staran 0.6.0__py3-none-any.whl → 1.0.0__py3-none-any.whl

staran/schemas/aum/__init__.py

@@ -1,210 +0,0 @@
-"""
-新疆工行代发长尾客户表结构定义模块
-
-包含新疆工行代发长尾客户相关的所有表结构：
-- 代发长尾客户行为特征表 (XinjiangICBCDaifaLongtailBehaviorSchema)  
-- 代发长尾客户资产平均值表 (XinjiangICBCDaifaLongtailAssetAvgSchema)
-- 代发长尾客户资产配置表 (XinjiangICBCDaifaLongtailAssetConfigSchema)
-- 代发长尾客户月度统计表 (XinjiangICBCDaifaLongtailMonthlyStatSchema)
-
-数据库: xinjiang_icbc_daifa_longtail
-业务范围: 代发长尾客户（资产10k-100k）
-
-这些表结构可以用于：
-1. 代发长尾客户特征工程
-2. 提升模型和防流失模型构建
-3. 业务文档生成
-4. 数据质量检查
-"""
-
-from typing import Dict
-from ...features.schema import TableSchema
-from ...tools.document_generator import SchemaDocumentGenerator
-
-
-class XinjiangICBCDaifaLongtailBehaviorSchema:
-    """新疆工行代发长尾客户行为特征表 - 严格按照已提供给行方的字段"""
-    
-    # 统计指标定义 - 与原始定义完全一致
-    _STATS = [
-        ("max", "最大值"),
-        ("min", "最小值"),
-        ("sum", "总和"),
-        ("avg", "均值"),
-        ("var", "方差"),
-        ("std", "标准差"),
-        ("rng", "极差"),
-        ("med", "中位数"),
-    ]
-    
-    @staticmethod
-    def create() -> TableSchema:
-        """创建新疆工行代发长尾客户行为特征表结构"""
-        schema = TableSchema('xinjiang_icbc_daifa_hlwj_dfcw_f1_f4_wy')
-        schema.add_primary_key('party_id', 'string')
-        schema.add_date_field('data_dt', 'string')
-        
-        # 基础字段 - 严格按照原始定义
-        schema.add_field("buy_ct", "int", comment="购买次数", aggregatable=True)
-        schema.add_field("recency", "int", comment="最近一次购买距今天数", aggregatable=True)
-        schema.add_field("tenure", "int", comment="客户关系持续时间", aggregatable=True)
-        schema.add_field("window1", "string", comment="时间窗口标记")
-        schema.add_field("freq", "float", comment="总购买频率", aggregatable=True)
-        schema.add_field("freq1", "float", comment="最近时间段购买频率", aggregatable=True)
-        schema.add_field("productidcount", "int", comment="产品种类数", aggregatable=True)
-        schema.add_field("orderidcount", "int", comment="订单数", aggregatable=True)
-        schema.add_field("label", "float", comment="标签值（如是否购买）", aggregatable=True)
-        
-        # productamount and m1 ~ m4 的含义描述
-        m_fields = {
-            "productamount": "购买金额",
-            "m1": "去重订单数",
-            "m2": "去重商品数",
-            "m3": "去重渠道数",
-            "m4": "去重产品品类数",
-        }
-
-        # 使用循环注册 productamount and m1~m4 各统计字段
-        for prefix, meaning in m_fields.items():
-            for stat_key, stat_desc in XinjiangICBCDaifaLongtailBehaviorSchema._STATS:
-                field_name = f"{prefix}_{stat_key}"
-                description = f"{meaning}的{stat_desc}"
-                schema.add_field(field_name, "float", comment=description, aggregatable=True)
-
-        # 其他字段
-        schema.add_field("life_day", "float", comment="客户生命周期天数", aggregatable=True)
-        schema.add_field("gender", "float", comment="性别（编码）", aggregatable=True)
-        schema.add_field("open_day", "float", comment="开户天数", aggregatable=True)
-        
-        schema.set_monthly_unique(False)  # 每人每日记录
-        return schema
-
-
-class XinjiangICBCDaifaLongtailAssetAvgSchema:
-    """新疆工行代发长尾客户资产平均值表"""
-    
-    @staticmethod
-    def create() -> TableSchema:
-        """创建新疆工行代发长尾客户资产平均值表结构"""
-        schema = TableSchema('xinjiang_icbc_daifa_hlwj_zi_chan_avg_wy')
-        schema.add_primary_key('party_id', 'string')
-        schema.add_date_field('data_dt', 'string')
-        
-        # 基础余额字段
-        schema.add_field("asset_total_bal", "decimal", comment="总资产余额", aggregatable=True)
-        schema.add_field("liab_total_bal", "decimal", comment="总负债余额", aggregatable=True)
-        schema.add_field("net_asset_bal", "decimal", comment="净资产余额", aggregatable=True)
-        
-        # 存款相关字段
-        schema.add_field("dep_bal", "decimal", comment="存款余额", aggregatable=True)
-        schema.add_field("current_dep_bal", "decimal", comment="活期存款余额", aggregatable=True)
-        schema.add_field("time_dep_bal", "decimal", comment="定期存款余额", aggregatable=True)
-        
-        # 理财投资字段
-        schema.add_field("wealth_bal", "decimal", comment="理财余额", aggregatable=True)
-        schema.add_field("fund_bal", "decimal", comment="基金余额", aggregatable=True)
-        schema.add_field("insurance_bal", "decimal", comment="保险余额", aggregatable=True)
-        
-        schema.set_monthly_unique(True)  # 每人每月一条记录
-        return schema
-
-
-class XinjiangICBCDaifaLongtailAssetConfigSchema:
-    """新疆工行代发长尾客户资产配置表"""
-    
-    @staticmethod
-    def create() -> TableSchema:
-        """创建新疆工行代发长尾客户资产配置表结构"""
-        schema = TableSchema('xinjiang_icbc_daifa_hlwj_zi_chan_config_wy')
-        schema.add_primary_key('party_id', 'string')
-        schema.add_date_field('data_dt', 'string')
-        
-        # 资产配置比例字段
-        schema.add_field("cash_ratio", "float", comment="现金类资产占比", aggregatable=True)
-        schema.add_field("fixed_income_ratio", "float", comment="固收类资产占比", aggregatable=True)
-        schema.add_field("equity_ratio", "float", comment="权益类资产占比", aggregatable=True)
-        schema.add_field("alternative_ratio", "float", comment="另类资产占比", aggregatable=True)
-        
-        # 风险偏好相关
-        schema.add_field("risk_level", "int", comment="风险偏好等级(1-5)", aggregatable=True)
-        schema.add_field("investment_experience", "int", comment="投资经验年限", aggregatable=True)
-        
-        # 配置变化指标
-        schema.add_field("config_change_freq", "int", comment="配置调整频率", aggregatable=True)
-        schema.add_field("rebalance_count", "int", comment="再平衡次数", aggregatable=True)
-        
-        schema.set_monthly_unique(True)
-        return schema
-
-
-class XinjiangICBCDaifaLongtailMonthlyStatSchema:
-    """新疆工行代发长尾客户月度统计表"""
-    
-    @staticmethod
-    def create() -> TableSchema:
-        """创建新疆工行代发长尾客户月度统计表结构"""
-        schema = TableSchema('xinjiang_icbc_daifa_hlwj_monthly_stat_wy')
-        schema.add_primary_key('party_id', 'string')
-        schema.add_date_field('data_dt', 'string')
-        
-        # 月度交易统计
-        schema.add_field("monthly_txn_count", "int", comment="月度交易笔数", aggregatable=True)
-        schema.add_field("monthly_txn_amount", "decimal", comment="月度交易金额", aggregatable=True)
-        schema.add_field("monthly_deposit_amount", "decimal", comment="月度存入金额", aggregatable=True)
-        schema.add_field("monthly_withdraw_amount", "decimal", comment="月度取出金额", aggregatable=True)
-        
-        # 代发工资相关统计
-        schema.add_field("salary_amount", "decimal", comment="月度代发工资金额", aggregatable=True)
-        schema.add_field("salary_date", "string", comment="代发工资日期")
-        schema.add_field("salary_stability", "float", comment="工资稳定性指数", aggregatable=True)
-        
-        # 长尾客户特征
-        schema.add_field("longtail_score", "float", comment="长尾客户评分", aggregatable=True)
-        schema.add_field("upgrade_potential", "float", comment="提升潜力评分", aggregatable=True)
-        schema.add_field("churn_risk", "float", comment="流失风险评分", aggregatable=True)
-        
-        # 活跃度指标
-        schema.add_field("login_days", "int", comment="月度登录天数", aggregatable=True)
-        schema.add_field("channel_usage", "string", comment="渠道使用情况")
-        
-        schema.set_monthly_unique(True)
-        return schema
-
-
-def get_xinjiang_icbc_daifa_longtail_schemas() -> Dict[str, TableSchema]:
-    """获取新疆工行代发长尾客户所有表结构"""
-    return {
-        'daifa_longtail_behavior': XinjiangICBCDaifaLongtailBehaviorSchema.create(),
-        'daifa_longtail_asset_avg': XinjiangICBCDaifaLongtailAssetAvgSchema.create(),
-        'daifa_longtail_asset_config': XinjiangICBCDaifaLongtailAssetConfigSchema.create(),
-        'daifa_longtail_monthly_stat': XinjiangICBCDaifaLongtailMonthlyStatSchema.create(),
-    }
-
-
-def export_xinjiang_icbc_daifa_longtail_docs(output_dir: str = "./docs") -> Dict[str, str]:
-    """导出新疆工行代发长尾客户表结构文档"""
-    generator = SchemaDocumentGenerator()
-    schemas = get_xinjiang_icbc_daifa_longtail_schemas()
-    exported_files = {}
-    
-    for table_type, schema in schemas.items():
-        file_path = generator.export_schema_doc(
-            schema, 
-            business_domain="新疆工行代发长尾客户", 
-            table_type=table_type,
-            output_dir=output_dir
-        )
-        exported_files[table_type] = file_path
-    
-    return exported_files
-
-
-# 导出主要组件
-__all__ = [
-    'XinjiangICBCDaifaLongtailBehaviorSchema',
-    'XinjiangICBCDaifaLongtailAssetAvgSchema', 
-    'XinjiangICBCDaifaLongtailAssetConfigSchema',
-    'XinjiangICBCDaifaLongtailMonthlyStatSchema',
-    'get_xinjiang_icbc_daifa_longtail_schemas',
-    'export_xinjiang_icbc_daifa_longtail_docs'
-]

staran/schemas/document_generator.py

@@ -1,350 +0,0 @@
-"""
-Schema文档生成器
-
-支持根据表结构定义生成多种格式的技术文档：
-- Markdown格式：适合开发团队和版本控制
-- PDF格式：适合正式交付和业务方审阅
-- HTML格式：适合在线查看和分享
-
-主要功能：
-1. 表结构自动解析
-2. 字段信息格式化
-3. 业务含义说明
-4. 技术规范文档
-5. 多格式导出支持
-"""
-
-import os
-from typing import Dict, List, Optional
-from datetime import datetime
-
-
-class SchemaDocumentGenerator:
-    """表结构文档生成器"""
-    
-    def __init__(self):
-        self.template_configs = {
-            'markdown': {
-                'extension': '.md',
-                'header_template': self._get_markdown_header_template(),
-                'field_template': self._get_markdown_field_template(),
-                'footer_template': self._get_markdown_footer_template()
-            },
-            'pdf': {
-                'extension': '.pdf', 
-                'requires_conversion': True,
-                'base_format': 'markdown'  # 先生成MD再转PDF
-            },
-            'html': {
-                'extension': '.html',
-                'header_template': self._get_html_header_template(),
-                'field_template': self._get_html_field_template(),
-                'footer_template': self._get_html_footer_template()
-            }
-        }
-    
-    def export_schema_doc(self, schema, business_domain: str, table_type: str, 
-                         output_dir: str = "./docs", format_type: str = "markdown") -> str:
-        """
-        导出表结构文档
-        
-        Args:
-            schema: TableSchema对象
-            business_domain: 业务域名称 (如: AUM, CRM, RISK)
-            table_type: 表类型 (如: behavior, asset_avg)
-            output_dir: 输出目录
-            format_type: 文档格式 ('markdown', 'pdf', 'html')
-            
-        Returns:
-            生成的文档文件路径
-        """
-        # 确保输出目录存在
-        os.makedirs(output_dir, exist_ok=True)
-        
-        # 生成文件名
-        timestamp = datetime.now().strftime("%Y%m%d")
-        filename = f"{business_domain}_{table_type}_schema_{timestamp}"
-        
-        if format_type.lower() == 'pdf':
-            # PDF格式先生成Markdown再转换
-            md_content = self._generate_markdown_content(schema, business_domain, table_type)
-            md_path = os.path.join(output_dir, f"{filename}.md")
-            
-            with open(md_path, 'w', encoding='utf-8') as f:
-                f.write(md_content)
-            
-            # 转换为PDF (这里可以集成pandoc或其他转换工具)
-            pdf_path = os.path.join(output_dir, f"{filename}.pdf")
-            self._convert_md_to_pdf(md_path, pdf_path)
-            return pdf_path
-            
-        elif format_type.lower() == 'html':
-            # HTML格式
-            html_content = self._generate_html_content(schema, business_domain, table_type)
-            html_path = os.path.join(output_dir, f"{filename}.html")
-            
-            with open(html_path, 'w', encoding='utf-8') as f:
-                f.write(html_content)
-            return html_path
-            
-        else:
-            # 默认Markdown格式
-            md_content = self._generate_markdown_content(schema, business_domain, table_type)
-            md_path = os.path.join(output_dir, f"{filename}.md")
-            
-            with open(md_path, 'w', encoding='utf-8') as f:
-                f.write(md_content)
-            return md_path
-    
-    def _generate_markdown_content(self, schema, business_domain: str, table_type: str) -> str:
-        """生成Markdown格式内容"""
-        content = []
-        
-        # 文档头部
-        content.append(f"# {business_domain} - {table_type.upper()}表结构文档")
-        content.append("")
-        content.append(f"## 基本信息")
-        content.append("")
-        content.append(f"- **表名**: `{schema.table_name}`")
-        content.append(f"- **业务域**: {business_domain}")
-        content.append(f"- **表类型**: {table_type}")
-        content.append(f"- **生成时间**: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
-        content.append(f"- **月度唯一性**: {'是' if getattr(schema, 'is_monthly_unique', False) else '否'}")
-        content.append("")
-        
-        # 表结构说明
-        content.append("## 表结构说明")
-        content.append("")
-        if hasattr(schema, 'description'):
-            content.append(f"{schema.description}")
-            content.append("")
-        
-        # 主键信息
-        if hasattr(schema, 'primary_key') and schema.primary_key:
-            content.append("### 主键字段")
-            content.append("")
-            content.append(f"- `{schema.primary_key}` (主键)")
-            content.append("")
-        
-        # 日期字段
-        if hasattr(schema, 'date_field') and schema.date_field:
-            content.append("### 日期字段")
-            content.append("")
-            content.append(f"- `{schema.date_field}` (日期字段)")
-            content.append("")
-        
-        # 字段详情表格
-        content.append("## 字段详情")
-        content.append("")
-        content.append("| 字段名 | 数据类型 | 业务含义 | 可聚合 | 备注 |")
-        content.append("|--------|----------|----------|--------|------|")
-        
-        if hasattr(schema, 'fields'):
-            for field_name, field in schema.fields.items():
-                # 简化数据类型显示
-                field_type_str = str(field.field_type) if hasattr(field, 'field_type') else 'string'
-                field_type = field_type_str.replace('FieldType.', '').lower()
-                comment = field.comment if hasattr(field, 'comment') else ''
-                aggregatable = '是' if getattr(field, 'aggregatable', False) else '否'
-                remarks = ''  # 可以从其他地方获取备注
-                
-                content.append(f"| `{field_name}` | {field_type} | {comment} | {aggregatable} | {remarks} |")
-        
-        content.append("")
-        
-        # 业务规则说明
-        content.append("## 业务规则")
-        content.append("")
-        content.append("### 数据更新规则")
-        if getattr(schema, 'is_monthly_unique', False):
-            content.append("- 每人每月一条记录")
-            content.append("- 月末批量更新")
-        else:
-            content.append("- 每人每日一条记录")
-            content.append("- 日终批量更新")
-        content.append("")
-        
-        content.append("### 数据质量要求")
-        content.append("- 主键字段不允许为空")
-        content.append("- 日期字段格式统一为YYYYMMDD")
-        content.append("- 金额字段精度保持2位小数")
-        content.append("- 比例字段取值范围[0,1]")
-        content.append("")
-        
-        # 使用说明
-        content.append("## 使用说明")
-        content.append("")
-        content.append("### 特征工程配置")
-        if table_type == 'behavior':
-            content.append("- 生成原始拷贝特征")
-            content.append("- 生成聚合特征")
-            content.append("- 不生成环比、同比特征")
-        else:
-            content.append("- 生成聚合特征")
-            content.append("- 生成5个月环比特征")
-            content.append("- 生成1年同比特征")
-        content.append("")
-        
-        content.append("### 示例SQL查询")
-        content.append("```sql")
-        content.append(f"-- 查询最新数据")
-        content.append(f"SELECT * FROM {schema.table_name}")
-        content.append(f"WHERE data_dt = (SELECT MAX(data_dt) FROM {schema.table_name})")
-        content.append(f"LIMIT 10;")
-        content.append("```")
-        content.append("")
-        
-        # 文档尾部
-        content.append("---")
-        content.append("*本文档由Staran Schema自动生成*")
-        
-        return "\n".join(content)
-    
-    def _generate_html_content(self, schema, business_domain: str, table_type: str) -> str:
-        """生成HTML格式内容"""
-        # 基础HTML模板，可以根据需要扩展
-        html_content = f"""
-<!DOCTYPE html>
-<html lang="zh-CN">
-<head>
-    <meta charset="UTF-8">
-    <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <title>{business_domain} - {table_type.upper()}表结构文档</title>
-    <style>
-        body {{ font-family: 'Microsoft YaHei', Arial, sans-serif; margin: 40px; }}
-        h1, h2, h3 {{ color: #333; }}
-        table {{ border-collapse: collapse; width: 100%; margin: 20px 0; }}
-        th, td {{ border: 1px solid #ddd; padding: 12px; text-align: left; }}
-        th {{ background-color: #f2f2f2; font-weight: bold; }}
-        code {{ background-color: #f4f4f4; padding: 2px 4px; border-radius: 3px; }}
-        .info-table {{ background-color: #f9f9f9; }}
-    </style>
-</head>
-<body>
-    <h1>{business_domain} - {table_type.upper()}表结构文档</h1>
-    
-    <h2>基本信息</h2>
-    <table class="info-table">
-        <tr><th>表名</th><td><code>{schema.table_name}</code></td></tr>
-        <tr><th>业务域</th><td>{business_domain}</td></tr>
-        <tr><th>表类型</th><td>{table_type}</td></tr>
-        <tr><th>生成时间</th><td>{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}</td></tr>
-        <tr><th>月度唯一性</th><td>{'是' if getattr(schema, 'is_monthly_unique', False) else '否'}</td></tr>
-    </table>
-    
-    <h2>字段详情</h2>
-    <table>
-        <thead>
-            <tr>
-                <th>字段名</th>
-                <th>数据类型</th>
-                <th>业务含义</th>
-                <th>可聚合</th>
-                <th>备注</th>
-            </tr>
-        </thead>
-        <tbody>
-"""
-        
-        # 添加字段行
-        if hasattr(schema, 'fields'):
-            for field_name, field in schema.fields.items():
-                # 简化数据类型显示
-                field_type_str = str(field.field_type) if hasattr(field, 'field_type') else 'string'
-                field_type = field_type_str.replace('FieldType.', '').lower()
-                comment = field.comment if hasattr(field, 'comment') else ''
-                aggregatable = '是' if getattr(field, 'aggregatable', False) else '否'
-                remarks = ''  # 可以从其他地方获取备注
-                
-                html_content += f"""
-            <tr>
-                <td><code>{field_name}</code></td>
-                <td>{field_type}</td>
-                <td>{comment}</td>
-                <td>{aggregatable}</td>
-                <td>{remarks}</td>
-            </tr>"""
-        
-        html_content += """
-        </tbody>
-    </table>
-    
-    <hr>
-    <p><em>本文档由Staran Schema自动生成</em></p>
-</body>
-</html>"""
-        
-        return html_content
-    
-    def _convert_md_to_pdf(self, md_path: str, pdf_path: str):
-        """将Markdown转换为PDF (需要安装pandoc或其他转换工具)"""
-        try:
-            import subprocess
-            # 尝试使用pandoc转换
-            subprocess.run([
-                'pandoc', md_path, '-o', pdf_path,
-                '--pdf-engine=xelatex',
-                '--variable=CJKmainfont:Microsoft YaHei'
-            ], check=True)
-        except (subprocess.CalledProcessError, FileNotFoundError):
-            # 如果pandoc不可用，创建一个说明文件
-            with open(pdf_path.replace('.pdf', '_conversion_note.txt'), 'w', encoding='utf-8') as f:
-                f.write(f"PDF转换说明：\\n")
-                f.write(f"原始Markdown文件：{md_path}\\n")
-                f.write(f"如需PDF格式，请安装pandoc工具：\\n")
-                f.write(f"pip install pandoc\\n")
-                f.write(f"或访问：https://pandoc.org/installing.html\\n")
-    
-    def _get_markdown_header_template(self) -> str:
-        return "# {title}\\n\\n## 基本信息\\n\\n"
-    
-    def _get_markdown_field_template(self) -> str:
-        return "| {name} | {type} | {comment} | {aggregatable} |\\n"
-    
-    def _get_markdown_footer_template(self) -> str:
-        return "\\n---\\n*文档生成时间: {timestamp}*\\n"
-    
-    def _get_html_header_template(self) -> str:
-        return "<h1>{title}</h1>\\n<h2>基本信息</h2>\\n"
-    
-    def _get_html_field_template(self) -> str:
-        return "<tr><td>{name}</td><td>{type}</td><td>{comment}</td><td>{aggregatable}</td></tr>\\n"
-    
-    def _get_html_footer_template(self) -> str:
-        return "<hr><p><em>文档生成时间: {timestamp}</em></p>\\n"
-
-
-def export_business_docs(business_domain: str, schemas_dict: Dict, output_dir: str = "./docs", 
-                        format_type: str = "markdown") -> Dict[str, str]:
-    """
-    批量导出业务域表结构文档
-    
-    Args:
-        business_domain: 业务域名称
-        schemas_dict: 表结构字典 {table_type: schema}
-        output_dir: 输出目录
-        format_type: 文档格式
-        
-    Returns:
-        生成的文档文件路径字典
-    """
-    generator = SchemaDocumentGenerator()
-    results = {}
-    
-    for table_type, schema in schemas_dict.items():
-        file_path = generator.export_schema_doc(
-            schema=schema,
-            business_domain=business_domain,
-            table_type=table_type,
-            output_dir=output_dir,
-            format_type=format_type
-        )
-        results[table_type] = file_path
-    
-    return results
-
-
-__all__ = [
-    'SchemaDocumentGenerator',
-    'export_business_docs'
-]