Module AmpliVision.src.objs.test_analyzer.test_analyzer
Classes
class TestAnalyzer (block)-
This class is responsible for getting and analyzing test results a.k.a phase B
Expand source code
class TestAnalyzer: "This class is responsible for getting and analyzing test results a.k.a phase B" def __init__(self, block): self.block = block # look only at the inner test square: self.test_square_img = block.get_test_area_img() # square used in csv export self.grid_index = block.index self.block_type = block.get_block_type() self.strip_sections = { "bkg": StripSection(self.test_square_img, 'bkg', block.rotation), "spot1": StripSection(self.test_square_img, 'spot1', block.rotation), "spot2": StripSection(self.test_square_img, 'spot2', block.rotation) } def analyze_test_result(self, double_thresh = False, display: bool = False): # should I name it main? "gets test results from a block, analyses them, and export them to csv" # find the positive spots with hsv mask # need to think about cases where mask for example return one pixel. # do you check for minimum contour size? do you only look for it manually? food for thought if display: print("rotation: ", self.block.rotation) # thresholds optimized for marker data rgb_spots = ColorContourExtractor.process_image( self.test_square_img, hsv_lower=[0, 40, 20], double_thresh=double_thresh, display=display ) if display: cv.waitKey(100) plt.close() cv.destroyAllWindows() self.add_positives_to_sections(rgb_spots, display=display) # find the negative spots "manually" through ratios self.add_negatives_to_sections(display=display) # get background color noise so we can remove it from other sections self.strip_sections['bkg'].set_total_avg_rgb() bkg_rgb_avg = self.strip_sections['bkg'].total_avg_rgb # remove background noise from other sections corrected_rgbs = [] for section in self.strip_sections.values(): if section.strip_type != 'bkg': if display: print(f"{section.strip_type}AVG RGB: {section.total_avg_rgb}") print("correcting: ", section.strip_type) corrected_rgbs.append(section.subtract_bkg(bkg_rgb_avg)) if display: print("\n") # validate results to catch any potential errors in the test "TODO: adapt validate_results to work with the new strip configuration" # self.validate_results() # export results to csv row = self.create_csv_row(corrected_rgbs) return row def add_positives_to_sections(self, rgb_spots, display: int = 0) -> None: "used to add positive result spots to appropriate strip section" # adds each spot to its strip section for spot in rgb_spots: # display the spot cpy = cv.drawContours(self.test_square_img.copy(), [spot], -1, (0, 255, 0), 1) #plt.imshow(f'TA/add_positives_to_sections', cv.resize(cpy, (400, 400))) for section in self.strip_sections.values(): if section.bounds_contour(spot): section.add_spot(self.block, spot, True, debug=display) # break # only adds to one section def add_negatives_to_sections(self, display: int = 0) -> None: "used to find negative result spots to appropriate strip section" for type, section in zip(self.strip_sections.keys(), self.strip_sections.values()): if len(section.spots) == 0: section.set_spots_manually(self.block, debug=display) def validate_results(self) -> None: "deals with test result potential positive, negative, false positive, error scenarios" results = self.get_section_results() # 1 test is properly positive (bkg, test, and spot2 line rgbs are > threshold) if results[1] & results[2]: print("Test worked properly and result is positive") # 2 test is properly negative (spot2 line rgb is > threshold) elif (not results[1]) & results[2]: print("Test worked properly and result is negative") # 3 spot2 error (bkg, and maybe test line rgbs are > threshold) else: print("Test may have not worked properly") def get_section_results(self) -> list[bool]: "returs a list of booleans representing the result (positive or negative) of each section bkg, test, spot2" results = [] # bkg, test, spot2 for strip in self.strip_sections.values(): strip_result = False # display # strip.print_spots() for spot in strip.spots: if spot["positive"] == True: strip_result = True break results.append(strip_result) print("\n") return results def create_csv_row(self, corrected_rgbs: list[list]) -> str: """ writes the test results to csv file row in format:\n date, time, grid_index, block_type, bkg_r, bkg_g, bkg_b, test_r, test_g, test_b, cntrl_r, cntrl_g, cntrl_b""" # get current date and time now = datetime.now() # format date and time date = now.strftime("%m/%d/%Y") time = now.strftime("%H:%M:%S") # setting all rgb values to None bkg_r = " None" bkg_g, bkg_b = bkg_r, bkg_r spot1_r, spot1_g, spot1_b = bkg_r, bkg_r, bkg_r spot2_r, spot2_g, spot2_b = bkg_r, bkg_r, bkg_r # get rgb values of each section if self.strip_sections['bkg'].total_avg_rgb != None: bkg_b, bkg_g, bkg_r = self.strip_sections['bkg'].total_avg_rgb # print("bkg rgb: ", bkg_r, bkg_g, bkg_b) if self.strip_sections['spot1'].total_avg_rgb != None: spot1_b, spot1_g, spot1_r = self.strip_sections['spot1'].total_avg_rgb #print("spot1 rgb: ", test_r, test_g, test_b) if self.strip_sections["spot2"].total_avg_rgb != None: spot2_b, spot2_g, spot2_r = self.strip_sections['spot2'].total_avg_rgb spot1_corr_b, spot1_corr_g, spot1_corr_r = corrected_rgbs[0] spot2_corr_b, spot2_corr_g, spot2_corr_r = corrected_rgbs[1] # create data to be written to csv data = [ date, time, self.grid_index, self.block_type, spot1_r, spot1_g, spot1_b, spot2_r, spot2_g, spot2_b, bkg_r, bkg_g, bkg_b, spot1_corr_r, spot1_corr_g, spot1_corr_b, spot2_corr_r, spot2_corr_g, spot2_corr_b ] return data def paint_spots(self, rgb_spot_results: dict[list]): """ colors the spots in the image accourding to the results Args: rgb_spot_results (dict[list]): dictionary containing the rgb values of the spots in the image format: {r: [mean1, std1, mean2, std2], g: [mean1, std1, mean2, std2], b: [mean1, std1, mean2, std2]} """ from numpy import random import time # can show improved performance if get_rgb_avg_of_contour TODO is done t = time.time() self.analyze_test_result() #print(f" analyze_test_result in {round(time.time() - t,2)}") image = self.test_square_img image_ = image.copy() for type, section in self.strip_sections.items(): # dont paint bkg if section.strip_type == 'bkg': continue # get the correct index for result i = 0 if type == 'spot1' else 2 rgb = [] means = [] for c in ('b', 'g', 'r'): mean, std = rgb_spot_results[c][i:i+2] means.append(mean) rgb.append(int(random.normal(mean, std))) rgb = tuple(rgb) if means == [0,0,0]: continue #print(f"Paiting {type} with {rgb}") image_ = section.paint_spot(image, rgb, display=False) self.block.set_test_area_img(image_) return self.blockMethods
def add_negatives_to_sections(self, display: int = 0) ‑> None-
used to find negative result spots to appropriate strip section
def add_positives_to_sections(self, rgb_spots, display: int = 0) ‑> None-
used to add positive result spots to appropriate strip section
def analyze_test_result(self, double_thresh=False, display: bool = False)-
gets test results from a block, analyses them, and export them to csv
def create_csv_row(self, corrected_rgbs: list[list]) ‑> str-
writes the test results to csv file row in format:
date, time, grid_index, block_type, bkg_r, bkg_g, bkg_b, test_r, test_g, test_b, cntrl_r, cntrl_g, cntrl_b
def get_section_results(self) ‑> list[bool]-
returs a list of booleans representing the result (positive or negative) of each section bkg, test, spot2
def paint_spots(self, rgb_spot_results: dict[list])-
colors the spots in the image accourding to the results
Args
rgb_spot_results:dict[list]- dictionary containing the rgb values of the spots in the image
format- {r: [mean1, std1, mean2, std2], g: [mean1, std1, mean2, std2], b: [mean1, std1, mean2, std2]}
def validate_results(self) ‑> None-
deals with test result potential positive, negative, false positive, error scenarios