MCP Sheet Parser

import logging from pathlib import Path from typing import Any from src.converters.cell_converter import CellConverter from src.converters.chart_converter import ChartConverter from src.converters.style_converter import StyleConverter from src.converters.table_structure_converter import TableStructureConverter from src.models.table_model import Sheet from src.utils.html_utils import compact_html, create_html_element logger = logging.getLogger(__name__) class HTMLConverter: """将 Sheet 对象转换为 HTML 文件的转换器。""" def __init__(self, compact_mode: bool = False, header_rows: int = 1, auto_detect_headers: bool = True): """ 初始化 HTML 转换器。 参数： compact_mode: 是否使用紧凑模式（减少空白）。 header_rows: 表头行数，默认 1。 auto_detect_headers: 是否自动检测多行表头。 """ self.compact_mode = compact_mode self.header_rows = header_rows self.auto_detect_headers = auto_detect_headers self.style_converter = StyleConverter() self.cell_converter = CellConverter(self.style_converter) self.table_converter = TableStructureConverter(self.cell_converter, self.style_converter) self.chart_converter = ChartConverter(self.cell_converter) def convert_to_files(self, sheets: list[Sheet], output_path: str) -> list[dict[str, Any]]: """ 批量将多个 Sheet 对象转换为多个 HTML 文件。 参数： sheets: Sheet 对象列表。 output_path: 输出文件路径模板。 返回： 转换结果信息列表。 """ results = [] output_p = Path(output_path) base_name = output_p.stem extension = output_p.suffix or '.html' for i, sheet in enumerate(sheets): if len(sheets) > 1: sheet_output_path = output_p.parent / f"{base_name}-{sheet.name}{extension}" else: sheet_output_path = output_p try: html_content = self._generate_html(sheet) sheet_output_path.parent.mkdir(parents=True, exist_ok=True) with open(sheet_output_path, 'w', encoding='utf-8') as f: f.write(html_content) file_size = sheet_output_path.stat().st_size results.append({ "status": "success", "output_path": str(sheet_output_path.absolute()), "file_size": file_size, "file_size_kb": round(file_size / 1024, 2), "sheet_name": sheet.name, "rows_converted": len(sheet.rows), "cells_converted": sum(len(row.cells) for row in sheet.rows), "has_styles": any(any(cell.style for cell in row.cells) for row in sheet.rows), "has_merged_cells": len(sheet.merged_cells) > 0 }) except Exception as e: logger.error(f"HTML conversion failed for sheet '{sheet.name}': {e}") results.append({"status": "error", "sheet_name": sheet.name, "error": str(e)}) return results def _generate_html(self, sheet: Sheet) -> str: """ 生成完整的 HTML 内容。 参数： sheet: Sheet 对象。 返回： HTML 字符串。 """ # 智能检测表头行数 effective_header_rows = self._detect_header_rows(sheet) if self.auto_detect_headers else self.header_rows # 检测是否有Excel标题行 has_excel_title = self._detect_title_row(sheet) styles = self.style_converter.collect_styles(sheet) css_content = self.style_converter.generate_css(styles, sheet) table_html = self.table_converter.generate_table(sheet, styles, effective_header_rows) overlay_charts_html = self.chart_converter.generate_overlay_charts_html(sheet) standalone_charts_html = self.chart_converter.generate_standalone_charts_html(sheet.charts) # 根据是否检测到Excel标题来决定是否包含<h1>标签 html_template = self._get_html_template(include_h1=not has_excel_title) # 修复：将图表覆盖层放在表格容器内部，确保正确的相对定位 table_with_charts = create_html_element('div', f'\n{table_html}\n{overlay_charts_html}\n', css_classes=['table-container']) html_content = html_template.format( title=f"Table: {sheet.name}", css_styles=css_content, table_content=table_with_charts, overlay_charts_content="", # 已经包含在table_content中 charts_content=standalone_charts_html, sheet_name=sheet.name ) if self.compact_mode: html_content = compact_html(html_content) return html_content def _detect_header_rows(self, sheet: Sheet) -> int: """ 智能检测表头行数（样式、内容、合并单元格等多种启发式规则）。 参数：sheet: Sheet对象 返回：检测到的表头行数 """ if not sheet.rows or len(sheet.rows) < 2: return min(1, len(sheet.rows)) # 如果只有0-1行，返回实际行数 # 默认至少有1行表头 header_rows = 1 max_check_rows = min(10, len(sheet.rows)) # 最多检查前10行 # 检查合并单元格 header_merged_cells = [] for merged_range in sheet.merged_cells: try: from src.utils.range_parser import parse_range_string start_row, _, end_row, _ = parse_range_string(merged_range) if start_row < max_check_rows: header_merged_cells.append((start_row, end_row)) except Exception: pass # 如果有跨行的合并单元格，可能是多行表头 for start_row, end_row in header_merged_cells: if end_row >= header_rows: header_rows = end_row + 1 # 检查样式差异 data_row_styles = [] header_row_styles = [] # 收集前几行的样式特征 for i in range(min(max_check_rows, len(sheet.rows))): row = sheet.rows[i] style_features = self._get_row_style_features(row) if i < header_rows: header_row_styles.append(style_features) else: data_row_styles.append(style_features) # 如果数据行样式与表头行显著不同，可能是多行表头 if data_row_styles and header_row_styles: for i in range(header_rows, min(max_check_rows, len(sheet.rows))): if self._is_header_style(sheet.rows[i], header_row_styles, data_row_styles): header_rows = i + 1 else: break return header_rows def _detect_title_row(self, sheet: Sheet) -> bool: """ 检测第一行是否为Excel标题行。 检测标准： 1. 强标题：字体大小 >= 18pt（无论其他样式） 2. 中等标题：字体大小 >= 14pt 且 (粗体 且 居中对齐) 3. 弱标题：字体大小 >= 16pt 且 (粗体 或 居中对齐) 参数： sheet: Sheet 对象 返回： bool: 如果检测到标题行返回True，否则返回False """ if not sheet.rows or not sheet.rows[0].cells: return False first_row = sheet.rows[0] # 检查第一行的所有单元格 for cell in first_row.cells: if cell.value and str(cell.value).strip() and cell.style: font_size = cell.style.font_size or 0 is_bold = cell.style.bold or False is_center = cell.style.text_align == 'center' # 强标题：字体大小 >= 18pt if font_size >= 18: return True # 中等标题：字体大小 >= 14pt 且 粗体 且 居中 if font_size >= 14 and is_bold and is_center: return True # 弱标题：字体大小 >= 16pt 且 (粗体 或 居中) if font_size >= 16 and (is_bold or is_center): return True return False def _get_row_style_features(self, row) -> dict: """提取行的样式特征。""" features = { 'bold_count': 0, 'bg_color_count': 0, 'text_count': 0, 'number_count': 0, 'empty_count': 0 } for cell in row.cells: # 检查样式 if cell.style: if cell.style.bold: features['bold_count'] += 1 if cell.style.background_color: features['bg_color_count'] += 1 # 检查内容类型 if cell.value is None or cell.value == '': features['empty_count'] += 1 elif isinstance(cell.value, (int, float)): features['number_count'] += 1 else: features['text_count'] += 1 return features def _is_header_style(self, row, header_styles, data_styles) -> bool: """判断行是否更像表头样式。""" row_features = self._get_row_style_features(row) # 计算与表头样式和数据样式的相似度 header_similarity = 0 data_similarity = 0 for header_style in header_styles: similarity = 0 for key in header_style: if header_style[key] > 0 and row_features[key] > 0: similarity += 1 header_similarity = max(header_similarity, similarity) for data_style in data_styles: similarity = 0 for key in data_style: if data_style[key] > 0 and row_features[key] > 0: similarity += 1 data_similarity = max(data_similarity, similarity) # 如果行更像表头，或者包含大量粗体/背景色，认为是表头 return (header_similarity > data_similarity or row_features['bold_count'] > len(row.cells) * 0.5 or row_features['bg_color_count'] > len(row.cells) * 0.5) def _get_html_template(self, include_h1: bool = True) -> str: """ 获取 HTML 模板字符串。 参数： include_h1: 是否包含<h1>标签，默认为True """ # 根据参数决定是否包含<h1>标签 h1_section = " <h1>{sheet_name}</h1>\n" if include_h1 else "" template = """<!DOCTYPE html> <html lang="zh-CN"> <head> <meta charset="UTF-8"> <meta name="viewport" content="width=device-width, initial-scale=1.0"> <title>{title}</title> <style> {css_styles} </style> </head> <body> """ + h1_section + """{table_content} {overlay_charts_content} {charts_content} <script> // 强制应用文字溢出样式 document.addEventListener('DOMContentLoaded', function() {{ const overflowCells = document.querySelectorAll('.text-overflow'); overflowCells.forEach(function(cell) {{ cell.style.setProperty('overflow', 'visible', 'important'); cell.style.setProperty('white-space', 'nowrap', 'important'); cell.style.setProperty('width', 'auto', 'important'); cell.style.setProperty('min-width', 'auto', 'important'); cell.style.setProperty('word-wrap', 'normal', 'important'); cell.style.setProperty('position', 'relative', 'important'); cell.style.setProperty('z-index', '5', 'important'); console.log('Applied overflow styles to:', cell.textContent.trim()); }}); }}); </script> </body> </html>""" return template