RAG MCP Tool

import os import mimetypes def is_text_file(file_path: str) -> bool: """ Check if a file is a text file based on extension and content. """ # Skip hidden files/dirs (except if explicitly handled, but usually we skip .git etc) if os.path.basename(file_path).startswith('.'): return False # Extensions to skip skip_exts = { '.jpg', '.jpeg', '.png', '.gif', '.bmp', '.tiff', '.svg', '.mp4', '.avi', '.mov', '.flv', '.mkv', '.mp3', '.wav', '.flac', '.aac', '.zip', '.rar', '.tar', '.gz', '.7z', '.exe', '.dll', '.sh', '.bat', '.apk', '.pdf', '.docx', '.xlsx', '.bin', '.pyc' } ext = os.path.splitext(file_path)[1].lower() if ext in skip_exts: return False # Extensions to include text_exts = { '.txt', '.md', '.json', '.yaml', '.xml', '.csv', '.log', '.ini', '.conf', '.py', '.js', '.html', '.css' } if ext in text_exts: return True # Heuristic check for other files try: with open(file_path, 'r', encoding='utf-8') as f: f.read(1024) return True except UnicodeDecodeError: return False except Exception: return False def read_file_content(file_path: str) -> str: try: with open(file_path, 'r', encoding='utf-8') as f: return f.read() except Exception as e: print(f"Error reading {file_path}: {e}") return "" def chunk_text(text: str, chunk_count: int) -> list[str]: """ Split text into approximately `chunk_count` parts, respecting boundaries. """ if not text: return [] if chunk_count <= 1: return [text] total_len = len(text) target_size = total_len // chunk_count if target_size == 0: return [c for c in text] # Very small text chunks = [] current_pos = 0 # Simple recursive splitter logic simulation # We want to find a split point near current_pos + target_size for _ in range(chunk_count - 1): if current_pos >= total_len: break search_start = current_pos + target_size # Look for a good split point around search_start # Priorities: \n\n, \n, space split_point = -1 # Search window: +/- 20% of target size? Or just look forward/backward # Let's look forward for the nearest newline # Try to find \n in the next chunk # We want to split roughly at `current_pos + target_size` candidate = min(current_pos + target_size, total_len) # If we are at the end, just take the rest if candidate == total_len: chunks.append(text[current_pos:]) current_pos = total_len break # Look for \n around candidate # Search range: [candidate - target_size/2, candidate + target_size/2] # But we must advance. # Let's just find the nearest newline after candidate, or before if it's too far. # Simple approach: Find last \n before candidate + margin? # Let's use a simpler approach: # Just split at `target_size` but back off to nearest newline. end = min(current_pos + target_size, total_len) # Try to extend to next newline if it's close next_newline = text.find('\n', end) prev_newline = text.rfind('\n', current_pos, end) if next_newline != -1 and next_newline - end < 100: split_point = next_newline + 1 elif prev_newline != -1 and end - prev_newline < 100: split_point = prev_newline + 1 else: # Try space next_space = text.find(' ', end) if next_space != -1 and next_space - end < 50: split_point = next_space + 1 else: split_point = end chunks.append(text[current_pos:split_point]) current_pos = split_point # Last chunk if current_pos < total_len: chunks.append(text[current_pos:]) return chunks