def test_inflection_point():
'''Test for function which obtains the melting point of all the samples'''
file_name = (
'../musical-robot/musicalrobot/data/10_17_19_PPA_Shallow_plate.tiff')
frames = edge_detection.input_file(file_name)
crop_frame = []
for frame in frames:
crop_frame.append(frame[35:85, 40:120])
# flip_frames = edge_detection.flip_frame(crop_frame)
n_samples = 9
n_rows = 3
n_columns = 3
labeled_samples = edge_detection.edge_detection(crop_frame, n_samples)
regprops = edge_detection.regprop(labeled_samples, crop_frame, n_rows,
n_columns)
sorted_regprops = edge_detection.sort_regprops(regprops, n_columns, n_rows)
temp, plate_temp = edge_detection.sample_temp(sorted_regprops, crop_frame)
s_peaks, s_infl = edge_detection.peak_detection(temp, plate_temp, 'Sample')
p_peaks, p_infl = edge_detection.peak_detection(temp, plate_temp, 'Plate')
inf_temp = edge_detection.inflection_point(temp, plate_temp, s_peaks,
p_peaks)
assert isinstance(inf_temp, list), 'Output is not a list'
assert len(
inf_temp) == n_samples, 'Wrong number of melting points determined'
return
def pixel_temp(frames, n_frames, n_columns, n_rows):
'''
Function to determine the temperature of the samples and plate locations by analysing
pixel values and finding peaks.
Parameters
-----------
frames: Array
The frames of a video to be analysed.
n_frames: Int
Number of frames in the video
n_columns: Int
Number of columns of samples in the image
n_rows: Int
Number of rows of samples in the image.
Returns
--------
m_df : Dataframe
A dataframe containing row and column coordinates of each sample
and its respective inflection point obtained.
'''
# flip_frames = edge_detection.flip_frame(frames)
#Function to obtained an equalized image using all the frames
#in the video.
img_eq = image_eq(n_frames, frames)
#Funtion to determine sum of pixels over all the rows and columns
#to obtain plots with peaks at the sample position in the array.
column_sum, row_sum = pixel_sum(img_eq)
# Function to find peaks from the column_sum and row_sum arrays
r_peaks, c_peaks = peak_values(column_sum, row_sum, n_columns, n_rows,
img_eq)
# Function to return a dataframe with sample locations and respective plate locations.
sample_location = locations(r_peaks, c_peaks, img_eq)
# Function to find pixel intensity at all sample locations
# and plate locations in each frame.
temp, plate_temp = pixel_intensity(sample_location,
frames,
x_name='Row',
y_name='Column',
plate_name='plate_location')
#Function to obtain the peaks in sample temperature profile
s_peaks, s_infl = edge_detection.peak_detection(temp, plate_temp, 'Sample')
# Function to obtain the peaks in plate location temperature profile
p_peaks, p_infl = edge_detection.peak_detection(temp, plate_temp, 'Plate')
# Function to obtain the inflection point(melting point) from the temperature profile.
inf_temp = inflection_point(temp, plate_temp, s_peaks, p_peaks)
# Dataframe with sample location (row and column coordinates) and respective inflection point.
m_df = pd.DataFrame({
'Row': sample_location.Row,
'Column': sample_location.Column,
'Melting point': inf_temp
})
return m_df
def final_result(temp, plate_temp, path):
'''
Funtion to classify temperature profiles as noisy or noiseless
Parameters
-----------
temp : List
Temperature of all the samples in
every frame of the video.
plate_temp : List
Temperature profiles of all the plate locations
path : String
Path to the location to temporarily store neural
network input images.
Returns
--------
result_df : Dataframe
Dataframe containing well number, predictions of noise net anf
inflection net and melting point.
'''
# Generating noise images
noise_image(temp, plate_temp, path)
# Making predictions using noise net
file_path = path + 'noise_images/'
result_df, nonoise_index = noise_prediction(file_path)
# Generating inflection images
inf_images(temp, plate_temp, 2, nonoise_index, path)
# Making prediction using inflection net
file_path = path + 'inf_images/'
inf_pred, inf_index = inf_prediction(file_path)
# Extracting melting point
s_peaks, s_infl = ed.peak_detection(temp, plate_temp, 'Sample')
melting_point = np.asarray(s_infl)[:, 0]
# Adding inflection and melting point data to the dataframe
result_df['Inf net'] = '-'
result_df['Melting point'] = '-'
for i in nonoise_index:
result_df['Inf net'].loc[i] = inf_pred[i]
for i in inf_index:
result_df['Melting point'].loc[i] = melting_point[i]
result_df['Inf net'].replace(0, 'Inflection', inplace=True)
result_df['Inf net'].replace(1, 'No Inflection', inplace=True)
result_df['Noise net'].replace(0, 'Noiseless', inplace=True)
result_df['Noise net'].replace(1, 'Noisy', inplace=True)
return result_df