aare/python/aare/transform.py

# SPDX-License-Identifier: MPL-2.0
import numpy as np
from . import _aare
from aare._aare import ReadoutMode 
from aare._aare import Matterhorn10 

class AdcSar04Transform64to16:
    def __call__(self, data):
        return _aare.adc_sar_04_decode64to16(data)

class AdcSar05Transform64to16:
    def __call__(self, data):
        return _aare.adc_sar_05_decode64to16(data)
    
class AdcSar05060708Transform64to16:
    def __call__(self, data):
        return _aare.adc_sar_05_06_07_08decode64to16(data)

class Moench05Transform:
    """
    Transforms Moench05 chip data from a buffer of bytes (uint8_t)
    to a numpy array of uint16. Assumes data taken with analog samples and assumes adc 1, 9, 13 are enabled. 
    (e.g. for 10g mode adc 0,1,2,3 and 8,9,10,11 and 12,13,14,15 are enabled but only adc 1,9,13 contain relevant data)  

    .. note:: 
        A moench05 chip has 160 rows and 50 cols per adc and has dynamic range 16 bit. Each adc sample is encoded in 16 bits. 
        The transformation thus requires 160*50*16/16 = 8000 analog samples per adc. 
    """
    #Could be moved to C++ without changing the interface
    def __init__(self):
        self.pixel_map = _aare.GenerateMoench05PixelMap()

    def __call__(self, data):
        return np.take(data.view(np.uint16), self.pixel_map)
    

class Moench03Transform: 
    def __init__(self): 
        self.pixel_map = _aare.GenerateMoench03PixelMap()

    def __call__(self, data):
        return np.take(data.view(np.uint16), self.pixel_map)
    

class Moench05Transform1g:
    #Could be moved to C++ without changing the interface
    def __init__(self):
        self.pixel_map = _aare.GenerateMoench05PixelMap1g()

    def __call__(self, data):
        return np.take(data.view(np.uint16), self.pixel_map)
    

class Moench05TransformOld:
    #Could be moved to C++ without changing the interface
    def __init__(self):
        self.pixel_map = _aare.GenerateMoench05PixelMapOld()

    def __call__(self, data):
        return np.take(data.view(np.uint16), self.pixel_map)
    
class Moench04AnalogTransform:
    #Could be moved to C++ without changing the interface
    def __init__(self):
        self.pixel_map = _aare.GenerateMoench04AnalogPixelMap()

    def __call__(self, data):
        return np.take(data.view(np.uint16), self.pixel_map)
    
class Matterhorn02TransceiverTransform: 
    #Could be moved to C++ without changing the interface
    def __init__(self):
        self.pixel_map = _aare.GenerateMH02SingleCounterPixelMap()

    def __call__(self, data):
        return np.take(data.view(np.uint16), self.pixel_map)

class Matterhorn10Transform:
    """
    Transforms Matterhorn10 chip data from a buffer of bytes (uint8_t)
    to a numpy array of uint8, uint16 depending on dynamic range.
    Assumes data taken with transceiver samples only.

    :param dynamic_range: How many bits a pixel is encoded dynamic range (4, 8, or 16)
    :type dynamic_range: int
    :param num_counters: num counters used (1 to 4)
    :type num_counters: int

    .. note::
        A matterhorn chip has 256 columns and 256 rows.
        A matterhornchip with dynamic range 16 and 2 counters thus requires 
        256*256*16*2/(2*64) = 1024 transceiver samples. (Per default 2 channels are enabled per transceiver sample, each channel storing 64 bits)

    .. note:: 
        Due to an artefact in the chip, the transformation only fully supports 2 or 4 counters. Also if you enable 2 counters you can only select counter 1 and 2 or 0, 3 to get reasonable results. 
        Otherwise only the first half of the image is correct. 
    """
    def __init__(self, dynamic_range : int, num_counters : int):
        self.pixel_map = _aare.GenerateMatterhorn10PixelMap(dynamic_range, num_counters)
        self.dynamic_range = dynamic_range
        self.num_counters = num_counters 

    def compatibility(self, readingmode : ReadoutMode):
        """
        checks if Matterhorn10Transform is compatible with given parameters

        :param readingmode: Reading mode set
        :type readingmode: ReadoutMode
        :raises ValueError: if not compatible
        """
        if(readingmode != ReadoutMode.TRANSCEIVER_ONLY): 
            raise ValueError(f"Incompatible Transformation. Matterhorn10Transform only requires transceiver samples. However reading mode is  {readingmode}.")
        
        pass

    def data_compatibility(self, data):
        """
        checks if data is compatible for transformation 
        
        :param data: data to be transformed, expected to be a 1D numpy array of uint8
        :type data: np.ndarray(n_counters, n_rows, n_cols) 
        :raises ValueError: if not compatible
        """
        expected_size = (Matterhorn10.nRows*Matterhorn10.nCols*self.num_counters*self.dynamic_range)//8 # read_frame returns data in uint8_t 

        if(data.size != expected_size): 
            raise ValueError(f"Data size {data.size} does not match expected size {expected_size} for Matterhorn10 with dynamic range {self.dynamic_range} and num_counters {self.num_counters}.")
        
        pass
        
    def __call__(self, data):
        self.data_compatibility(data)
        if self.dynamic_range == 16:
            return np.take(data.view(np.uint16), self.pixel_map).reshape(self.num_counters, Matterhorn10.nRows, Matterhorn10.nCols)
        elif self.dynamic_range == 8: 
            return np.take(data.view(np.uint8), self.pixel_map).reshape(self.num_counters, Matterhorn10.nRows, Matterhorn10.nCols)
        else: #dynamic range 4
            return np.take(_aare.expand4to8bit(data.view(np.uint8)), self.pixel_map).reshape(self.num_counters, Matterhorn10.nRows, Matterhorn10.nCols)
            
class Mythen302Transform:
    """
    Transform Mythen 302 test chip data from a buffer of bytes (uint8_t)
    to a uint32 numpy array of [64,3] representing channels and counters.
    Assumes data taken with rx_dbitlist 17 6, rx_dbitreorder 1 and Digital
    Samples = 2310 [(64x3x24)/2 + some extra]

    .. note::

        The offset is in number of bits 0-7

    """
    _n_channels = 64
    _n_counters = 3

    def __init__(self, offset=4):
        self.offset = offset

    def __call__(self, data : np.ndarray):
        """
        Transform buffer of data to a [64,3] np.ndarray of uint32. 

        Parameters
        ----------
        data : np.ndarray
            Expected dtype: uint8

        Returns
        ----------
        image : np.ndarray
            uint32 array of size  64, 3
        """
        res = _aare.decode_my302(data, self.offset)
        res = res.reshape(
            Mythen302Transform._n_channels, Mythen302Transform._n_counters
        )
        return res

#on import generate the pixel maps to avoid doing it every time
moench05 = Moench05Transform()
moench05_1g = Moench05Transform1g()
moench05_old = Moench05TransformOld()
matterhorn02 = Matterhorn02TransceiverTransform()
adc_sar_04_64to16 = AdcSar04Transform64to16()
adc_sar_05_64to16 = AdcSar05Transform64to16()
adc_sar_05_06_07_08_64to16 = AdcSar05060708Transform64to16()