OCR MCP Service

"""Prompt template loader and parser.""" import re from pathlib import Path from typing import Dict, List, Optional import importlib.resources def get_prompt_doc_path() -> Optional[Path]: """Get the path to the prompt documentation file.""" # Try to find the document in multiple locations # 1. In the installed package (if packaged) try: # Try to access via importlib.resources (Python 3.9+) with importlib.resources.files("ocr_mcp_service") as package_path: # Check if prompt_template.md exists in package root template_path = package_path.parent.parent / "prompt_template.md" if template_path.exists(): return template_path except Exception: pass # 2. In the project root (development mode) # Get the package directory try: package_path = Path(__file__).parent.parent.parent template_path = package_path / "prompt_template.md" if template_path.exists(): return template_path except Exception: pass # 3. Try relative to current file current_file = Path(__file__) project_root = current_file.parent.parent.parent.parent template_path = project_root / "prompt_template.md" if template_path.exists(): return template_path return None def load_prompt_doc() -> str: """Load the prompt documentation file.""" doc_path = get_prompt_doc_path() if doc_path is None: raise FileNotFoundError( "无法找到Prompt文档。请确保文档已正确安装或位于prompt_template.md" ) try: with open(doc_path, "r", encoding="utf-8") as f: return f.read() except Exception as e: raise IOError(f"无法读取Prompt文档: {e}") def parse_scenarios(doc_content: str) -> Dict[str, str]: """Parse scenarios from the documentation.""" scenarios = {} # Pattern to match scenario sections # Look for "#### 场景 X：" or "### 场景 X：" followed by scenario name # Examples: "#### 场景 1：快速识别" or "#### 场景 2：文档页面分析" scenario_pattern = re.compile( r'^####\s+场景\s+\d+[：:]\s*(.+?)$', re.MULTILINE ) # Find all scenario headers matches = list(scenario_pattern.finditer(doc_content)) for i, match in enumerate(matches): scenario_name = match.group(1).strip() start_pos = match.end() # Find the end of this scenario (next scenario or next major section) if i + 1 < len(matches): end_pos = matches[i + 1].start() else: # Last scenario, find next major section (## or ###) next_section = re.search(r'^##+\s+', doc_content[start_pos:], re.MULTILINE) if next_section: end_pos = start_pos + next_section.start() else: end_pos = len(doc_content) scenario_content = doc_content[start_pos:end_pos].strip() # Extract the code block if exists (look for code blocks with or without language) code_block_pattern = re.compile(r'```(?:\w+)?\n(.*?)```', re.DOTALL) code_match = code_block_pattern.search(scenario_content) if code_match: scenarios[scenario_name] = code_match.group(1).strip() else: # If no code block, use the full content scenarios[scenario_name] = scenario_content return scenarios def get_scenario_names() -> List[str]: """Get list of available scenario names.""" return ["通用模板"] def get_template_file_path() -> Optional[Path]: """Get the path to the prompt template file.""" # Try multiple locations try: # 1. In the installed package with importlib.resources.files("ocr_mcp_service") as package_path: template_path = package_path.parent.parent / "prompt_template.md" if template_path.exists(): return template_path except Exception: pass # 2. In the project root (development mode) try: package_path = Path(__file__).parent.parent.parent template_path = package_path / "prompt_template.md" if template_path.exists(): return template_path except Exception: pass # 3. Try relative to current file try: current_file = Path(__file__) project_root = current_file.parent.parent.parent.parent template_path = project_root / "prompt_template.md" if template_path.exists(): return template_path except Exception: pass return None def get_scenario_template(scenario_name: str = None) -> str: """Get the general prompt template from file.""" template_path = get_template_file_path() if template_path is None: raise FileNotFoundError( "无法找到Prompt模板文件。请确保文件已正确安装或位于prompt_template.md" ) try: with open(template_path, "r", encoding="utf-8") as f: content = f.read() # Try to extract template section if file contains documentation # Look for "## 📋 模板内容" or "## 模板内容" section template_section_pattern = re.compile( r'##+\s*[📋]*\s*模板内容\s*\n(.*?)(?=\n---|\n##|$)', re.DOTALL | re.IGNORECASE ) match = template_section_pattern.search(content) if match: # Extract template content from the section template_content = match.group(1).strip() if template_content: return template_content # If no section found, check if content starts with template directly # (for backward compatibility with simple template files) if content.strip().startswith("请分析这张图片") or content.strip().startswith("# 请分析"): # Extract content before first "---" separator parts = content.split("---", 1) if parts and parts[0].strip(): return parts[0].strip() # Return entire content if no pattern matches content_stripped = content.strip() if not content_stripped: raise ValueError("Prompt模板文件为空") return content_stripped except FileNotFoundError: raise except Exception as e: raise IOError(f"无法读取Prompt模板文件: {e}") def get_full_document() -> str: """Get the full prompt documentation.""" return load_prompt_doc()