Module gonioanalysis.drosom.analysing
Main MAnalyser classes for goniometric motion analysis.
Classes
class AnalyserBase-
Common base class for all analysers.
Features
- setting vertical and horizontal ange limits
- exposing selected attributes as UI options
Attributes
ui_options:dict- Info about attributes should be exposed in a user interface. Each item is another dictionary with keys "help" (help string) and "type" (type conversion function such as int).
va_limits,ha_limits:list- Vertical angle and horizontal angle limits for get_3d_vectors.
Expand source code
class AnalyserBase: '''Common base class for all analysers. Features -------- - setting vertical and horizontal ange limits - exposing selected attributes as UI options Attributes ---------- ui_options : dict Info about attributes should be exposed in a user interface. Each item is another dictionary with keys "help" (help string) and "type" (type conversion function such as int). va_limits, ha_limits : list Vertical angle and horizontal angle limits for get_3d_vectors. ''' def __init__(self): self.ui_options = {} self.va_limits = [None, None] self.ha_limits = [None, None] self.alimits_reverse = False def get_ui_options(self): '''Lists UI options for the analyser: Their values, help and convert. ''' dictionary = [] for key in self.ui_options: dictionary[key] = {} dictionary[key]['value'] = getattr(self, key, None) dictionary[key]['help'] = self.ui_options[key]['help'] dictionary[key]['type'] = self.ui_options[key]['type'] return dictionary def set_ui_options(self, dictionary): ''' ''' for key, item in dictionary.items(): if key in self.ui_options: convert = self.ui_options[key].get('type', str) setattr(self, key, convert(item)) else: valid_keys = list(self.ui_options.keys()) raise KeyError(f'Key "{key}" not valid. Valid keys are {valid_keys}') def set_angle_limits(self, va_limits=(None, None), reverse=False): ''' Limit get_3d_vectors All units in degrees. ''' self.va_limits = va_limits self.alimits_reverse = reverseSubclasses
Methods
def get_ui_options(self)-
Lists UI options for the analyser: Their values, help and convert.
def set_angle_limits(self, va_limits=(None, None), reverse=False)-
Limit get_3d_vectors
All units in degrees.
def set_ui_options(self, dictionary)
class MAnalyser (data_path, folder, clean_tmp=False, no_data_load=False, active_analysis='')-
Motion analysis for GonioImsoft data.
Attributes
data_path:string- A full file path to the specimen folder's location.
folder:string- The name of the specimen folder.
active_analysis:string- Name of the active analysis
ROIsmovements:dict-
Nested dictionary of the measured 2D movements from Movemeter.
Nested structure
self.movements[eye][angle][i_repeat][x/y/time] eye = "left" or "right" angle = recording_name.lstrip('pos') // for example angle="(0, 0)_uv"
eyes:tupleofstrings- Default ("left", "right").
vector_rotation:floatorNone- Rotation of 2D vectors (affects 3D)
imagefolder_skiplist:dict
Initialize the MAnalyser object.
Arguments
no_data_load:bool- If True, skips loading data at constructing the object (use if needing many short lived objects.
active_analysis:string- Name of the activated analysis
Expand source code
class MAnalyser(AnalyserBase): '''Motion analysis for GonioImsoft data. Attributes ---------- data_path : string A full file path to the specimen folder's location. folder : string The name of the specimen folder. active_analysis : string Name of the active analysis ROIs movements : dict Nested dictionary of the measured 2D movements from Movemeter. Nested structure ----------------- self.movements[eye][angle][i_repeat][x/y/time] eye = "left" or "right" angle = recording_name.lstrip('pos') // for example angle="(0, 0)_uv" eyes : tuple of strings Default ("left", "right"). vector_rotation : float or None Rotation of 2D vectors (affects 3D) imagefolder_skiplist : dict ''' def __init__(self, data_path, folder, clean_tmp=False, no_data_load=False, active_analysis=''): '''Initialize the MAnalyser object. Arguments --------- no_data_load : bool If True, skips loading data at constructing the object (use if needing many short lived objects. active_analysis : string Name of the activated analysis ''' super().__init__() self._no_data_load = no_data_load self.ROIs = None self.data_path = data_path self.folder = folder # Skip image_folders. i_repeat self.imagefolder_skiplist = {} # Python dictionary for linked data self.linked_data = {} self.manalysers = [self] self.eyes = ("left", "right") self.vector_rotation = None # Different file or folder paths self._rois_skelefn = 'rois_{}{}.json' # specimen_name active_analysis self._movements_skelefn = 'movements_{}_{}{}.json' # specimen_name, eye, active_analysis self._skiplist_savefn = os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', folder, 'imagefolder_skiplist.json') self._crops_savefn = os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', folder, self._rois_skelefn.format(folder, '')) self._movements_savefn = os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', folder, self._movements_skelefn.format(folder, '{}', '')) self._link_savedir= os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', folder, 'linked_data') # Load some things if no_data_load == False: self.stacks = load_data(os.path.join(self.data_path, self.folder)) if os.path.isfile(self._skiplist_savefn): with open(self._skiplist_savefn, 'r') as fp: self.imagefolder_skiplist = json.load(fp) self.antenna_level_correction = self._getAntennaLevelCorrection(folder) self.load_linked_data() # Ensure the directories where the crops and movements are saved exist os.makedirs(os.path.dirname(self._crops_savefn), exist_ok=True) os.makedirs(os.path.dirname(self._movements_savefn), exist_ok=True) # Set the active analysis, and loads ROIs and movements if available self.active_analysis = active_analysis self.stop_now = False # No data load to affect only this constructor self._no_data_load = False # Info data about available UI options self.ui_options = { 'vector_rotation': {'help': 'Vector rotation in the 2D imaging plane', 'type': float} } @property def name(self): return self.folder @name.setter def name(self, name): self.folder=name @property def active_analysis(self): if self.__active_analysis == '': return 'default' else: return self.__active_analysis @active_analysis.setter def active_analysis(self, name): '''Sets the active analysis and loads ROIs an movements. Arguments --------- name : string The analysis default is "default" or "" (empty string). ''' if name == 'default': name = '' self.__active_analysis = name if name == '': self._crops_savefn = os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', self.folder, self._rois_skelefn.format(self.folder, '')) self._movements_savefn = os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', self.folder, self._movements_skelefn.format(self.folder, '{}', '')) else: self._crops_savefn = os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', self.folder, self._rois_skelefn.format(self.folder, '_'+name)) self._movements_savefn = os.path.join(PROCESSING_TEMPDIR, 'MAnalyser_data', self.folder, self._movements_skelefn.format(self.folder, '{}', '_'+name)) if self._no_data_load == False: if self.are_rois_selected(): self.load_ROIs() else: try: del self.ROIs except AttributeError: pass if self.is_measured(): self.load_analysed_movements() else: try: del self.movements except AttributeError: pass def list_analyses(self): '''Returns a list of existing analysis names. ''' manalyser_dir = os.path.dirname(self._movements_savefn) if os.path.isdir(manalyser_dir): fns = [fn for fn in os.listdir(manalyser_dir) if self._movements_skelefn.split('{')[0] in fn and self.eyes[0] in fn] else: fns = [] names = [] for fn in fns: secondlast, last = fn.split('.')[0].split('_')[-2:] if fn.split('.')[0].split('_')[-1] in self.eyes: names.append('default') else: analysis = fn.split(self.eyes[0])[1].split('.')[0].removeprefix('_') names.append(analysis) return names def __fileOpen(self, fn): with open(fn, 'r') as fp: data = json.load(fp) return data def __fileSave(self, fn, data): with open(fn, 'w') as fp: json.dump(data, fp) def mark_bad(self, image_folder, i_repeat): ''' Marks image folder and repeat to be bad and excluded when loading movements. i_repeat : int or 'all' ''' if self.imagefolder_skiplist.get(image_folder, None) is None: self.imagefolder_skiplist[image_folder] = [] self.imagefolder_skiplist[image_folder].append(i_repeat) with open(self._skiplist_savefn, 'w') as fp: json.dump(self.imagefolder_skiplist, fp) def list_rotations(self, list_special=True, special_separated=False, horizontal_condition=None, vertical_condition=None, _return_imagefolders=False): ''' List all the imaged vertical-horizontal pair rotations. Arguments --------- list_special : bool If false, include only rotations whose folders have no suffix special_separated : bool If true, return standard and special image folders separetly. horizontal_condition : callable or None A callable, that when supplied with horizontal (in steps), returns either True (includes) or False (excludes). vertical_condition : callable or None Same as horizontal condition but for vertical rotation Returns ------- rotations : list of tuples List of rotations. ''' def check_conditions(vertical, horizontal): if callable(horizontal_condition): if not horizontal_condition(horizontal): return False if callable(vertical_condition): if not vertical_condition(vertical): return False return True standard = [] special = [] for key in self.stacks.keys(): try: horizontal, vertical = ast.literal_eval(key) if check_conditions(vertical, horizontal) == False: continue except (SyntaxError, ValueError): # This is now a special folder, ie. with a suffix or something else # Try to get the angle anyhow splitted = key.replace('(', ')').split(')') if len(splitted) == 3: try: horizontal, vertical = splitted[1].replace(' ', '').split(',') horizontal = int(horizontal) vertical = int(vertical) if check_conditions(vertical, horizontal) == False: continue except: pass if _return_imagefolders: special.append('pos'+key) else: special.append((horizontal, vertical)) continue except TypeError: # Fully custom folder name special.append(key) continue if _return_imagefolders: standard.append('pos'+key) else: standard.append((horizontal, vertical)) if not list_special: special = [] if special_separated: return standard, special else: return standard + special def list_imagefolders(self, endswith='', only_measured=False, **kwargs): ''' Returns a list of the image folders (specimen subfolders that contain the images). Arguments --------- only_measured : bool Return only image_folders with completed movement analysis See list_rotations for other allowed keyword arguments. Returns ------- image_folders : list of strings ''' image_folders, special_image_folders = self.list_rotations( special_separated=True, _return_imagefolders=True, **kwargs) all_folders = [fn for fn in sorted(image_folders) + sorted(special_image_folders) if fn.endswith(endswith)] if only_measured: all_folders = [fn for fn in all_folders if self.folder_has_movements(fn)] return all_folders def get_horizontal_vertical(self, image_folder, degrees=True): ''' Tries to return the horizontal and vertical for an image folder. image_folder degrees If true, return in degrees ''' # Trusting that ( and ) only reserved for the angle splitted = key.replace('(', ')').split(')') if len(splitted) == 3: horizontal, vertical = splitted[1].replace(' ', '').split(',') horizontal = int(horizontal) vertical = int(vertical) if degrees: return step2degree(horizontal), step2degree(vertical) else: return horizontal, vertical def get_specimen_directory(self): return os.path.join(self.data_path, self.folder) def list_images(self, image_folder, absolute_path=False): ''' List all image filenames in an image folder FIXME: Alphabetical order not right because no zero padding image_folder Name of the image folder absolute_path If true, return filenames with absolute path instead of relative ''' #fns = [fn for fn in os.listdir(os.path.join(self.data_path, self.folder, image_folder)) if fn.endswith('.tiff') or fn.endswith('.tif')] #fns = arange_fns(fns) #if absolute_path: # fns = [os.path.join(self.data_path, self.folder, image_folder, fn) for fn in fns] fns = [] # Flatten out the i_repeat structure for repetitions_images in self.stacks[image_folder.removeprefix('pos')]: fns.extend(repetitions_images) if not absolute_path: fns = [os.path.basename(fn) for fn in fns] return fns def get_specimen_name(self): ''' Return the name of the data (droso) folder, such as DrosoM42 ''' return self.folder def get_imagefolder(self, image_fn): '''Gets the image containing folder (for example, /a/b/c/image -> c). Arguments --------- image_fn : string Full path to the image. ''' return os.path.split(os.path.dirname(image_fn))[1] @staticmethod def _getAntennaLevelCorrection(fly_name): fn = os.path.join(ANALYSES_SAVEDIR, 'antenna_levels', fly_name+'.txt') if os.path.exists(fn): with open(fn, 'r') as fp: antenna_level_offset = float(fp.read()) else: antenna_level_offset = False return antenna_level_offset def _correctAntennaLevel(self, angles): ''' angles In degrees, tuples, (horizontal, pitch) ''' if self.antenna_level_correction != False: for i in range(len(angles)): angles[i][1] -= self.antenna_level_correction return angles def get_antenna_level_correction(self): ''' Return the antenna level correction or if no correction exists, False. ''' return self._getAntennaLevelCorrection(self.folder) def get_imaging_parameters(self, image_folder): ''' Returns a dictionary of the Gonio Imsoft imaging parameters. The dictionary is empty if the descriptions file is missing. image_folder : string ''' parameters = {} fn = os.path.join(self.data_path, self.folder, image_folder, 'description.txt') if not os.path.isfile(fn): # Fallback for older Imsoft data where only # one descriptions file for each imaging old_fn = os.path.join(self.data_path, self.folder, self.folder+'.txt') if os.path.isfile(old_fn): fn = old_fn else: return {} with open(fn, 'r') as fp: for line in fp: if line.startswith('#') or line in ['\n', '\r\n']: continue split= line.strip('\n\r').split(' ') if len(split) >= 1: parameters[split[0]] = split[1] else: parameters[split[0]] = '' return parameters def get_specimen_age(self): ''' Returns age of the specimen, or None if unkown. If many age entries uses the latest for that specimen. ''' try: self.descriptions_file except AttributeError: self.descriptions_file = self._load_descriptions_file() for line in self.descriptions_file[::-1]: if line.startswith('age '): return line.lstrip('age ') return None def get_specimen_sex(self): ''' Returns sex of the specimen, or None if unkown. If many sex entries uses the latest for that specimen. ''' try: self.descriptions_file except AttributeError: self.descriptions_file = self._load_descriptions_file() for line in self.descriptions_file[::-1]: if line.startswith('sex '): return line.lstrip('sex ').strip(' ').strip('\n') return None def get_imaging_frequency(self, image_folder, fallback_value=100.): ''' Return imaging frequency (how many images per second) for an image folder by searching for frame_length field in the descriptions file. Arguments --------- image_folder : string or None Name. If None (not recommended), returns the imaging frequency from any image folder that is available (~random). fallback_value : float or None What to return if fs cannot be determined. Returns ------- fs : float The imaging frequency. Default of 100 HzNone if the imaging frequency could not be determined. ''' if image_folder is None: folders = self.list_imagefolders() else: folders = [image_folder] for folder in folders: fs = self.get_imaging_parameters(folder).get('frame_length', None) if fs is not None: break if fs is None: # FIXME return fallback_value else: return 1/float(fs) def get_pixel_size(self, image_folder): ''' Return the pixel size of the imaging. Currently always returns the same static value of 1.22375. ''' # Based on the stage micrometer; # 0.8 µm in the images 979 pixels return 1/1.22376 def get_rotstep_size(self): ''' Returns how many degrees one rotation encoder step was (the return value * steps == rotation in degrees) ''' return 360/DEFAULT_STEPS_PER_REVOLUTION def get_snap_fn(self, i_snap=0, absolute_path=True): ''' Returns the first snap image filename taken (or i_snap'th if specified). Many time I took a snap image of the fly at (0,0) horizontal/vertical, so this can be used as the "face photo" of the fly. ''' snapdir = os.path.join(self.data_path, self.folder, 'snaps') fns = [fn for fn in os.listdir(snapdir) if fn.endswith('.tiff')] fns.sort() if absolute_path: fns = [os.path.join(snapdir, fn) for fn in fns] return fns[i_snap] def load_ROIs(self): ''' Load ROIs (selected before) for the left/right eye. INPUT ARGUMENTS DESCRIPTION eye 'left' or 'right' DETAILS While selecting ROIs, both eyes are selcted simultaneously. There's no explicit information about from which eye each selected ROI is from. Here we reconstruct the distinction to left/right using following way: 1 ROI: horizontal angle determines 2 ROIs: being left/right in the image determines Notice that this means that when the horizontal angle is zero (fly is facing towards the camera), image rotation has to be so that the eyes are on image's left and right halves. ''' self.ROIs = {'left': {}, 'right': {}} with open(self._crops_savefn, 'r') as fp: marker_markings = json.load(fp) for image_fn, ROIs in marker_markings.items(): # Since use to relative filenames in the ROIs savefile image_fn = os.path.join(self.data_path, self.folder, image_fn) # ROIs smaller than 7 pixels a side are not loaded good_rois = [] for i_roi in range(len(ROIs)): if not (ROIs[i_roi][2] < 7 and ROIs[i_roi][3] < 7): good_rois.append(ROIs[i_roi]) ROIs = good_rois pos = self.get_imagefolder(image_fn) try: horizontal, pitch = angles_from_fn(pos) except: horizontal, pitch = (0, 0) pos = pos.removeprefix('pos') if '_cam' in os.path.basename(image_fn): # Allows cameras 0-9 try: i_camera = int(image_fn[image_fn.index('_cam') + 4]) pos += f'_cam{i_camera}' except: pass # ROI belonging to the eft/right eye is determined solely by # the horizontal angle when only 1 ROI exists for the position if len(ROIs) == 1: if horizontal > 0: self.ROIs['left'][pos] = ROIs[0] else: self.ROIs['right'][pos] = ROIs[0] # If there's two ROIs elif len(ROIs) == 2: if ROIs[0][0] > ROIs[1][0]: self.ROIs['left'][pos] = ROIs[0] self.ROIs['right'][pos] = ROIs[1] else: self.ROIs['left'][pos]= ROIs[1] self.ROIs['right'][pos] = ROIs[0] elif len(ROIs) > 2: # With drift_correction, many ROIs are possible #if self.active_analysis == 'drift_correction': # for i_roi, roi in enumerate(ROIs): # self.ROIs[f'roi{i_roi}'] = roi print('Warning. len(ROIs) == {} for {}'.format(len(ROIs), image_fn)) self.N_folders_having_rois = len(marker_markings) def select_ROIs(self, **kwargs): ''' Selecting the ROIs from the loaded images. Currently, only the first frame of each recording is shown. kwargs Passed to the marker constructor ''' to_cropping = [stacks[0][0] for str_angles, stacks in self.stacks.items()] fig, ax = plt.subplots() marker = Marker(fig, ax, to_cropping, self._crops_savefn, relative_fns_from=os.path.join(self.data_path, self.folder), **kwargs) marker.run() def are_rois_selected(self): ''' Returns True if a file for crops/ROIs is found. ''' return os.path.exists(self._crops_savefn) def count_roi_selected_folders(self): ''' Returns the number of imagefolders that have ROIs selected ''' if self.are_rois_selected(): return self.N_folders_having_rois else: return 0 def folder_has_rois(self, image_folder): ''' Returns True if for specified image_folder at least one ROI exsits. Otherwise False. ''' try: self.ROIs except AttributeError: return False if self.get_rois(image_folder) != []: return True return False def get_rois(self, image_folder): rois = [] for eye in ['left', 'right']: try: roi = self.ROIs[eye][image_folder.removeprefix('pos')] rois.append(roi) except: continue return rois def is_measured(self): ''' Returns (True, True) if analyseMovement results can be found for the fly and bot eyes. ''' return all((os.path.exists(self._movements_savefn.format('left')), os.path.exists(self._movements_savefn.format('right')))) def folder_has_movements(self, image_folder): ''' Returns True if for specified image_folder has movements measured. Otherwise False. ''' try: self.movements except AttributeError: return False if any([image_folder.removeprefix('pos') in self.movements[eye].keys()] for eye in ['left', 'right']): if len(self.get_displacements_from_folder(image_folder)) > 0: return True return False def load_analysed_movements(self): self.movements = {} with open(self._movements_savefn.format('right'), 'r') as fp: self.movements['right'] = json.load(fp) with open(self._movements_savefn.format('left'), 'r') as fp: self.movements['left'] = json.load(fp) if self.__active_analysis == 'drift_correction': for key in self.movements['right']: if key not in self.movements['left']: self.movements['left'] = self.movements['right'] # Special analysis name: "drift_correction". If present, subtracted # from all the active analysis if 'drift_correction' in self.list_analyses() and self.__active_analysis != 'drift_correction': dc_analyser = MAnalyser( self.data_path, self.folder, active_analysis='drift_correction') for eye in ['left', 'right']: for image_folder, data in dc_analyser.movements[eye].items(): if image_folder not in self.movements[eye]: continue for i_repeat in range(len(data)): A = self.movements[eye][image_folder][i_repeat] B = data[i_repeat] for key in ['x', 'y']: dcx = np.array(list(range(len(B[key])))) fit = np.polynomial.polynomial.Polynomial.fit(dcx, B[key],1) corrected = (np.array(A[key]) - fit(dcx) ).tolist() #corrected = [a-b for a, b in zip(A[key], B[key])] self.movements[eye][image_folder][i_repeat][key] = corrected if self.imagefolder_skiplist: for image_folder, skip_repeats in self.imagefolder_skiplist.items(): for eye in self.eyes: if self.movements[eye].get(image_folder.removeprefix('pos'), None) is None: continue # Iterate repeats reversed so we can just pop things for i_repeat in sorted(skip_repeats)[::-1]: self.movements[eye][image_folder.removeprefix('pos')].pop(i_repeat) def measure_both_eyes(self, **kwargs): ''' Wrapper to self.measure_movement() for both left and right eyes. ''' for eye in self.ROIs: self.measure_movement(eye, **kwargs) def measure_movement(self, eye, only_folders=None, max_movement=30, absolute_coordinates=False, join_repeats=False, stop_event=None): ''' Performs cross-correlation analysis for the selected ROIs (regions of interest) using Movemeter package. If ROIs haven't been selected, calls method self.selectROIs. Movements are saved into a tmp directory. INPUT ARGUMENTS DESCRIPTION eye 'left' or 'right' only_folders Analyse only image folders in the given list (that is only_folders). max_movement Maximum total displacement in x or y expected. Lower values faster. absolute_coordinates Return movement values in absolute image coordinates join_repeats Join repeats together as if they were one long recording. stop_event None or threading.Event for stopping the movement measurement Cross-correlation analysis is the slowest part of the DrosoM pipeline. ''' self.movements = {} if not os.path.exists(self._crops_savefn): self.selectROIs() self.load_ROIs() angles = [] stacks = [] ROIs = [] if not self.ROIs[eye] == {}: for angle in self.stacks: #if angle in str(self.ROIs[eye].keys()): # Continue if no ROI for this eye exists try : self.ROIs[eye][angle] except KeyError: continue # Continue if only_folders set and the angle is not in # the only folders if only_folders and not 'pos'+angle in only_folders: continue if join_repeats: fused = [] for i_repetition in range(len(self.stacks[angle])): fused += self.stacks[angle][i_repetition] self.stacks[angle] = [fused] for i_repetition in range(len(self.stacks[angle])): angles.append(angle) stacks.append( self.stacks[angle][i_repetition] ) ROIs.append( [self.ROIs[eye][angle]] ) if ROIs == []: return None # Old upscale was 4 meter = Movemeter(upscale=10, absolute_results=absolute_coordinates) meter.set_data(stacks, ROIs) for stack_i, angle in enumerate(angles): if stop_event and stop_event.is_set(): self.stop_now = True if self.stop_now: self.stop_now = False self.movements = {} print('{} EYE CANCELLED'.format(eye.upper())) return None print('Analysing {} eye, motion from position {}, done {}/{} for this eye'.format(eye.upper(), angle, stack_i+1, len(ROIs))) print("Calculating ROI's movement...") x, y = meter.measure_movement(stack_i, max_movement=max_movement)[0] print('Done.') try: self.movements[angle] except KeyError: self.movements[angle] = [] tags = meter.get_metadata(stack_i)['Image ImageDescription'].values.split('"') # GonioImsoft start time tag in the images if 'start_time' in tags: time = tags[tags.index('start_time') + 2] else: time = None self.movements[angle].append({'x': x, 'y':y, 'time': time}) else: self.movements = {} # If only_folders set ie. only some angles were (re)measured, # load previous movements also for saving if only_folders: with open(self._movements_savefn.format(eye), 'r') as fp: previous_movements = json.load(fp) # Update previous movements with the new movements and set # the updated previous movements to be the current movements previous_movements.update(self.movements) self.movements = previous_movements # Save movements with open(self._movements_savefn.format(eye), 'w') as fp: json.dump(self.movements, fp) #for key, data in self.movements.items(): # plt.plot(data['x']) # plt.plot(data['y']) # plt.show() def get_time_ordered(self, angles_in_degrees=True, first_frame_only=False, exclude_imagefolders=[]): ''' Get images, ROIs and angles, ordered in recording time for movie making. exclude_imagefolders : list Imagefolders to exclude Returns 3 lists: image_fns, ROIs, angles image_fns ''' self.load_ROIs() times_and_data = [] seen_angles = [] for eye in self.movements: for angle in self.movements[eye]: if 'pos'+angle in exclude_imagefolders: continue if not angle in seen_angles: time = self.movements[eye][angle][0]['time'] fn = self.stacks[angle][0] ROI = self.get_moving_ROIs(eye, angle) deg_angle = [list(ast.literal_eval(angle.split(')')[0]+')' ))] if angles_in_degrees: to_degrees(deg_angle) deg_angle = [deg_angle[0] for i in range(len(fn))] times_and_data.append([time, fn, ROI, deg_angle]) seen_angles.append(angle) # Everything gets sorted according to the time times_and_data.sort(key=lambda x: x[0]) image_fns = [] ROIs = [] angles = [] if not first_frame_only: for time, fns, ROI, angle in times_and_data: image_fns.extend(fns) ROIs.extend(ROI) angles.extend(angle) else: for time, fns, ROI, angle in times_and_data: image_fns.append(fns[0]) ROIs.append(ROI[0]) angles.append(angle[0]) return image_fns, ROIs, angles def get_movements_from_folder(self, image_folder): ''' ''' data = {} for eye in ['left', 'right']: try: data[eye] = self.movements[eye][image_folder.removeprefix('pos')] except KeyError: pass return data def get_displacements_from_folder(self, image_folder): ''' Returns a list of 1D numpy arrays, which give displacement over time for each repeat. If no displacement data, returns an empty list Calculated from separete (x,y) data ''' displacements = [] for eye, data in self.get_movements_from_folder(image_folder).items(): for repetition_data in data: x = repetition_data['x'] y = repetition_data['y'] mag = np.sqrt(np.asarray(x)**2 + np.asarray(y)**2) displacements.append(mag) return displacements def get_raw_xy_traces(self, eye): ''' Return angles, values angles Each recorded fly orientation in steps values X and Y ''' angles = [list(ast.literal_eval(angle)) for angle in self.movements[eye]] movement_dict = [self.movements[eye][str(angle)] for angle in angles] return angles, movement_dict def get_2d_vectors(self, eye, mirror_horizontal=True, mirror_pitch=True, correct_level=True, repeats_separately=False, mirror_movements=False): ''' Creates 2D vectors from the movements analysis data. Vector start point: ROI's position at the first frame Vector end point: ROI's position at the last frame mirror_pitch Should make so that the negative values are towards dorsal and positive towards frontal (this is how things on DrosoX were) mirror_movements : bool If True, mirrors the movement directions (X=-X and Y=-Y) ''' # Make the order of angles deterministic sorted_angle_keys = sorted(self.movements[eye]) angles = [list(ast.literal_eval(angle.split(')')[0]+')' )) for angle in sorted_angle_keys] values = [self.movements[eye][angle] for angle in sorted_angle_keys] to_degrees(angles) if correct_level: angles = self._correctAntennaLevel(angles) if mirror_horizontal: for i in range(len(angles)): angles[i][0] *= -1 xdirchange = -1 else: xdirchange = 1 if mirror_pitch: for i in range(len(angles)): angles[i][1] *= -1 # Vector X and Y components # Fix here if repetitions are needed to be averaged # (don't take only x[0] but average) if repeats_separately: tmp_angles = [] X = [] Y = [] for angle, val in zip(angles, values): for repeat in val: tmp_angles.append(angle) X.append( xdirchange*(repeat['x'][-1]-repeat['x'][0]) ) Y.append( repeat['y'][-1]-repeat['y'][0] ) angles = tmp_angles else: X = [xdirchange*(x[0]['x'][-1]-x[0]['x'][0]) for x in values] Y = [x[0]['y'][-1]-x[0]['y'][0] for x in values] if mirror_movements: X = [-x for x in X] Y = [-y for y in Y] if self.vector_rotation: r = math.radians(self.vector_rotation) for i in range(len(X)): x = X[i] y = Y[i] if angles[i][1] > 0: sr = -1 * r else: sr = 1 * r if eye == 'left': sr = sr elif eye == 'right': sr = -sr X[i] = x * math.cos(sr) - y * math.sin(sr) Y[i] = x * math.sin(sr) + y * math.cos(sr) return angles, X, Y def get_magnitude_traces(self, eye, image_folder=None, mean_repeats=False, mean_imagefolders=False, microns=False, _phase=False, _derivative=False): ''' Get all movement magnitudes (sqrt(x**2+y**2)) from the specified eye. The results are returned as a dictionary where the keys are the angle pairs (self.movements keys) eye : string or None "left" or "right". None leads to taking mean where eyes overlap image_folder : string If specified, return movements from this image folder. Otherwise by default None, movements from all image folders. mean_repeats : bool Wheter take mean of the mean repeats. mean_imagefolders : bool Only makes sense when image_folder is None mean_eyes : bool Only makes sense when eye is None microns : bool Call self.get_pixel_size(image_folder) to convert from pixel units to micrometers. _phase: bool If true return phase in degrees instead. Returns if mean_repeats == True magnitude_traces = {angle_01: [mag_mean], ...} if mean_repeats == False magnitude_traces = {angle_01: [mag_rep1, mag_rep2,...], ...} if mean_imagefolders, there's only one key 'mean' ''' alleye_magnitude_traces = {} if eye is None: eyes = self.eyes else: eyes = [eye] if image_folder is None: movement_keys = set().union(*[list(self.movements[eye].keys()) for eye in eyes]) else: movement_keys = [image_folder.removeprefix('pos')] for eye in eyes: magnitude_traces = {} for angle in movement_keys: if self.movements[eye].get(angle, None) is None: # Continue if data for this eye continue magnitude_traces[angle] = [] for i_repeat in range(len(self.movements[eye][angle])): x = self.movements[eye][angle][i_repeat]['x'] y = self.movements[eye][angle][i_repeat]['y'] if _phase: mag = np.degrees(np.arctan2(y, -np.asarray(x))) else: mag = np.sqrt(np.asarray(x)**2 + np.asarray(y)**2) if _derivative: mag = np.diff(mag) magnitude_traces[angle].append( mag ) if mean_repeats: magnitude_traces[angle] = [np.mean(magnitude_traces[angle], axis=0)] if magnitude_traces == {}: # If nothing for this eye continue if mean_imagefolders: tmp = np.mean([val for val in magnitude_traces.values()], axis=0) magnitude_traces = {} magnitude_traces['imagefoldersmean'] = tmp alleye_magnitude_traces[eye] = magnitude_traces if len(eyes) > 1: merge = {} # Merge (mean) eyes where one imagefolder hols data from both eyes angles = [list(val.keys()) for val in alleye_magnitude_traces.values()] angles = set().union(*angles) for angle in angles: data = [alleye_magnitude_traces.get(eye, {}).get(angle, None) for eye in eyes] data = [d for d in data if d is not None] merge[angle] = np.mean(data, axis=0) magnitude_traces = merge if microns and not _phase: for image_folder in magnitude_traces: pixel_size = self.get_pixel_size(image_folder) magnitude_traces[image_folder] = [t*pixel_size for t in magnitude_traces[image_folder]] return magnitude_traces def get_moving_ROIs(self, eye, angle, i_repeat=0): ''' Returns a list of ROIs how they move over time. Useful for visualizing. ''' moving_ROI = [] if not self.ROIs: self.load_ROIs() movements = self.movements[eye][angle][i_repeat] rx,ry,rw,rh = self.ROIs[eye][angle] for i in range(len(movements['x'])): x = -movements['x'][i] y = -movements['y'][i] moving_ROI.append([rx+x,ry+y,rw,rh]) return moving_ROI def get_3d_vectors(self, eye, return_angles=False, correct_level=True, repeats_separately=False, normalize_length=0.1, strict=None, vertical_hardborder=None): ''' Returns 3D vectors and their starting points. correct_level Use estimated antenna levels va_limits Vertical angle limits in degrees, (None, None) for no limits ''' angles, X, Y = self.get_2d_vectors(eye, mirror_pitch=False, mirror_horizontal=True, correct_level=False, repeats_separately=repeats_separately) N = len(angles) points = np.zeros((N,3)) vectors = np.zeros((N,3)) for i, (angle, x, y) in enumerate(zip(angles, X, Y)): horizontal, pitch = angle point0 = camera2Fly(horizontal, pitch) point1 = camvec2Fly(x, y, horizontal, pitch, normalize=normalize_length) if correct_level: rotation = -self.antenna_level_correction point0 = rotate_about_x(point0, rotation) point1 = rotate_about_x(point1, rotation) x0,y0,z0 = point0 x1,y1,z1 = point1 #vectors.append( (tuple(angle), (x0,x1), (y0,y1), (z0, z1)) ) points[i] = np.array(point0) vectors[i] = np.array(point1) - points[i] # Vertical/horizontal angle limiting booleans = where_vertical_between(points, lower=self.va_limits[0], upper=self.va_limits[1], reverse=self.alimits_reverse) points = points[booleans] vectors = vectors[booleans] if return_angles: return points, vectors, angles else: return points, vectors def get_recording_time(self, image_folder, i_rep=0): ''' Returns the timestamp of a recording, if measure_movement() method has been run for the recording. If no time is found, returns None recording_Name Recording name i_rep Return the time for recording repeat i_rep By default, i_rep=0 ''' angle = image_folder.lstrip('pos') for eye in ['left', 'right']: try: return self.movements[eye][angle][i_rep]['time'] except (KeyError, IndexError): pass except AttributeError: print('No time for {} because movements not analysed'.format(image_folder)) return None return None def stop(self): ''' Stop long running activities (now measurement). ''' self.stop_now = True # ------------ # LINKED DATA # linking external data such as ERGs to the DPP data (MAnalyser) # ------------ def link_data(self, key, data): ''' Data linked to the MAnalyser ''' self.linked_data[key] = data def save_linked_data(self): ''' Attempt saving the linked data on disk in JSON format. ''' os.makedirs(self._link_savedir, exist_ok=True) for key, data in self.linked_data.items(): with open(os.path.join(self._link_savedir, "{}.json".format(key)), 'w') as fp: json.dump(data, fp) def load_linked_data(self): ''' Load linked data from specimen datadir. ''' # Initialize linked data to an empty dict self.linked_data = {} # Check if linked data directory exsists, if not, the no linked data for this specimen if os.path.exists(self._link_savedir): dfiles = [fn for fn in os.listdir(self._link_savedir) if fn.endswith('.json')] for dfile in dfiles: with open(os.path.join(self._link_savedir, dfile), 'r') as fp: data = json.load(fp) self.linked_data[dfile.replace('.json', '')] = dataAncestors
Subclasses
Instance variables
prop active_analysis-
Expand source code
@property def active_analysis(self): if self.__active_analysis == '': return 'default' else: return self.__active_analysis prop name-
Expand source code
@property def name(self): return self.folder
Methods
def are_rois_selected(self)-
Returns True if a file for crops/ROIs is found.
def count_roi_selected_folders(self)-
Returns the number of imagefolders that have ROIs selected
def folder_has_movements(self, image_folder)-
Returns True if for specified image_folder has movements measured. Otherwise False.
def folder_has_rois(self, image_folder)-
Returns True if for specified image_folder at least one ROI exsits. Otherwise False.
def get_2d_vectors(self, eye, mirror_horizontal=True, mirror_pitch=True, correct_level=True, repeats_separately=False, mirror_movements=False)-
Creates 2D vectors from the movements analysis data. Vector start point: ROI's position at the first frame Vector end point: ROI's position at the last frame
mirror_pitch Should make so that the negative values are towards dorsal and positive towards frontal (this is how things on DrosoX were) mirror_movements : bool If True, mirrors the movement directions (X=-X and Y=-Y)
def get_3d_vectors(self, eye, return_angles=False, correct_level=True, repeats_separately=False, normalize_length=0.1, strict=None, vertical_hardborder=None)-
Returns 3D vectors and their starting points.
correct_level Use estimated antenna levels
va_limits Vertical angle limits in degrees, (None, None) for no limits
def get_antenna_level_correction(self)-
Return the antenna level correction or if no correction exists, False.
def get_displacements_from_folder(self, image_folder)-
Returns a list of 1D numpy arrays, which give displacement over time for each repeat.
If no displacement data, returns an empty list
Calculated from separete (x,y) data
def get_horizontal_vertical(self, image_folder, degrees=True)-
Tries to return the horizontal and vertical for an image folder.
image_folder degrees If true, return in degrees
def get_imagefolder(self, image_fn)-
Gets the image containing folder (for example, /a/b/c/image -> c).
Arguments
image_fn:string- Full path to the image.
def get_imaging_frequency(self, image_folder, fallback_value=100.0)-
Return imaging frequency (how many images per second) for an image folder by searching for frame_length field in the descriptions file.
Arguments
image_folder:stringorNone- Name. If None (not recommended), returns the imaging frequency from any image folder that is available (~random).
fallback_value:floatorNone- What to return if fs cannot be determined.
Returns
fs:float- The imaging frequency. Default of 100 HzNone if the imaging frequency could not be determined.
def get_imaging_parameters(self, image_folder)-
Returns a dictionary of the Gonio Imsoft imaging parameters. The dictionary is empty if the descriptions file is missing.
image_folder : string
def get_magnitude_traces(self, eye, image_folder=None, mean_repeats=False, mean_imagefolders=False, microns=False)-
Get all movement magnitudes (sqrt(x2+y2)) from the specified eye. The results are returned as a dictionary where the keys are the angle pairs (self.movements keys)
eye : string or None "left" or "right". None leads to taking mean where eyes overlap image_folder : string If specified, return movements from this image folder. Otherwise by default None, movements from all image folders. mean_repeats : bool Wheter take mean of the mean repeats. mean_imagefolders : bool Only makes sense when image_folder is None mean_eyes : bool Only makes sense when eye is None microns : bool Call self.get_pixel_size(image_folder) to convert from pixel units to micrometers. _phase: bool If true return phase in degrees instead.
Returns if mean_repeats == True magnitude_traces = {angle_01: [mag_mean], …} if mean_repeats == False magnitude_traces = {angle_01: [mag_rep1, mag_rep2,…], …}
if mean_imagefolders, there's only one key 'mean' def get_movements_from_folder(self, image_folder)def get_moving_ROIs(self, eye, angle, i_repeat=0)-
Returns a list of ROIs how they move over time. Useful for visualizing.
def get_pixel_size(self, image_folder)-
Return the pixel size of the imaging. Currently always returns the same static value of 1.22375.
def get_raw_xy_traces(self, eye)-
Return angles, values angles Each recorded fly orientation in steps values X and Y
def get_recording_time(self, image_folder, i_rep=0)-
Returns the timestamp of a recording, if measure_movement() method has been run for the recording.
If no time is found, returns None
recording_Name Recording name i_rep Return the time for recording repeat i_rep By default, i_rep=0
def get_rois(self, image_folder)def get_rotstep_size(self)-
Returns how many degrees one rotation encoder step was (the return value * steps == rotation in degrees)
def get_snap_fn(self, i_snap=0, absolute_path=True)-
Returns the first snap image filename taken (or i_snap'th if specified).
Many time I took a snap image of the fly at (0,0) horizontal/vertical, so this can be used as the "face photo" of the fly.
def get_specimen_age(self)-
Returns age of the specimen, or None if unkown.
If many age entries uses the latest for that specimen.
def get_specimen_directory(self)def get_specimen_name(self)-
Return the name of the data (droso) folder, such as DrosoM42
def get_specimen_sex(self)-
Returns sex of the specimen, or None if unkown.
If many sex entries uses the latest for that specimen.
def get_time_ordered(self, angles_in_degrees=True, first_frame_only=False, exclude_imagefolders=[])-
Get images, ROIs and angles, ordered in recording time for movie making.
exclude_imagefolders : list Imagefolders to exclude
Returns 3 lists: image_fns, ROIs, angles image_fns
def is_measured(self)-
Returns (True, True) if analyseMovement results can be found for the fly and bot eyes.
def link_data(self, key, data)-
Data linked to the MAnalyser
def list_analyses(self)-
Returns a list of existing analysis names.
def list_imagefolders(self, endswith='', only_measured=False, **kwargs)-
Returns a list of the image folders (specimen subfolders that contain the images).
Arguments
only_measured:bool- Return only image_folders with completed movement analysis
See list_rotations for other allowed keyword arguments.
Returns
image_folders:listofstrings
def list_images(self, image_folder, absolute_path=False)-
List all image filenames in an image folder
FIXME: Alphabetical order not right because no zero padding
image_folder Name of the image folder absolute_path If true, return filenames with absolute path instead of relative
def list_rotations(self, list_special=True, special_separated=False, horizontal_condition=None, vertical_condition=None)-
List all the imaged vertical-horizontal pair rotations.
Arguments
list_special:bool- If false, include only rotations whose folders have no suffix
special_separated:bool- If true, return standard and special image folders separetly.
horizontal_condition:callableorNone- A callable, that when supplied with horizontal (in steps), returns either True (includes) or False (excludes).
vertical_condition:callableorNone- Same as horizontal condition but for vertical rotation
Returns
rotations:listoftuples- List of rotations.
def load_ROIs(self)-
Load ROIs (selected before) for the left/right eye.
INPUT ARGUMENTS DESCRIPTION eye 'left' or 'right'
DETAILS While selecting ROIs, both eyes are selcted simultaneously. There's no explicit information about from which eye each selected ROI is from. Here we reconstruct the distinction to left/right using following way: 1 ROI: horizontal angle determines 2 ROIs: being left/right in the image determines
Notice that this means that when the horizontal angle is zero (fly is facing towards the camera), image rotation has to be so that the eyes are on image's left and right halves.
def load_analysed_movements(self)def load_linked_data(self)-
Load linked data from specimen datadir.
def mark_bad(self, image_folder, i_repeat)-
Marks image folder and repeat to be bad and excluded when loading movements.
i_repeat : int or 'all'
def measure_both_eyes(self, **kwargs)-
Wrapper to self.measure_movement() for both left and right eyes.
def measure_movement(self, eye, only_folders=None, max_movement=30, absolute_coordinates=False, join_repeats=False, stop_event=None)-
Performs cross-correlation analysis for the selected ROIs (regions of interest) using Movemeter package.
If ROIs haven't been selected, calls method self.selectROIs. Movements are saved into a tmp directory.
INPUT ARGUMENTS DESCRIPTION eye 'left' or 'right' only_folders Analyse only image folders in the given list (that is only_folders). max_movement Maximum total displacement in x or y expected. Lower values faster. absolute_coordinates Return movement values in absolute image coordinates join_repeats Join repeats together as if they were one long recording. stop_event None or threading.Event for stopping the movement measurement
Cross-correlation analysis is the slowest part of the DrosoM pipeline.
def save_linked_data(self)-
Attempt saving the linked data on disk in JSON format.
def select_ROIs(self, **kwargs)-
Selecting the ROIs from the loaded images. Currently, only the first frame of each recording is shown.
kwargs Passed to the marker constructor
def stop(self)-
Stop long running activities (now measurement).
Inherited members
class MAverager (manalysers)-
Combining and averaging results from many MAnalyser objects.
MAverager acts like MAnalyser object for getting data (like get_2d_vectors) but lacks the movement analysis (cross-correlation) related parts.
Expand source code
class MAverager(AnalyserBase): ''' Combining and averaging results from many MAnalyser objects. MAverager acts like MAnalyser object for getting data (like get_2d_vectors) but lacks the movement analysis (cross-correlation) related parts. ''' def __init__(self, manalysers): super().__init__() self.manalysers = manalysers self.interpolation = {} self.intp_step = (5, 5) self.eyes = manalysers[0].eyes self.vector_rotation = manalysers[0].vector_rotation self.interpolated_raw= {} # Info data about available UI options self.ui_options = manalysers[0].ui_options.copy() def get_N_specimens(self): return len(self.manalysers) def get_specimen_name(self): return 'averaged_'+'_'.join([manalyser.folder for manalyser in self.manalysers]) @property def name(self): return self.get_specimen_name() @name.setter def name(self, name): return def setInterpolationSteps(self, horizontal_step, vertical_step): ''' Set the resolution of the N-nearest neighbour interpolation in Maverager.get_2d_vectors. INPUT ARGUMENTS horizontal_step vertical_step Arguments horizontal_step and vertical_step refer to the rotation stages. ''' self.intp_step = (horizontal_step, vertical_step) def get_2d_vectors(self, eye, **kwargs): ''' Get's the 2D movement vectors (in the camera coordinate system) using N_nearest neighbour interpolation and averaging. ''' #Modified from get_3d_vectors interpolated = [[],[],[]] points_2d = [] vectors_2d = [] for analyser in self.manalysers: angles, X, Y = analyser.get_2d_vectors(eye, mirror_horizontal=False, mirror_pitch=False) vecs = [[x,y] for x,y in zip(X, Y)] points_2d.append(np.array(angles)) vectors_2d.append( np.array(vecs) ) vectors_2d = np.array(vectors_2d) kdtrees = [KDTree(points) for points in points_2d] hmin, hmax = (-90, 90) vmax, hmax = (-180, 180) intp_points = [] for h in np.arange(hmin, hmax+0.01, 10): for v in np.arange(hmin, hmax+0.1, 10): intp_points.append((h,v)) for intp_point in intp_points: nearest_vectors = [] for kdtree, vectors in zip(kdtrees, vectors_2d): distance, index = kdtree.query(intp_point) if distance < math.sqrt(self.intp_step[0]**2+self.intp_step[1]**2): nearest_vectors.append(vectors[index]) if len(nearest_vectors) > len(vectors_2d)/2: avec = np.mean(nearest_vectors, axis=0) avec /= np.linalg.norm(avec) interpolated[0].append(np.array(intp_point)) interpolated[1].append(avec[0]) interpolated[2].append(avec[1]) angles, x, y = interpolated return angles, x, y def get_3d_vectors(self, eye, correct_level=True, normalize_length=0.1, recalculate=False, strict=False, vertical_hardborder=False, repeats_separately=False, **kwargs): ''' Equivalent to MAnalysers get_3d_vectors but interpolates with N-nearest neughbours. repeats_separately : bool If True, return underlying MAnalyser vectors separetly (same points get repeated many times) ''' cachename = ';'.join([str(item) for item in [self.vector_rotation, correct_level, normalize_length, strict, vertical_hardborder]]) if self.interpolation.get(eye, {}).get(cachename) is None or recalculate: interpolated = [[],[]] self.interpolated_raw[eye] = [] # key points str, value list of vectors R = 1 intp_dist = (2 * R * np.sin(math.radians(self.intp_step[0]))) vectors_3d = [] for analyser in self.manalysers: analyser.vector_rotation = self.vector_rotation vec = analyser.get_3d_vectors(eye, correct_level=True, normalize_length=normalize_length, **kwargs) vectors_3d.append(vec) if not strict: intp_points = optimal_sampling(np.arange(-90, 90.01, self.intp_step[0]), np.arange(0, 360.01, self.intp_step[1])) else: if eye == 'left': intp_points = optimal_sampling(np.arange(-90, 0.01, self.intp_step[0]), np.arange(0, 360.01, self.intp_step[1])) else: intp_points = optimal_sampling(np.arange(0, 90.01, self.intp_step[0]), np.arange(0, 360.01, self.intp_step[1])) for intp_point in intp_points: nearest_vectors = [] for vectors in vectors_3d: i_nearest = nearest_neighbour(intp_point, vectors[0], max_distance=intp_dist) if not i_nearest is False: if vertical_hardborder: if np.sign(intp_point[2]) != np.sign(vectors[0][i_nearest][2]): continue nearest_vectors.append(vectors[1][i_nearest]) if len(nearest_vectors) > len(vectors_3d)/2: avec = mean_vector(intp_point, nearest_vectors) interpolated[0].append(np.array(intp_point)) interpolated[1].append(avec) self.interpolated_raw[eye].append(nearest_vectors) self.interpolation[eye] = {} self.interpolation[eye][cachename] = np.array(interpolated[0]), np.array(interpolated[1]) else: pass points, vectors = self.interpolation[eye][cachename] if repeats_separately: newpoints = [] newvecs = [] for i_point, point in enumerate(points): for vec in self.interpolated_raw[eye][i_point]: newpoints.append(point) newvecs.append(vec) points = np.array(newpoints) vectors = np.array(newvecs) # Vertical/horizontal angle limiting booleans = where_vertical_between(points, lower=self.va_limits[0], upper=self.va_limits[1], reverse=self.alimits_reverse) points = points[booleans] vectors = vectors[booleans] return points, vectorsAncestors
Instance variables
prop name-
Expand source code
@property def name(self): return self.get_specimen_name()
Methods
def get_2d_vectors(self, eye, **kwargs)-
Get's the 2D movement vectors (in the camera coordinate system) using N_nearest neighbour interpolation and averaging.
def get_3d_vectors(self, eye, correct_level=True, normalize_length=0.1, recalculate=False, strict=False, vertical_hardborder=False, repeats_separately=False, **kwargs)-
Equivalent to MAnalysers get_3d_vectors but interpolates with N-nearest neughbours.
repeats_separately : bool If True, return underlying MAnalyser vectors separetly (same points get repeated many times)
def get_N_specimens(self)def get_specimen_name(self)def setInterpolationSteps(self, horizontal_step, vertical_step)-
Set the resolution of the N-nearest neighbour interpolation in Maverager.get_2d_vectors.
INPUT ARGUMENTS horizontal_step vertical_step
Arguments horizontal_step and vertical_step refer to the rotation stages.
Inherited members