def raj_cropper(csv_file):
### IMPORT ###
data = pd.read_csv(csv_file)
### BOUNDING ###
# Remove the excess first column, an artifact from exporting a csv.
data = data.drop('Unnamed: 0', axis=1)
# Filters the data set by the area to remove files that are above or below the thresholding limits.
data = area_filter(data, 1, 2)
data = data.reset_index()
data = data.drop('index', axis=1)
### PROCESSING ###
# Generate a file list from the data. As there are repeated filenames this compares the next file name to the current
# and if they are different adds the next filename to the list.
file_list = [data["file_name"][0]]
for i in range(0, data.shape[0] - 1):
if data["file_name"][i + 1] == data["file_name"][i]:
continue
else:
file_list.append(data["file_name"][i + 1])
# Create an empty array to add the data to.
cutout_dataframe = pd.DataFrame(
columns=['frame', 'X', 'Y', 'filename', 'cutout_array'])
for i in range(0, len(file_list)):
img = genr.load_img(file_list[i]) # loads in the file
print(file_list[i])
for j in range(0, data.shape[0]):
if data["file_name"][j] == file_list[i]:
y = data["centroid-0"][j]
x = data["centroid-1"][j]
print(x, y)
cutouts = np.array(square_area(img, x, y, 7))
cutout_current = pd.DataFrame(
{
'frame': [j],
'X': [x],
'Y': [y],
'filename': [file_list[i]],
'cutout_array': [cutouts]
},
index=["{}".format(j)])
cutout_dataframe = pd.concat(
[cutout_dataframe, cutout_current],
axis=0) # Final dataframe.
# Currently Data is offset by the rounding issue with the localisation. Data needs to be adjusted to match.
cutout_dataframe.to_csv('particle_position_crops.csv')
return cutout_dataframe
def matt_cropper(csv_pathname='{}/panda_data_molecule_locs{}'.format(
os.getcwd(), '.csv')):
# Iterate through indices in imported locations dataframe
molecules, save_data = load_csv(csv_pathname)
file = ""
for index in molecules.index:
# If the file name for current index does not match the current working file, load the file
if file is not "{}".format(molecules.loc[index, "file_name"]):
file = genr.load_img("{}".format(molecules.loc[index,
"file_name"]))
# Define crop area from centre of mass data in imported locations table
centre_y = int(np.floor(molecules.loc[index, "centroid-0"]))
centre_x = int(np.floor(molecules.loc[index, "centroid-1"]))
# Call cropping function
molecule_crop = point_crop(file, centre_y, centre_x, 7)
# Define variables for saving out data
file_nm, centroid_one, centroid_zero = molecules.loc[
index, "file_name"], molecules.loc[
index, "centroid-1"], molecules.loc[index, "centroid-0"]
# Format variables into a pandas series to allow saving to dataframe
molecule_crop_id = pd.Series({
"file_name": file_nm,
"x-coord": centroid_one,
"y-coord": centroid_zero,
"cutout_square": molecule_crop
})
# Convert series to dataframe, with index relating to current value of index variable
save_row = pd.DataFrame([molecule_crop_id], index=["{}".format(index)])
# Concatenate the data from this iteration to the created data frame
save_data = pd.concat([save_row, save_data], axis=0)
# Save finished dataframe to .csv in working directory
save_data.sort_index(axis=0)
save_data.to_csv("molecules_positions_crops.csv")
return save_data
# For each file in the above list, execute the chosen filter and threshold and then save it out as a numpy array
a = 0 # Set up a counter
localised_data = pd.DataFrame(
columns=['area', 'centroid-0', 'centroid-1', 'file_name'])
folder = "{}/storm_output_data".format(params["directory:"])
if not os.path.exists(folder):
os.mkdir(folder) # Makes a directory is there isn't one there.
for name in file_list:
a += 1
file_name = "{}/{}".format(params["directory:"], name)
print(file_name)
img = genr.load_img(file_name)
### FILTERING ###
# This takes the data and the filter params information and pulls out the relevant information to choose which
# function to run. Based on the "filter type:", and uses the parameters a and b as required.
# Matt, at the moment it does not account for your above if statement.
filtered_data = filters.filter_switcher(img, params)
### THRESHOLDING ###
# As above, with switcher adapted to thresholds
thresholded_data = thresholds.threshold_switcher(filtered_data, params)
# np.save('{}/thresholded_img_{}_{}'.format(folder, a, datetime.datetime.now()), thresholded_data)
### LOCALISATION ###