* Added test_frelon_spectroscopy_lima_tango.py:

  + Precise readout time measurement of the different HW RoI modes
    and comparation with the theoritical value reported by the camera
  + Configure the Kinetics RoI mode for fast Spectroscopy measurements
  + Stripe concatenation readout using 3D data from readImageSeq

test/test_frelon_spectroscopy_lima_tango.py

+import time
+import PyTango
+import numpy as np
+import struct
+
+domain = 'isg'
+server = 'frelon1'
+limaccds_name = '%s/limaccds/%s' % (domain, server)
+limafrelon_name = '%s/frelon/%s' % (domain, server)
+
+vert_bin = 64
+input_channel = None
+throw_lines = 5
+
+exp_time = 1.1e-6
+slow_nb_frames = 10
+fast_nb_frames = 1000
+
+limaccds = PyTango.DeviceProxy(limaccds_name)
+limafrelon = PyTango.DeviceProxy(limafrelon_name)
+
+def get_serial_chr_delay():
+    # CR-LF terminator is not mandatory at the end of command
+    # It is added here for correctness in the real serial string sent
+    cmd = 'VER?\r\n'  
+    t0 = time.time()
+    ans = limafrelon.execSerialCommand(cmd)
+    t1 = time.time()
+    return (t1 - t0) / len(cmd + ans)
+
+serial_chr_delay = get_serial_chr_delay()
+
+def get_cam_readout_time():
+    cmd = 'TRD?'
+    ans = limafrelon.execSerialCommand(cmd)
+    return float(ans.strip()[4:]) * 1e-6
+
+def get_espia_xfer_time():
+    dtype, depth, width, height = limaccds.image_sizes
+    image_size_mb = float(width * height * depth) / 1024**2
+    xfer_speed_mb = 180
+    return image_size_mb / xfer_speed_mb
+    
+def bench_acq(exp_time, nb_frames):
+    limaccds.acq_expo_time = exp_time
+    limaccds.acq_nb_frames = nb_frames
+    height = limaccds.image_sizes[3]
+    if height > 1:
+        limaccds.acq_mode = 'SINGLE'
+    else:
+        limaccds.acq_mode = 'CONCATENATION'
+        limaccds.concat_nb_frames = nb_frames
+    limaccds.prepareAcq()
+    t0 = time.time()
+    limaccds.startAcq()
+    while limaccds.acq_status == 'Running':
+        pass
+    t1 = time.time()
+    theo_readout = get_cam_readout_time()
+    start_cmd = '>S\r\n'
+    start_cmd_delay = len(start_cmd) * serial_chr_delay
+    espia_xfer_time = get_espia_xfer_time()
+    frame_time = (t1 - t0 - start_cmd_delay - espia_xfer_time) / nb_frames
+    meas_readout = frame_time - exp_time
+    print '  Theory: %.2f ms, Measure: %.2f ms' % (theo_readout / 1e-3,
+                                                   meas_readout / 1e-3)
+    
+def get_throw_lines():
+    cmd = 'RLB?'
+    ans = limafrelon.execSerialCommand(cmd)
+    return int(ans.strip()[4:])
+
+def decode_data_array(s):
+    pack_str = '<IHHIIHHHHHHHHIIIIIIII'
+    header_len = struct.calcsize(pack_str)
+    u = struct.unpack(pack_str, s[:header_len])
+    magic = u[0]				# 4 bytes I - magic number
+    version = u[1]     				# 2 bytes H - version
+    header_header_len = u[2]			# 2 bytes H - this header length
+    category = u[3]    				# 4 bytes I - category (enum)
+    da_type = u[4]				# 4 bytes I - data type (enum)
+    endian = u[5]      				# 2 bytes H - endianness
+    nb_dims = u[6]     				# 2 bytes H - nb of dims
+    dim = u[7:13]				# 12 bytes H x 6 - dims
+    step = u[13:19]				# 24 bytes I x 6 - stepsbytes
+    pad = u[19:21]   				# padding 2 x 4 bytes
+    if magic != 0x44544159:
+        raise ValueError, 'Invalid magic 0x%x' % magic
+    if version < 2:
+        print 'Warning: using an old version: %d (expected %d)' % (version, 2)
+    DataArrayNumpyType = [
+        'u1', 'u2', 'u4', 'u8', 'i1', 'i2', 'i4', 'i8', 'f4', 'f8',
+    ]
+    endian_prefix = '>' if endian else '<'
+    np_type = endian_prefix + DataArrayNumpyType[da_type]
+    data = np.fromstring(s[header_header_len:], dtype=np.dtype(np_type))
+    shape = [dim[i] for i in xrange(nb_dims - 1, -1, -1)]
+    return data.reshape(shape)
+    
+def read_concat_stripes():
+    buffer_frames = limaccds.concat_nb_frames
+    codec, s = limaccds.readImageSeq(np.array([0, buffer_frames], 'u4'))
+    if codec != 'DATA_ARRAY':
+        raise RuntimeError, 'Unexpected DevEncoded type %s' % codec
+    data = decode_data_array(s)
+    lines, height, cols = data.shape
+    bigEndian = np.dtype(data.dtype.byteorder + 'i4') == np.dtype('>i4')
+    endianess = 'Big' if bigEndian else 'Little' 
+    print 'Read %d frames of %d columns, %d lines each [%s-Endian]' % \
+          (lines, cols, height, endianess)
+
+limafrelon.input_channel = '1-2-3-4'
+print 'Full frame %s:' % limafrelon.input_channel
+limaccds.image_bin = np.array((1, 1))
+limaccds.image_roi = np.array((0, 0, 0, 0))
+bench_acq(exp_time, slow_nb_frames)
+
+dtype, depth, width, height = limaccds.image_sizes
+binned_size = height / vert_bin
+max_line = binned_size - 1
+
+def run_stripe(input_channel):
+    line = 1
+    if input_channel == '3-4':
+        line = max_line - line
+    
+    limafrelon.roi_mode = 'SLOW'
+    bin = np.array((1, vert_bin))
+    roi = np.array((0, line, width, 1))
+    limaccds.image_bin = bin
+    limaccds.image_roi = roi
+    limafrelon.input_channel = input_channel
+    mode = 'Stripe %s %s' % (limafrelon.input_channel, limafrelon.roi_mode)
+    print '%s:' % mode
+    bench_acq(exp_time, slow_nb_frames)
+    
+    limafrelon.roi_mode = 'FAST'
+    mode = 'Stripe %s %s' % (limafrelon.input_channel, limafrelon.roi_mode)
+    print '%s:' % mode
+    bench_acq(exp_time, slow_nb_frames)
+    
+    limafrelon.roi_mode = 'KINETIC'
+    limafrelon.roi_bin_offset = 0
+    real_thrown = get_throw_lines()
+    mode = 'Stripe %s %s' % (limafrelon.input_channel, limafrelon.roi_mode)
+    print '%s - [%s lines thrown]:' % (mode, real_thrown)
+    bench_acq(exp_time, fast_nb_frames)
+    
+    if input_channel == '1-2':
+        roi = np.array((0, 0, width, 1))
+        limaccds.image_roi = roi
+        limafrelon.roi_bin_offset = throw_lines
+    else:
+        limafrelon.roi_bin_offset = vert_bin - throw_lines
+    
+    real_thrown = get_throw_lines()
+    mode = 'Stripe %s %s' % (limafrelon.input_channel, limafrelon.roi_mode)
+    print '%s - [%s lines thrown]:' % (mode, real_thrown)
+    bench_acq(exp_time, fast_nb_frames)
+
+    read_concat_stripes()
+
+for input_channel in ['1-2', '3-4']:
+    run_stripe(input_channel)
+