yms-kan 0.0.2__py3-none-any.whl → 0.0.4__py3-none-any.whl

yms_kan/tool.py

@@ -1,8 +1,155 @@
 import os
+import re
+from datetime import datetime, timezone, timedelta
+from typing import Optional, Dict, List
 
+import click
 import numpy as np
-import torch
+import pandas as pd
 import wandb
+from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix, \
+    classification_report
+from tqdm import tqdm
+
+
+# 读取txt内两个不同表格的数据，并将结果转换为字典列表输出
+def read_multi_table_txt(file_path):
+    # 读取原始内容
+    with open(file_path, 'r') as f:
+        content = f.read()
+
+    # 按表格标题分割内容（假设每个新表格以"epoch"开头）
+    table_blocks = re.split(r'\n(?=epoch\s)', content.strip())
+
+    # 处理每个表格块
+    table_dicts = []
+    for block in table_blocks:
+        lines = [line.strip() for line in block.split('\n') if line.strip()]
+
+        # 解析列名（处理制表符和混合空格）
+        columns = re.split(r'\s{2,}|\t', lines[0])
+
+        # 解析数据行（处理混合分隔符）
+        data = []
+        for line in lines[1:]:
+            # 使用正则表达式分割多个连续空格/制表符
+            row = re.split(r'\s{2,}|\t', line)
+            data.append(row)
+
+        # 创建DataFrame并自动转换数值类型
+        df = pd.DataFrame(data, columns=columns)
+        df = df.apply(pd.to_numeric, errors='coerce')  # 自动识别数值列，非数值转换为NaN
+
+        # 将DataFrame转换为字典，每列以列表形式保存
+        table_dict = df.to_dict(orient='list')
+        table_dicts.append(table_dict)
+
+    return table_dicts
+
+
+def get_current_time(format_str="%Y-%m-%d %H:%M:%S"):
+    """
+    获取东八区（UTC+8）的当前时间，并返回指定格式的字符串
+    :param format_str: 时间格式（默认为 "%Y-%m-%d %H:%M:%S"）
+    :return: 格式化后的时间字符串
+    """
+
+    # 创建东八区的时区对象
+    utc8_timezone = timezone(timedelta(hours=8))
+
+    # 转换为东八区时间
+    utc8_time = datetime.now(utc8_timezone)
+
+    # 格式化为字符串
+    formatted_time = utc8_time.strftime(format_str)
+    return formatted_time
+
+
+# val和test时的相关结果指标计算
+def calculate_results(all_labels, all_predictions, classes, average='macro'):
+    results = {
+        'accuracy': accuracy_score(y_true=all_labels, y_pred=all_predictions),
+        'precision': precision_score(y_true=all_labels, y_pred=all_predictions, average=average),
+        'recall': recall_score(y_true=all_labels, y_pred=all_predictions, average=average),
+        'f1_score': f1_score(y_true=all_labels, y_pred=all_predictions, average=average),
+        'cm': confusion_matrix(y_true=all_labels, y_pred=all_predictions, labels=np.arange(len(classes)))
+    }
+    return results
+
+
+def calculate_metric(all_labels, all_predictions, classes, class_metric=False, average='macro avg'):
+    metric = classification_report(y_true=all_labels, y_pred=all_predictions,
+                                   target_names=classes, digits=4, output_dict=True, zero_division=0)
+    if not class_metric:
+        metric = {
+            'accuracy': metric.get('accuracy'),
+            'precision': metric.get(average).get('precision'),
+            'recall': metric.get(average).get('recall'),
+            'f1-score': metric.get(average).get('f1-score'),
+        }
+        return metric
+    else:
+        return metric
+
+
+def dict_to_classification_report(report_dict, digits=2):
+    headers = ["precision", "recall", "f1-score", "support"]
+    target_names = list(report_dict.keys())
+    target_names.remove('accuracy') if 'accuracy' in target_names else None
+    longest_last_line_heading = "weighted avg"
+    name_width = max(len(cn) for cn in target_names)
+    width = max(name_width, len(longest_last_line_heading), digits)
+    head_fmt = "{:>{width}s} " + " {:>9}" * len(headers)
+    report = head_fmt.format("", *headers, width=width)
+    report += "\n\n"
+    row_fmt = "{:>{width}s} " + " {:>9.{digits}f}" * 3 + " {:>9}\n"
+    for target_name in target_names:
+        scores = [report_dict[target_name][h] for h in headers]
+        report += row_fmt.format(target_name, *scores, width=width, digits=digits)
+    report += "\n"
+
+    average_options = ["micro avg", "macro avg", "weighted avg"]
+    if 'samples avg' in report_dict:
+        average_options.append('samples avg')
+    for average in average_options:
+        if average in report_dict:
+            scores = [report_dict[average][h] for h in headers]
+            if average == "accuracy":
+                row_fmt_accuracy = (
+                        "{:>{width}s} "
+                        + " {:>9.{digits}}" * 2
+                        + " {:>9.{digits}f}"
+                        + " {:>9}\n"
+                )
+                report += row_fmt_accuracy.format(
+                    average, "", "", *scores[2:], width=width, digits=digits
+                )
+            else:
+                report += row_fmt.format(average, *scores, width=width, digits=digits)
+
+    if 'accuracy' in report_dict:
+        row_fmt_accuracy = (
+                "{:>{width}s} "
+                + " {:>9.{digits}}" * 2
+                + " {:>9.{digits}f}"
+                + " {:>9}\n"
+        )
+        report += row_fmt_accuracy.format(
+            "accuracy", "", "", report_dict["accuracy"], "", width=width, digits=digits
+        )
+
+    return report
+
+
+# def append_metrics(metrics, metric, result, lr):
+#     metrics['train_losses'].append(result['train_loss'])
+#     metrics['val_losses'].append(result['val_loss'])
+#     metrics['accuracies'].append(metric['accuracy'])
+#     metrics['precisions'].append(metric['precision'])
+#     metrics['recalls'].append(metric['recall'])
+#     metrics['f1-scores'].append(metric['f1-score'])
+#     metrics['lrs'].append(lr)
+#     return metrics
 
 
 def initialize_results_file(results_file, result_info):
@@ -12,9 +159,10 @@ def initialize_results_file(results_file, result_info):
     参数:
         results_file (str): 结果文件的路径。
         result_info (list): 需要写入的第一行内容列表。
+        space:列名间隔（默认两个空格的距离）
     """
     # 处理 result_info，在每个单词后添加两个空格
-    result_info_str = "  ".join(result_info) + '\n'
+    result_info_str = '  '.join(result_info) + '\n'
     # 检查文件是否存在
     if os.path.exists(results_file):
         # 如果文件存在，读取第一行
@@ -35,86 +183,45 @@ def initialize_results_file(results_file, result_info):
         print(f"文件 {results_file} 已创建并写入 result_info。")
 
 
-def write_results_file(file_path: str,
-                       data_dict: dict,
-                       column_order: list,
-                       float_precision: int = 5) -> None:
+def is_similar_key(key1, key2):
     """
-    通用格式化文本行写入函数（支持列表形式数据）
+    检查两个键是否相似，考虑复数形式的转换。
 
-    参数：
-    file_path: 目标文件路径
-    data_dict: 包含数据的字典，键为列名，值为列表
-    column_order: 列顺序列表，元素为字典键
-    float_precision: 浮点数精度位数 (默认5位)
+    Args:
+        key1 (str): 第一个键
+        key2 (str): 第二个键
+
+    Returns:
+        bool: 如果两个键相似（包括复数形式的转换），返回 True，否则返回 False
     """
-    # 验证数据格式
-    rows = None
-    for key in data_dict:
-        if not isinstance(data_dict[key], list):
-            raise ValueError(f"Value for key '{key}' is not a list")
-        if rows is None:
-            rows = len(data_dict[key])
-        else:
-            if len(data_dict[key]) != rows:
-                raise ValueError("All lists in data_dict must have the same length")
+    if key1 == key2:
+        return True
 
-    # 辅助函数：格式化单个值
-    def format_value(value, column_name):
-        if isinstance(value, (int, np.integer)):
-            return f"{value:d}"
-        elif isinstance(value, (float, np.floating)):
-            if column_name in ['train_losses', 'val_losses']:
-                return f"{value:.{float_precision + 1}f}"
-            elif column_name == 'lrs':
-                return f"{value:.8f}"
-            else:
-                return f"{value:.{float_precision}f}"
-        elif isinstance(value, str):
-            return value
-        else:
-            return str(value)
+    # 检查 key2 是否是复数形式
+    if key2.endswith("ies"):
+        singular_candidate = key2.removesuffix("ies") + "y"
+        if key1 == singular_candidate:
+            return True
 
-    # 计算列宽
-    column_widths = []
-    for col in column_order:
-        dict_key = 'val_accuracies' if col == 'accuracies' else col
-        if dict_key not in data_dict:
-            raise ValueError(f"Missing required column: {dict_key}")
-        values = data_dict[dict_key]
-
-        max_width = len(col)
-        for val in values:
-            fmt_val = format_value(val, col)
-            max_width = max(max_width, len(fmt_val))
-        column_widths.append(max_width)
-
-    # 生成格式化行
-    lines = []
-    for i in range(rows):
-        row = []
-        for j, col in enumerate(column_order):
-            dict_key = 'val_accuracies' if col == 'accuracies' else col
-            val = data_dict[dict_key][i]
-            fmt_val = format_value(val, col)
-
-            # 对齐处理
-            if j == len(column_order) - 1:
-                fmt_val = fmt_val.ljust(column_widths[j])
-            else:
-                fmt_val = fmt_val.rjust(column_widths[j])
-            row.append(fmt_val)
-        lines.append("  ".join(row) + '\n')
+    if key2.endswith("es"):
+        singular_candidate = key2.removesuffix("es")
+        if key1 == singular_candidate:
+            return True
 
-    # 写入文件
-    with open(file_path, 'a', encoding='utf-8') as f:
-        f.writelines(lines)
+    if key2.endswith("s"):
+        singular_candidate = key2.removesuffix("s")
+        if key1 == singular_candidate:
+            return True
+
+    return False
 
 
 def append_to_results_file(file_path: str,
                            data_dict: dict,
                            column_order: list,
-                           float_precision: int = 5) -> None:
+                           float_precision: int = 4,
+                           more_float: int = 2,
+                           custom_column_widths: dict = None) -> None:
     """
     通用格式化文本行写入函数
 
@@ -123,182 +230,340 @@ def append_to_results_file(file_path: str,
     data_dict: 包含数据的字典，键为列名
     column_order: 列顺序列表，元素为字典键
     float_precision: 浮点数精度位数 (默认5位)
+    more_float: 额外的浮点数精度位数
+    custom_column_widths: 自定义列宽的字典，键为列名，值为列宽
     """
-    # 检查 data_dict 中的值是否为列表
-    all_values_are_lists = all(isinstance(value, list) for value in data_dict.values())
-    if all_values_are_lists:
-        num_rows = len(next(iter(data_dict.values())))
-        # 逐行处理
-        for row_index in range(num_rows):
-            formatted_data = []
-            column_widths = []
-            for col in column_order:
-                # 处理字典键的别名
-                dict_key = 'val_accuracies' if col == 'accuracies' else col
-                # 如果键不存在，跳过该列
-                if dict_key not in data_dict:
-                    continue
-                value_list = data_dict[dict_key]
-                if row_index >= len(value_list):
-                    continue
-                value = value_list[row_index]
-
-                # 根据数据类型进行格式化
-                if isinstance(value, (int, np.integer)):
-                    fmt_value = f"{value:d}"
-                elif isinstance(value, (float, np.floating)):
-                    if col in ['train_losses', 'val_losses']:  # 如果列名是'train_losses'或'val_losses'，保留浮点数精度位数+1位
-                        fmt_value = f"{value:.{float_precision + 1}f}"
-                    elif col == 'lrs':  # 如果列名是'lrs'，保留8位小数
-                        fmt_value = f"{value:.8f}"
-                    else:
-                        fmt_value = f"{value:.{float_precision}f}"
-                elif isinstance(value, str):
-                    fmt_value = value
-                else:  # 处理其他类型转换为字符串
-                    fmt_value = str(value)
-
-                # 取列名长度和数值长度的最大值作为列宽
-                column_width = max(len(col), len(fmt_value))
-                column_widths.append(column_width)
-
-                # 应用列宽对齐
-                if col == column_order[-1]:  # 最后一列左边对齐
-                    fmt_value = fmt_value.ljust(column_width)
-                else:
-                    fmt_value = fmt_value.rjust(column_width)
-
-                formatted_data.append(fmt_value)
-
-            # 构建文本行并写入，列之间用两个空格分隔
-            if formatted_data:
-                line = "  ".join(formatted_data) + '\n'
-                with open(file_path, 'a', encoding='utf-8') as f:
-                    f.write(line)
-    else:
-        # 非列表情况，原逻辑处理
-        # 计算每列的最大宽度
-        column_widths = []
-        formatted_data = []
-        for col in column_order:
-            # 处理字典键的别名
-            dict_key = 'val_accuracies' if col == 'accuracies' else col
-            # 如果键不存在，跳过该列
-            if dict_key not in data_dict:
-                continue
-
-            value = data_dict[dict_key]
-
-            # 根据数据类型进行格式化
-            if isinstance(value, (int, np.integer)):
-                fmt_value = f"{value:d}"
-            elif isinstance(value, (float, np.floating)):
-                if col in ['train_losses', 'val_losses']:  # 如果列名是'train_losses'或'val_losses'，保留浮点数精度位数+1位
-                    fmt_value = f"{value:.{float_precision + 1}f}"
-                elif col == 'lrs':  # 如果列名是'lrs'，保留8位小数
-                    fmt_value = f"{value:.8f}"
-                else:
-                    fmt_value = f"{value:.{float_precision}f}"
-            elif isinstance(value, str):
-                fmt_value = value
-            else:  # 处理其他类型转换为字符串
-                fmt_value = str(value)
+    # 计算每列的最大宽度
+    column_widths = []
+    formatted_data = []
+    for col in column_order:
+        # 查找 data_dict 中相似的键
+        dict_key = None
+        for key in data_dict:
+            if is_similar_key(key, col):
+                dict_key = key
+                break
+        if dict_key is None:
+            raise ValueError(f"Missing required column: {col}")
+
+        value = data_dict[dict_key]
+
+        # 根据数据类型进行格式化
+        if isinstance(value, (int, np.integer)):
+            fmt_value = f"{value:d}"
+        elif isinstance(value, (float, np.floating)):
+            if col in ['train_losses', 'val_losses']:  # 如果列名是'train_losses'或'val_losses'，保留浮点数精度位数+1位
+                fmt_value = f"{value:.{float_precision + more_float}f}"
+            elif col == 'lrs':
+                fmt_value = f"{value:.8f}"
+            else:
+                fmt_value = f"{value:.{float_precision}f}"
+        elif isinstance(value, str):
+            fmt_value = value
+        else:  # 处理其他类型转换为字符串
+            fmt_value = str(value)
 
+        # 确定列宽
+        if custom_column_widths and col in custom_column_widths:
+            column_width = custom_column_widths[col]
+        else:
             # 取列名长度和数值长度的最大值作为列宽
             column_width = max(len(col), len(fmt_value))
-            column_widths.append(column_width)
+        column_widths.append(column_width)
 
-            # 应用列宽对齐
-            if col == column_order[-1]:  # 最后一列左边对齐
-                fmt_value = fmt_value.ljust(column_width)
-            else:
-                fmt_value = fmt_value.rjust(column_width)
-
-            formatted_data.append(fmt_value)
-
-        # 构建文本行并写入，列之间用两个空格分隔
-        if formatted_data:
-            line = "  ".join(formatted_data) + '\n'
-            with open(file_path, 'a', encoding='utf-8') as f:
-                f.write(line)
-
-
-# def append_to_results_file(file_path: str,
-#                            data_dict: dict,
-#                            column_order: list,
-#                            column_widths: list = None,
-#                            float_precision: int = 5) -> None:
-#     """
-#     通用格式化文本行写入函数
-#
-#     参数：
-#     file_path: 目标文件路径
-#     data_dict: 包含数据的字典，键为列名
-#     column_order: 列顺序列表，元素为字典键
-#     column_widths: 每列字符宽度列表 (可选)
-#     float_precision: 浮点数精度位数 (默认4位)
-#     """
-#     formatted_data = []
-#
-#     # 遍历指定列顺序处理数据
-#     for i, col in enumerate(column_order):
-#         # 处理字典键的别名
-#         if col == 'accuracies':
-#             dict_key = 'val_accuracies'
-#         else:
-#             dict_key = col
-#
-#         if dict_key not in data_dict:
-#             raise ValueError(f"Missing required column: {dict_key}")
-#
-#         value = data_dict[dict_key]
-#
-#         # 根据数据类型进行格式化
-#         if isinstance(value, (int, np.integer)):
-#             fmt_value = f"{value:d}"
-#         elif isinstance(value, (float, np.floating)):
-#             if col in ['train_losses', 'val_losses']:  # 如果列名是'train_losses'或'val_losses'，保留浮点数精度位数+1位
-#                 fmt_value = f"{value:.{float_precision + 1}f}"
-#             elif col == 'lr':  # 如果列名是'lr'，保留8位小数
-#                 fmt_value = f"{value:.8f}"
-#             else:
-#                 fmt_value = f"{value:.{float_precision}f}"
-#         elif isinstance(value, str):
-#             fmt_value = value
-#         else:  # 处理其他类型转换为字符串
-#             fmt_value = str(value)
-#
-#         # 应用列宽对齐
-#         if column_widths and i < len(column_widths):
-#             try:
-#                 if i == len(column_order) - 1:  # 最后一列左边对齐
-#                     fmt_value = fmt_value.ljust(column_widths[i])
-#                 else:
-#                     fmt_value = fmt_value.rjust(column_widths[i])
-#             except TypeError:  # 处理非字符串类型
-#                 if i == len(column_order) - 1:  # 最后一列左边对齐
-#                     fmt_value = str(fmt_value).ljust(column_widths[i])
-#                 else:
-#                     fmt_value = str(fmt_value).rjust(column_widths[i])
-#
-#         formatted_data.append(fmt_value)
-#
-#     # 构建文本行并写入
-#     line = '\t'.join(formatted_data) + '\n'
-#     with open(file_path, 'a', encoding='utf-8') as f:
-#         f.write(line)
-
-
-def get_wandb_key(key_path='tools/wandb_key.txt'):
+        # 应用列宽对齐
+        if col == column_order[-1]:  # 最后一列左边对齐
+            fmt_value = fmt_value.ljust(column_width)
+        else:
+            fmt_value = fmt_value.rjust(column_width)
+
+        formatted_data.append(fmt_value)
+
+    # 构建文本行并写入，列之间用两个空格分隔
+    line = "  ".join(formatted_data) + '\n'
+    with open(file_path, 'a', encoding='utf-8') as f:
+        f.write(line)
+
+
+def get_wandb_key(key_path):
     with open(key_path, 'r', encoding='utf-8') as f:
         key = f.read()
     return key
 
 
-def wandb_use(project=None, name=None, key_path='tools/wandb_key.txt'):
+def wandb_init(project=None, key_path=None, name=None):
     run = None
     if project is not None:
+        if key_path is None:
+            raise ValueError("When 'project' is not None, 'key_path' should also not be None.")
         wandb_key = get_wandb_key(key_path)
         wandb.login(key=wandb_key)
         run = wandb.init(project=project, name=name)
     return run
+
+
+def check_wandb_login_required():
+    """兼容旧版的登录检查函数"""
+    # 优先检查环境变量
+    if os.environ.get("WANDB_API_KEY"):
+        return False
+
+    try:
+        api = wandb.Api()
+        # 方法 1：通过 settings 检查（适用于旧版）
+        if hasattr(api, "settings") and api.settings.get("entity"):
+            return False
+
+        # 方法 2：通过 projects() 验证（通用性强）
+        api.projects(per_page=1)  # 仅请求第一页的第一个项目
+        return False
+    except Exception as e:
+        print(f"检测到意外错误: {str(e)}")
+        return True  # 保守返回需要登录
+
+
+def get_wandb_runs(
+        project_path: str,
+        default_name: str = "未命名",
+        api_key: Optional[str] = None,
+        per_page: int = 1000
+) -> List[Dict[str, str]]:
+    """
+    获取指定 WandB 项目的所有运行信息（ID 和 Name）
+
+    Args:
+        project_path (str): 项目路径，格式为 "username/project_name"
+        default_name (str): 当运行未命名时的默认显示名称（默认："未命名"）
+        api_key (str, optional): WandB API 密钥，若未设置环境变量则需传入
+        per_page (int): 分页查询每页数量（默认1000，用于处理大量运行）
+
+    Returns:
+        List[Dict]: 包含运行信息的字典列表，格式 [{"id": "...", "name": "..."}]
+
+    Raises:
+        ValueError: 项目路径格式错误
+        wandb.errors.UsageError: API 密钥无效或未登录
+    """
+    # 参数校验
+    if "/" not in project_path or len(project_path.split("/")) != 2:
+        raise ValueError("项目路径格式应为 'username/project_name'")
+
+    # 登录（仅在需要时）
+    if api_key:
+        wandb.login(key=api_key)
+    elif not wandb.api.api_key:
+        raise wandb.errors.UsageError("需要提供API密钥或预先调用wandb.login()")
+
+    # 初始化API
+    api = wandb.Api()
+
+    try:
+        # 分页获取所有运行（自动处理分页逻辑）
+        runs = api.runs(project_path, per_page=per_page)
+        print(f'共获取{len(runs)}个run')
+        return [
+            {
+                "id": run.id,
+                "name": run.name or default_name,
+                "url": run.url,  # 增加实用字段
+                "state": run.state  # 包含运行状态
+            }
+            for run in runs
+        ]
+
+    except wandb.errors.CommError as e:
+        raise ConnectionError(f"连接失败: {str(e)}") from e
+    except Exception as e:
+        raise RuntimeError(f"获取运行数据失败: {str(e)}") from e
+
+
+def delete_runs(
+        project_path: str,
+        run_ids: Optional[List[str]] = None,
+        run_names: Optional[List[str]] = None,
+        delete_all: bool = False,
+        dry_run: bool = True,
+        api_key: Optional[str] = None,
+        per_page: int = 500
+) -> dict:
+    """
+    多功能WandB运行删除工具
+
+    :param project_path: 项目路径（格式：username/project_name）
+    :param run_ids: 指定要删除的运行ID列表（无视状态）
+    :param run_names: 指定要删除的运行名称列表（无视状态）
+    # :param preserve_states: 保护状态列表（默认保护 finished/running）
+    :param delete_all: 危险模式！删除所有运行（默认False）
+    :param dry_run: 模拟运行模式（默认True）
+    :param api_key: WandB API密钥
+    :param per_page: 分页查询数量
+    :return: 操作统计字典
+
+    使用场景：
+    1. 删除指定运行：delete_runs(..., run_ids=["abc","def"])
+    2. 默认删除失败运行：delete_runs(...)
+    3. 删除所有运行：delete_runs(..., delete_all=True)
+    """
+    preserve_states: List[str] = ["finished", "running"]
+    # 参数校验
+    if not project_path.count("/") == 1:
+        raise ValueError("项目路径格式应为 username/project_name")
+    if delete_all and (run_ids or run_names):
+        raise ValueError("delete_all模式不能与其他筛选参数同时使用")
+
+    # 身份验证
+    if api_key:
+        wandb.login(key=api_key)
+    elif not wandb.api.api_key:
+        raise wandb.errors.UsageError("需要API密钥或预先登录")
+
+    api = wandb.Api()
+    stats = {
+        "total": 0,
+        "candidates": 0,
+        "deleted": 0,
+        "failed": 0,
+        "dry_run": dry_run
+    }
+
+    try:
+        runs = api.runs(project_path, per_page=per_page)
+        stats["total"] = len(runs)
+
+        # 确定删除目标
+        if delete_all:
+            targets = runs
+            click.secho("\n⚠️ 危险操作：将删除项目所有运行！", fg="red", bold=True)
+        elif run_ids or run_names:
+            targets = [
+                run for run in runs
+                if run.id in (run_ids or []) or run.name in (run_names or [])
+            ]
+            print(f"\n找到 {len(targets)} 个指定运行")
+        else:
+            targets = [run for run in runs if run.state not in preserve_states]
+            print(f"\n找到 {len(targets)} 个非正常状态运行")
+
+        stats["candidates"] = len(targets)
+
+        if not targets:
+            print("没有符合条件的运行")
+            return stats
+
+        # 打印预览
+        print("\n待删除运行示例：")
+        for run in targets[:3]:
+            state = click.style(run.state, fg="green" if run.state == "finished" else "red")
+            print(f" • {run.id} | {run.name} | 状态：{state}")
+        if len(targets) > 3:
+            print(f" ...（共 {len(targets)} 条）")
+
+        # 安全确认
+        if dry_run:
+            click.secho("\n模拟运行模式：不会实际删除", fg="yellow")
+            return stats
+
+        if delete_all:
+            msg = click.style("确认要删除所有运行吗？此操作不可逆！", fg="red", bold=True)
+        else:
+            msg = f"确认要删除 {len(targets)} 个运行吗？"
+
+        if not click.confirm(msg, default=False):
+            print("操作已取消")
+            return stats
+
+        # 执行删除
+        print("\n删除进度：")
+        for i, run in enumerate(targets, 1):
+            try:
+                run.delete()
+                stats["deleted"] += 1
+                print(click.style(f"  [{i}/{len(targets)}] 已删除 {run.id}", fg="green"))
+            except Exception as e:
+                stats["failed"] += 1
+                print(click.style(f"  [{i}/{len(targets)}] 删除失败 {run.id}: {str(e)}", fg="red"))
+
+        return stats
+
+    except wandb.errors.CommError as e:
+        raise ConnectionError(f"网络错误: {str(e)}")
+    except Exception as e:
+        raise RuntimeError(f"操作失败: {str(e)}")
+
+
+def get_all_artifacts_from_project(project_path, max_runs=None, run_id=None):
+    """获取WandB项目或指定Run的所有Artifact
+
+    Args:
+        project_path (str): 项目路径，格式为 "entity/project"
+        max_runs (int, optional): 最大获取Run数量（仅当未指定run_id时生效）
+        run_id (str, optional): 指定要查询的Run ID
+
+    Returns:
+        list: 包含所有Artifact对象的列表
+    """
+    api = wandb.Api()
+    all_artifacts = []
+    seen_artifacts = set()  # 用于去重
+
+    try:
+        if run_id:
+            # 处理单个Run的情况
+            run = api.run(f"{project_path}/{run_id}")
+            artifacts = run.logged_artifacts()
+
+            for artifact in artifacts:
+                artifact_id = f"{artifact.name}:{artifact.version}"
+                if artifact_id not in seen_artifacts:
+                    all_artifacts.append(artifact)
+                    seen_artifacts.add(artifact_id)
+
+            print(f"Found {len(all_artifacts)} artifacts in run {run_id}")
+        else:
+            # 处理整个项目的情况
+            runs = api.runs(project_path, per_page=500)
+            run_iterator = tqdm(runs[:max_runs] if max_runs else runs,
+                                desc=f"Scanning {project_path}")
+
+            for run in run_iterator:
+                try:
+                    artifacts = run.logged_artifacts()
+                    for artifact in artifacts:
+                        artifact_id = f"{artifact.name}:{artifact.version}"
+                        if artifact_id not in seen_artifacts:
+                            all_artifacts.append(artifact)
+                            seen_artifacts.add(artifact_id)
+                except Exception as run_error:
+                    print(f"Error processing run {run.id}: {str(run_error)}")
+
+    except Exception as e:
+        print(f"Error: {str(e)}")
+        return []
+
+    return all_artifacts
+
+
+def upload_model_dataset(
+        artifact_dir: str,
+        artifact_name: str,
+        artifact_type: str) -> None:
+    run_id = f'yms_upload_{artifact_type}_' + get_current_time('%y%m%d_%H%M%S')
+    run = wandb.init(project='upload_model_dataset', name=artifact_name, id=run_id)
+    artifact = wandb.Artifact(artifact_name, artifact_type)
+    artifact.add_dir(artifact_dir)
+    run.log_artifact(artifact)
+    run.finish()
+
+
+def download_model_dataset(
+    download_name: str,
+    run_name: str,
+    artifact_type: str,
+    download_dir: str = None,
+    entity: str = 'YNA-DeepLearning'
+) -> str:
+    run_id = f'yms_download_{artifact_type}_' + get_current_time('%y%m%d_%H%M%S')
+    run = wandb.init(project='download_model_dataset', name=run_name, id=run_id)
+    artifact = run.use_artifact(entity + '/upload_model_dataset/' + download_name, type=artifact_type)
+    artifact_dir = artifact.download(root=download_dir)
+    return artifact_dir

yms_kan/train_eval_utils.py

@@ -1,25 +1,52 @@
 import math
 import os
 import sys
-from enum import Enum, auto
 
 import numpy as np
 import torch
 from matplotlib import pyplot as plt
+from sklearn.metrics import classification_report
 from torch.optim.lr_scheduler import ReduceLROnPlateau
 from tqdm import tqdm
 
 from yms_kan import LBFGS
+from yms_kan.tool import initialize_results_file, append_to_results_file
 
 
-def train_val(model, dataset: dict, batch_size, batch_size_test, opt="LBFGS", epochs=100, lamb=0.,
-              lamb_l1=1., label=None, lamb_entropy=2., lamb_coef=0.,
+def calculate_metric(all_labels, all_predictions, classes, class_metric=False, average='macro avg'):
+    metric = classification_report(y_true=all_labels, y_pred=all_predictions,
+                                   target_names=classes, digits=4, output_dict=True, zero_division=0)
+    if not class_metric:
+        metric = {
+            'accuracy': metric.get('accuracy'),
+            'precision': metric.get(average).get('precision'),
+            'recall': metric.get(average).get('recall'),
+            'f1-score': metric.get(average).get('f1-score'),
+        }
+        return metric
+    else:
+        return metric
+
+
+def train_val(model, dataset: dict, batch_size, batch_size_test, save_path, txt_file=None, opt="LBFGS", epochs=100,
+              lamb=0.,
+              lamb_l1=1., label=None, class_dict=None, lamb_entropy=2., lamb_coef=0.,
               lamb_coefdiff=0., update_grid=True, grid_update_num=10, loss_fn=None, lr=1., start_grid_update_step=-1,
               stop_grid_update_step=100,
               save_fig=False, in_vars=None, out_vars=None, beta=3, save_fig_freq=1, img_folder='./video',
               singularity_avoiding=False, y_th=1000., reg_metric='edge_forward_spline_n'):
-    # all_predictions = []
-    # all_labels = []
+    all_predictions = []
+    all_labels = []
+    best = -1
+    column_order = ['epoch', 'train_losses', 'val_losses', 'accuracies', 'precisions', 'recalls',
+                    'f1-scores', 'lrs']
+    custom_column_widths = {'epoch': 5, 'train_loss': 12, 'val_loss': 10, 'accuracy': 10, 'precision': 9,
+                            'recall': 7,
+                            'f1-score': 8,
+                            'lr': 3}
+    if txt_file is not None:
+        initialize_results_file(txt_file, column_order)
+
     if lamb > 0. and not model.save_act:
         print('setting lamb=0. If you want to set lamb > 0, set model.save_act=True')
 
@@ -44,8 +71,8 @@ def train_val(model, dataset: dict, batch_size, batch_size_test, opt="LBFGS", ep
 
     lr_scheduler = ReduceLROnPlateau(optimizer, mode='min', factor=0.1, patience=5, min_lr=1e-9)
 
-    results = {'train_loss': .0, 'val_loss': .0, 'regularize': .0, 'all_predictions': [],
-               'all_labels': []}
+    results = {'train_losses': [], 'val_losses': [], 'accuracies': [], 'precisions': [], 'recalls': [], 'f1-scores': [],
+               'lrs': [], 'all_predictions': [], 'all_labels': []}
 
     steps = math.ceil(dataset['train_input'].shape[0] / batch_size)
 
@@ -118,7 +145,6 @@ def train_val(model, dataset: dict, batch_size, batch_size_test, opt="LBFGS", ep
                 optimizer.step()
             train_pbar.set_postfix(loss=train_loss.item())
 
-        # print(f'{epoch}/{epochs}:train_loss:{train_loss.item()}')
         val_loss = torch.zeros(1).to(model.device)
         with torch.no_grad():
             test_indices = np.arange(dataset['test_input'].shape[0])
@@ -142,14 +168,34 @@ def train_val(model, dataset: dict, batch_size, batch_size_test, opt="LBFGS", ep
                     diffs = torch.abs(outputs - label)
                     closest_indices = torch.argmin(diffs, dim=1)
                     closest_values = label[closest_indices]
-                    results['all_predictions'].extend(closest_values.detach().cpu().numpy())
-                    results['all_labels'].extend(batch_test_label.detach().cpu().numpy())
+                    all_predictions.extend(closest_values.detach().cpu().numpy())
+                    all_labels.extend(batch_test_label.detach().cpu().numpy())
 
+            train_lr = lr_scheduler.get_last_lr()[0]
             lr_scheduler.step(val_loss)
 
-        results['train_loss'] = train_loss.item()
-        results['val_loss'] = val_loss.item()
-        results['regularize'] = reg_.item()
+        if label is not None:
+            m = calculate_metric(all_labels, all_predictions, class_dict)
+            print(m)
+            results["accuracies"].append(m["accuracy"])
+            results["precisions"].append(m["precision"])
+            results["recalls"].append(m["recall"])
+            results["f1-scores"].append(m["f1-score"])
+            results["lrs"].append(train_lr)
+            if best < m["f1-score"]:
+                best = m["f1-score"]
+                results['all_predictions'] = all_predictions
+                results['all_labels'] = all_labels
+                if save_path is not None:
+                    model.saveckpt(path=save_path + '/' + 'model')
+            if txt_file is not None:
+                m.update({'lr': train_lr, 'epoch': epoch, 'train_loss': train_loss.item(), 'val_loss': val_loss.item()})
+                append_to_results_file(txt_file, m, column_order,
+                                       custom_column_widths=custom_column_widths)
+
+        results["train_losses"].append(train_loss.item())
+        results["val_losses"].append(val_loss.item())
+        results["regularize"].append(reg_.item())
 
         if save_fig and epoch % save_fig_freq == 0:
             model.plot(folder=img_folder, in_vars=in_vars, out_vars=out_vars, title="Step {}".format(epoch),

yms_kan/version.py

@@ -1 +1 @@
-__version__ = "0.0.2"  # 初始版本
+__version__ = "0.0.4"  # 初始版本

{yms_kan-0.0.2.dist-info → yms_kan-0.0.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: yms_kan
-Version: 0.0.2
+Version: 0.0.4
 Summary: My awesome package
 Author-email: yms <11@qq.com>
 License-Expression: MIT

{yms_kan-0.0.2.dist-info → yms_kan-0.0.4.dist-info}/RECORD

@@ -9,14 +9,14 @@ yms_kan/experiment.py,sha256=VnZq7hmcvRk08GNI7VIpkOjkaRZBoIw1C8SU_f1KbaA,1682
 yms_kan/feynman.py,sha256=Eisf69K49s4C6UlPEi5LnNK_p5TUJQLBKxMp-sW0a9w,33687
 yms_kan/hypothesis.py,sha256=Ec20xadfgOSSWeZHQaGn-h9F2PY7LWFU3iniNI2Zd_4,23165
 yms_kan/spline.py,sha256=ZXyGwl2Sc-UrnrcuUXeUQkBOMnetaWcHrbpZaqatCvs,4345
-yms_kan/tool.py,sha256=CLIsOYWwG-A5PJvoyIP8cRBzX8iRhEssW-2uXdLfi-U,12124
-yms_kan/train_eval_utils.py,sha256=73pA3-HDPDik_yCsDW0oF1dIvVu_vPeHbvJ08o26ygQ,14867
+yms_kan/tool.py,sha256=rkRpqF3EcsAq7a3k1F1zKlxfJ4U9n-FzHyNCJgN4URY,21159
+yms_kan/train_eval_utils.py,sha256=Cqw0heB7gOIK3pvOPBx0OIIWi2glfimPpyDqboFq2Tk,17186
 yms_kan/utils.py,sha256=J07L-tgmc1OfU6Tl6mGwHJRizjFN75EJK8BxejaZLUc,23860
-yms_kan/version.py,sha256=qeSnHAh3t9Zb2L0FPUF5OaQvWEJcfTki6FmrfynjWz4,39
+yms_kan/version.py,sha256=eECtaVYZj2CuGnsLuv9pAmxQhOb0PZcTisjYg4JgF5c,39
 yms_kan/assets/img/mult_symbol.png,sha256=2f4xUKdweft-qUbHjFI5h9-smnEtc0FWq8hNYZhPAXY,6392
 yms_kan/assets/img/sum_symbol.png,sha256=94QkMUzmEjlCq_yf14nMEQmettaq86FmlGfdl22b4XE,6210
-yms_kan-0.0.2.dist-info/licenses/LICENSE,sha256=BJXDWyF4Groqtnp4Gi9puH4aLg7A2IC3MpHmC-cSxwc,1067
-yms_kan-0.0.2.dist-info/METADATA,sha256=jTD-nNMWFF64GiFO2-bYQePNxJh4J1-yi4eniWT1djQ,240
-yms_kan-0.0.2.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
-yms_kan-0.0.2.dist-info/top_level.txt,sha256=Z_JDh6yZf-EiW1eKgL6ADsN2yqEMRMspi-o29JZ1WPo,8
-yms_kan-0.0.2.dist-info/RECORD,,
+yms_kan-0.0.4.dist-info/licenses/LICENSE,sha256=BJXDWyF4Groqtnp4Gi9puH4aLg7A2IC3MpHmC-cSxwc,1067
+yms_kan-0.0.4.dist-info/METADATA,sha256=VsDT6gWg7lsWcP424XM-o-jsTQ6eAlmGQWbb-hLxGmk,240
+yms_kan-0.0.4.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
+yms_kan-0.0.4.dist-info/top_level.txt,sha256=Z_JDh6yZf-EiW1eKgL6ADsN2yqEMRMspi-o29JZ1WPo,8
+yms_kan-0.0.4.dist-info/RECORD,,