""" ASCII 2D Loader """ from sas.sascalc.dataloader.data_info import Data2D from sas.sascalc.file_converter.nxcansas_writer import NXcanSASWriter import numpy as np # ISIS 2D ASCII File Format # http://www.isis.stfc.ac.uk/instruments/loq/software/colette-ascii-file-format-descriptions9808.pdf # line: property # 0: File header # 1: q_x axis label and units # 2: q_y axis label and units # 3: Intensity axis label and units # 4: nUseRec - number of lines of user content following this line # 5 - 5+nUseRec: user content # Number of qx points # List of qx points # Number of qy points # List of qy points # numqx numqy scale class ASCII2DLoader(object): def __init__(self, data_path): """ :param data_path: The path to the file to load """ self.data_path = data_path def load(self): """ Load the data from the file into a Data2D object :return: A Data2D instance containing data from the file :raises ValueError: Raises a ValueError if the file is incorrectly formatted """ file_handle = open(self.data_path, 'r') file_buffer = file_handle.read() all_lines = file_buffer.splitlines() # Load num_points line-by-line from lines into a numpy array, starting # on line number start_line def _load_points(lines, start_line, num_points): qs = np.zeros(num_points) n = start_line filled = 0 while filled < num_points: row = np.fromstring(lines[n], dtype=np.float32, sep=' ') qs[filled:filled+len(row)] = row filled += len(row) n += 1 return n, qs current_line = 4 try: # Skip nUseRec lines nUseRec = int(all_lines[current_line].strip()[0]) current_line += nUseRec + 1 # Read qx data num_qs = int(all_lines[current_line].strip()) current_line += 1 current_line, qx = _load_points(all_lines, current_line, num_qs) # Read qy data num_qs = int(all_lines[current_line].strip()) current_line += 1 current_line, qy = _load_points(all_lines, current_line, num_qs) except ValueError as e: err_msg = "File incorrectly formatted.\n" if str(e).find('broadcast') != -1: err_msg += "Incorrect number of q data points provided. " err_msg += "Expected {}.".format(num_qs) elif str(e).find('invalid literal') != -1: err_msg += "Expected integer on line {}. Instead got '{}'".format(current_line + 1, all_lines[current_line]) else: err_msg += str(e) raise ValueError(err_msg) # dimensions: [width, height, scale] try: dimensions = np.fromstring(all_lines[current_line], dtype=np.float32, sep=' ') if len(dimensions) != 3: raise ValueError() width = int(dimensions[0]) height = int(dimensions[1]) except ValueError as e: err_msg = "File incorrectly formatted.\n" err_msg += "Expected line {} to be of the form: .".format(current_line + 1) err_msg += " Instead got '{}'.".format(all_lines[current_line]) raise ValueError(err_msg) if width > len(qx) or height > len(qy): err_msg = "File incorrectly formatted.\n" err_msg += ("Line {} says to use {}x{} points. " "Only {}x{} provided.").format(current_line + 1, width, height, len(qx), len(qy)) raise ValueError(err_msg) # More qx and/or qy points can be provided than are actually used qx = qx[:width] qy = qy[:height] current_line += 1 # iflag = 1 => Only intensity data (not dealt with here) # iflag = 2 => q axis and intensity data # iflag = 3 => q axis, intensity and error data try: iflag = int(all_lines[current_line].strip()[0]) if iflag <= 0 or iflag > 3: raise ValueError() except: err_msg = "File incorrectly formatted.\n" iflag = all_lines[current_line].strip()[0] err_msg += "Expected iflag on line {} to be 1, 2 or 3. Instead got '{}'.".format(current_line+1, iflag) raise ValueError(err_msg) current_line += 1 try: current_line, I = _load_points(all_lines, current_line, width*height) dI = np.zeros(width*height) # Load error data if it's provided if iflag == 3: _, dI = _load_points(all_lines, current_line, width*height) except Exception as e: err_msg = "File incorrectly formatted.\n" if str(e).find("list index") != -1: err_msg += ("Incorrect number of data points. Expected {} intensity").format(width*height) if iflag == 3: err_msg += " and error" err_msg += " points." else: err_msg += str(e) raise ValueError(err_msg) # Format data for use with Data2D qx = list(qx) * height qy = np.array([[y] * width for y in qy]).flatten() data = Data2D(qx_data=qx, qy_data=qy, data=I, err_data=dI) return data