 'return.{sub_key}'")
elif isinstance(sub, dict):
if not sub.get("value"):
errors.append(f"'return.{sub_key}.value' 不能为空")
if not sub.get("description"):
errors.append(f"'return.{sub_key}.description' 不能为空")
return errors
def validate_all_signatures(signatures: List[dict]) -> Dict[str, List[str]]:
"""
批量校验函数签名列表。
注意：此函数接受的是纯签名列表（即顶层文档的 "functions" 字段），
而非包含 project/description 的顶层文档。
Args:
signatures: 函数签名 dict 列表
Returns:
{函数名: [错误信息, ...]} 字典，仅包含有错误的条目
"""
report: Dict[str, List[str]] = {}
for sig in signatures:
name = sig.get("name", f"unknown_{id(sig)}")
errs = validate_signature_schema(sig)
if errs:
report[name] = errs
return report
"""批量校验，返回 {函数名: [错误]} 字典（仅含有错误的条目）"""
return {
sig.get("name", f"unknown_{i}"): errs
for i, sig in enumerate(signatures)
if (errs := validate_signature_schema(sig))
}