How to use hierarchy_tags method in Lemoncheesecake
Best Python code snippet using lemoncheesecake
feature_extraction_helpers.py
Source:feature_extraction_helpers.py 
1# Imports2import Cluster3import extract4import glob5import partition6import pickle7import re8import timeit9import Visualiser10import numpy as np11import math12from pysal.esda.getisord import G_Local13from pysal.weights.Distance import DistanceBand14from descartes import PolygonPatch15import matplotlib.pyplot as plt16from tqdm import tqdm17import seaborn as sb18from matplotlib import cm19from matplotlib.ticker import LinearLocator, FormatStrFormatter20from shapely.geometry import Point, Polygon, LineString21from mpl_toolkits.mplot3d import Axes3D22def get_tags(hierarchy_tags):23 """24 Gets the x-y coordinates of the piecewise line defining the invasive margin.25 """26 tags = []27 x_tags, y_tags = [], []28 for tag in hierarchy_tags:29 if "X=" in tag and "Y=" in tag:30 x = tag.split("\"")[1]31 y = tag.split("\"")[3]32 x_tags.append(int(x))33 y_tags.append(int(y))34 for i, j in zip(x_tags, y_tags):35 tags.append((int(i), int(j)))36 return tags37def load_linestring(polygon_file):38 """39 Loads the x-y coordinates and creates a LineString object from them.40 This is useful for establishing the distance of cells from the41 invasive margin.42 """43 poly = None44 with open(polygon_file, "r") as f:45 hierarchy = f.read()46 p = re.compile("<*")47 hierarchy_tags = p.split(hierarchy)48 tags = get_tags(hierarchy_tags)49 poly = LineString(tags)50 if poly is None:51 raise Exception52 else:53 return poly54def load_shape(polygon_file):55 """56 Loads the x-y coordinates and creates a Polygon object from them.57 This is useful for establishing whether a given object is within58 the invasive margin or not.59 """60 poly = None61 with open(polygon_file, "r") as f:62 hierarchy = f.read()63 p = re.compile("<*")64 hierarchy_tags = p.split(hierarchy)65 tags = get_tags(hierarchy_tags)66 poly = Polygon(tags)67 if poly is None:68 raise Exception69 else:70 return poly71def plot_line_and_shape(ax, linestring, shape):72 """73 Given a matplotlib axes object, plots the Shapely LineString and Polygon74 objects onto the axes. Returns the axes.75 """76 patch = PolygonPatch(shape, fc="red", alpha=0.5, zorder=2)77 ax.add_patch(patch)78 xs = [x for (x, y) in linestring.coords]79 ys = [y for (x, y) in linestring.coords]80 ax.plot(xs, ys)81 return ax82def get_margin_cells(all_cells, linestring, my_distance):83 """84 Gets all cells that are within my_distance of the invasive margin.85 This includes TB, CD3, and CD8.86 """87 margin_cell_list = []88 assert (type(all_cells) == np.ndarray and all_cells.shape[1] >= 4), "type of all_cells is {}, all_cells.shape is {}".format(type(all_cells), all_cells.shape)89 x_avgs = (all_cells[:, 0] + all_cells[:, 1]) / 290 y_avgs = (all_cells[:, 2] + all_cells[:, 3]) / 291 xy_avgs = np.array((x_avgs, y_avgs)).transpose()92 for i in range(len(all_cells)):93 p = Point(xy_avgs[i])94 if p.distance(linestring) <= my_distance:95 margin_cell_list.append(all_cells[i])96 return np.asarray(margin_cell_list)97def get_core_cells(all_cells, shape, linestring, my_distance):98 """99 Gets all cells that are inside the invasive margin, but not within my_distance to it.100 This includes TB, CD3, and CD8.101 """102 core_cell_list = []103 assert (type(all_cells) == np.ndarray and all_cells.shape[1] >= 4), "type of all_cells is {}, all_cells.shape is {}".format(type(all_cells), all_cells.shape)104 x_avgs = (all_cells[:, 0] + all_cells[:, 1]) / 2105 y_avgs = (all_cells[:, 2] + all_cells[:, 3]) / 2106 xy_avgs = np.array((x_avgs, y_avgs)).transpose()107 for i in range(len(all_cells)):108 p = Point(xy_avgs[i])109 if p.distance(linestring) > my_distance and p.within(shape):110 core_cell_list.append(all_cells[i])111 return np.asarray(core_cell_list)112def get_margin_core_cells(all_cells, shape, linestring, my_distance):113 """114 Excludes all cells that are more than my_distance outside the115 invasive margin. This includes TB, CD3, and CD8.116 """117 margin_cell_list, core_cell_list = [], []118 assert (type(all_cells) == np.ndarray and all_cells.shape[1] >= 4), "type of all_cells is {}, all_cells.shape is {}".format(type(all_cells), all_cells.shape)119 x_avgs = (all_cells[:, 0] + all_cells[:, 1]) / 2120 y_avgs = (all_cells[:, 2] + all_cells[:, 3]) / 2121 xy_avgs = np.array((x_avgs, y_avgs)).transpose()122 for i in range(len(all_cells)):123 p = Point(xy_avgs[i])124 if p.distance(linestring) <= my_distance:125 margin_cell_list.append(all_cells[i])126 elif p.within(shape):127 core_cell_list.append(all_cells[i])128 return np.asarray(margin_cell_list), np.asarray(core_cell_list)129def get_lymphocytes_on_margin(cd3cd8, annotations_file):130 """131 Gets all lymphocytes (CD3/CD8) within 500um of the invasive margin.132 """133 all_cd3, all_cd8 = cd3cd8134 linestring = load_linestring(annotations_file)135 margin_cd3, margin_cd8 = [], []136 margin_cd3, margin_cd8 = np.asarray(get_margin_cells(all_cd3, linestring, 500)), np.asarray(get_margin_cells(all_cd8, linestring, 500))137 return margin_cd3, margin_cd8138def get_lymphocytes_on_margin_and_core(cd3cd8, annotations_file):139 """140 Gets all lymphocytes (CD3/CD8) except those that are more than 500um outside141 the invasive margin142 """143 all_cd3, all_cd8 = cd3cd8144 shape = load_shape(annotations_file)145 linestring = load_linestring(annotations_file)146 margin_cd3, core_cd3 = get_margin_core_cells(all_cd3, shape, linestring, 500)147 margin_cd8, core_cd8 = get_margin_core_cells(all_cd8, shape, linestring, 500)148 return np.asarray(margin_cd3), np.asarray(core_cd3), np.asarray(margin_cd8), np.asarray(core_cd8)149def partition_lymphocytes_for_heatmap(items, tile_size=886, to_list=True, input_type="clean",150 by_metric=True, scale=1):151 """152 Given appropriate parameters and a list of lymphocytes,153 divide them into a (T x T) grid and take the counts of each type154 of object for each tile. Return these 'heatmaps', rotated accordingly.155 """156 partitioned_items, _, _, _, _, _, _, _ = partition.partition(items, tile_size=tile_size,157 to_list=to_list, input_type=input_type,158 by_metric=by_metric, scale=scale)159 heatmap_cd3 = np.zeros(partitioned_items.shape, dtype=int)160 heatmap_cd8 = np.zeros(partitioned_items.shape, dtype=int)161 heatmap_cd3cd8 = np.zeros(partitioned_items.shape, dtype=int)162 (width, height) = partitioned_items.shape163 for col in range(width):164 for row in range(height):165 cells = partitioned_items[col][row]166 heatmap_cd3cd8[col][row] += len(cells)167 for cell in cells:168 if cell[4] > 0:169 heatmap_cd3[col][row] += 1170 elif cell[5] > 0:171 heatmap_cd8[col][row] += 1172 heatmap_cd3 = np.flip(np.rot90(heatmap_cd3), 0)173 heatmap_cd8 = np.flip(np.rot90(heatmap_cd8), 0)174 heatmap_cd3cd8 = np.flip(np.rot90(heatmap_cd3cd8), 0)175 return heatmap_cd3, heatmap_cd8, heatmap_cd3cd8176def partition_tb_for_heatmap(items, tile_size=886, to_list=True, input_type="clean",177 by_metric=True, scale=1):178 """179 Given appropriate parameters and a list of tumour buds,180 divide them into a (T x T) grid and take the counts of tumour buds181 for each tile. Return this 'heatmap', rotated/flipped accordingly.182 """183 partitioned_tb, _, _, _, _, _, _, _ = partition.partition(items, tile_size=tile_size,184 to_list=to_list, input_type=input_type,185 by_metric=by_metric, scale=scale)186 heatmap_tb = np.zeros(partitioned_tb.shape, dtype=int)187 for i in range(partitioned_tb.shape[0]):188 for j in range(partitioned_tb.shape[1]):189 heatmap_tb[i][j] += len(partitioned_tb[i][j])190 heatmap_tb = np.flip(np.rot90(heatmap_tb), 0)191 return heatmap_tb192def partition_nearby_lymphocytes_tb_for_heatmap(items, tile_size=886, to_list=True, input_type="mixed",193 by_metric=True, scale=1):194 """195 Given appropriate parameters and a list of tumour buds and lymphocytes within 50um of any tumour bud,196 divide them into a (T x T) grid and take the counts of tumour buds and lymphocytes197 for each tile. Return these 'heatmaps', rotated accordingly.198 """199 partitioned_items, tiles, \200 lymph_w, lymph_h, \201 clust_w, clust_h, \202 x_step, y_step = partition.partition(items, tile_size=tile_size,203 to_list=to_list, input_type=input_type,204 by_metric=by_metric, scale=scale)205 counts = Cluster.get_lymphocyte_cluster_density_heatmap(partitioned_items, partitioned_items.shape, tiles,206 lymph_w, lymph_h, clust_w, clust_h, x_step, y_step)207 heatmap_tb = np.zeros(counts.shape, dtype=int)208 heatmap_nearby_cd3 = np.zeros(counts.shape, dtype=int)209 heatmap_nearby_cd8 = np.zeros(counts.shape, dtype=int)210 heatmap_nearby_cd3cd8 = np.zeros(counts.shape, dtype=int)211 for i in range(counts.shape[0]):212 for j in range(counts.shape[1]):213 heatmap_nearby_cd3[i][j] += counts[i][j][0]214 heatmap_nearby_cd8[i][j] += counts[i][j][1]215 heatmap_nearby_cd3cd8[i][j] += counts[i][j][2]216 heatmap_tb[i][j] += counts[i][j][3]217 heatmap_nearby_cd3 = np.flip(np.rot90(heatmap_nearby_cd3), 0)218 heatmap_nearby_cd8 = np.flip(np.rot90(heatmap_nearby_cd8), 0)219 heatmap_nearby_cd3cd8 = np.flip(np.rot90(heatmap_nearby_cd3cd8), 0)220 heatmap_tb = np.flip(np.rot90(heatmap_tb), 0)221 return heatmap_tb, heatmap_nearby_cd3, heatmap_nearby_cd8, heatmap_nearby_cd3cd8222def get_cluster_code(cluster_name):223 """224 Path/to/file/.../ -> split on the slash and get the last token225 SN---.p -> just truncate the last 2 characters226 """227 filename = cluster_name.split('/')[-1]228 return filename[0:-2]229def get_annotations_code(annotations_name):230 """231 SN---_----_job----.annotations -> get the first 5 characters232 """233 filename = annotations_name.split('/')[-1]234 return filename[0:5]235def load_cluster_list(cluster_files):236 """237 Cluster file, load through pickle.238 """239 for cluster_filename in sorted(glob.glob(cluster_files)):240 cluster_data = pickle.load(open(cluster_filename, "rb"))241 yield cluster_filename, cluster_data242def load_lymphocyte_list(lymphocyte_files):243 """244 Lymphocyte file, load through csv.245 """246 for lymphocyte_filename in sorted(glob.glob(lymphocyte_files)):247 lymphocyte_data = extract.load(lymphocyte_filename)248 yield lymphocyte_filename, lymphocyte_data249def load_len_lymphocyte_list(lymphocyte_files):250 """251 Lymphocyte file, load through csv.252 """253 for lymphocyte_filename in sorted(glob.glob(lymphocyte_files)):254 lymphocyte_data = extract.load(lymphocyte_filename)255 yield lymphocyte_filename, len(lymphocyte_data)256def load_cluster_tb_metadata(cluster_metadata_files):257 """258 Cluster metadata file, load through .txt file.259 Get the appropriate line, get the val at the end.260 """261 for metadata_filename in tqdm(sorted(glob.glob(cluster_metadata_files))):262 f = open(metadata_filename, 'r')263 contents = f.readlines()264 cluster_count_val = -1265 tb_count_val = -1266 for line in contents:267 if "CLU" in line:268 cluster_count_val = int(line.split(": ")[-1])269 if "TB" in line:270 tb_count_val = int(line.split(": ")[-1])271 if cluster_count_val != -1 and tb_count_val != -1:272 yield metadata_filename, cluster_count_val, tb_count_val273 else:274 raise Exception("No cluster/TB value associated in file " + metadata_filename)275def load_lymphocyte_metadata(lymphocyte_metadata_files):276 """277 Lymphocyte metadata file, load through .txt file.278 Get the appropriate line, get the val at the end.279 """280 for metadata_filename in tqdm(sorted(glob.glob(lymphocyte_metadata_files))):281 f = open(metadata_filename, 'r')282 contents = f.readlines()283 cd3cd8_count_val = -1284 cd3_count_val = -1285 cd8_count_val = -1286 for line in contents:287 if "CD3&CD8" in line:288 cd3cd8_count_val = int(line.split(": ")[-1])289 continue290 if "CD3" in line:291 cd3_count_val = int(line.split(": ")[-1])292 continue293 if "CD8" in line:294 cd8_count_val = int(line.split(": ")[-1])295 continue296 if cd3cd8_count_val > -1 and cd3_count_val > -1 and cd8_count_val > -1:297 yield metadata_filename, cd3_count_val, cd8_count_val, cd3cd8_count_val298 else:299 raise Exception("No CD3/CD8/CD3CD8 value associated in file " + metadata_filename)300def load_margin(margin_files):301 """302 Margin defining shape of invasive margin, load through helpers.303 """304 for margin_filename in sorted(glob.glob(margin_files)):305 linestring, shape = load_linestring(margin_filename), load_shape(margin_filename)306 yield margin_filename, linestring, shape307def load_meta(obj_type, meta_data_files, code_extract_func):308 """309 meta_data_values = [(meta_code1, meta_val1), (meta_code2, meta_val2), ...]310 """311 meta_data_values = []312 for meta_data_filename in tqdm(sorted(glob.glob(meta_data_files))):313 f = open(meta_data_filename, 'r')314 contents = f.readlines()315 obj_count_val = -1316 for line in contents:317 if obj_type in line:318 obj_count_val = int(line.split(": ")[-1])319 meta_data_values.append((code_extract_func(meta_data_filename), obj_count_val))320 return meta_data_values321def code_extract_func(filename):322 return filename.split('/')[-1].split('.')[0]323def visualisation_block(visualise_func, data, linestring, shape, title, image_filename):324 fig, ax = plt.subplots(1, 1, figsize=(6, 6))325 ax = visualise_func(data, ax, title=title)326 xs = [x for (x, y) in linestring.coords]327 ys = [y for (x, y) in linestring.coords]328 patch = PolygonPatch(shape, fc="red", alpha=0.2, zorder=2)329 ax.add_patch(patch)330 ax.plot(xs, ys)331 fig.savefig(image_filename, dpi=100, bbox_inches="tight")332def visualisation_interaction_block(data, linestring, shape, title, image_filename):333 fig, ax = plt.subplots(1, 1, figsize=(6, 6))334 ax = Visualiser.visualise_clusters_and_lymphocytes_on_ax(data, ax, title=title)335 xs = [x for (x, y) in linestring.coords]336 ys = [y for (x, y) in linestring.coords]337 patch = PolygonPatch(shape, fc="red", alpha=0.2, zorder=2)338 ax.add_patch(patch)339 ax.plot(xs, ys)340 fig.savefig(image_filename, dpi=100, bbox_inches="tight")341def statistic_generation_block(partition_func, statistic_funcs, data, meta_data_value, d, heatmap_filename_func, target_directory_stub, filename):342 heatmap = partition_func(data, meta_data_value, d)343 assert heatmap.shape[2] == 4, "Output of transform should be 3D numpy array with 4 layers" # Layers for TB, CD3, CD8, CD3CD8 respectively344 for (statistic_name, statistic_func, statistic_mask) in statistic_funcs:345 transformed_data = statistic_func(heatmap)346 for i, obj_type in enumerate(["TB_CD3", "TB_CD8", "TB_CD3CD8"]):347 heatmap_filename = heatmap_filename_func(target_directory_stub + obj_type + "_" + str(d) + "/", filename, statistic_name)348 p_file = open(heatmap_filename, "wb")349 pickle.dump(transformed_data[:, :, i+1], p_file)350 p_file.close()351 fig, ax = plt.subplots(1, 1, figsize=(6, 6))352 title = "Heatmap for " + target_directory_stub.split('/')[-1] + obj_type + " with d=" + str(d)353 ax = Visualiser.visualise_heatmaps(transformed_data[:, :, i+1], ax, title, statistic_mask(transformed_data[:, :, i+1]))354 image_filename = heatmap_filename_func(target_directory_stub + obj_type + "_IMAGES/", filename, "_heatmap" + "_" + str(d) + "_" + statistic_name)[0:-2] + ".png"355 fig.savefig(image_filename, dpi=100, bbox_inches="tight")356def get_images_with_margins(obj_type, obj_files, expected_num_files, meta_data_files, load_obj_func, target_directory_stub, convert_func, visualise_func, image_filename_func, partition_func, heatmap_filename_func, margin_files, d):357 meta_data_pairs = load_meta(obj_type, meta_data_files, code_extract_func)358 meta_data_codes = [code for (code, val) in meta_data_pairs]359 meta_data_values = [val for (code, val) in meta_data_pairs]360 num_files = 0361 obj_file_type = obj_type if obj_type != "CD3&CD8" else "CD3CD8"362 for (filename, raw_data), (margin_filename, linestring, shape) in zip(load_obj_func(obj_files), load_margin(margin_files)):363 meta_data_value = 0364 try:365 meta_data_value = meta_data_values[meta_data_codes.index(code_extract_func(filename))]366 except:367 raise Exception("Meta-data does not have a value for file " + filename + "\nwith obj_type " + obj_type)368 all_data = convert_func(raw_data, meta_data_value)369 assert type(all_data) == np.ndarray370 ##Â ALL371 print(code_extract_func(filename) + ": " + "All {}".format(obj_type))372 visualisation_block(visualise_func, all_data, linestring, shape, "All {}".format(obj_type), image_filename_func(target_directory_stub + "ALL_{}/".format(obj_file_type), filename))373 statistic_generation_block(partition_func, [("_density", density)], all_data, meta_data_value, [heatmap_filename_func(target_directory_stub + "ALL_{}/".format(obj_file_type), filename)])374 margin_data, core_data = get_margin_core_cells(all_data, shape, linestring, 500)375 assert margin_data.shape[0] > 0 and margin_data.shape[1] >= 4, "Something wrong with {} margin data".format(obj_type)376 assert core_data.shape[0] > 0 and core_data.shape[1] >= 4, "Something wrong with {} core data".format(obj_type)377 ## MARGIN378 print(code_extract_func(filename) + ": " + "{} (IM)".format(obj_type))379 visualisation_block(visualise_func, margin_data, linestring, shape, "{} (IM)".format(obj_type), image_filename_func(target_directory_stub + "MARGIN_{}/".format(obj_file_type), filename))380 statistic_generation_block(partition_func, [("_density", density)], margin_data, meta_data_value, [heatmap_filename_func(target_directory_stub + "MARGIN_{}/".format(obj_file_type), filename)])381 ## CORE382 print(code_extract_func(filename) + ": " + "{} (CT)".format(obj_type))383 visualisation_block(visualise_func, core_data, linestring, shape, "{} (CT)".format(obj_type), image_filename_func(target_directory_stub + "CORE_{}/".format(obj_file_type), filename))384 statistic_generation_block(partition_func, [("_density", density)], core_data, meta_data_value, [heatmap_filename_func(target_directory_stub + "CORE_{}/".format(obj_file_type), filename)])385 ## IMCT386 print(code_extract_func(filename) + ": " + "{} (IMCT)".format(obj_type))387 imct_data = np.concatenate((margin_data, core_data), axis=0)388 assert imct_data.shape[0] == (margin_data.shape[0] + core_data.shape[0]), "Margin and core cells not added properly"389 del(margin_data) # numpy creates a copy on concat, so safe to delete390 del(core_data) # these two arrays391 visualisation_block(visualise_func, imct_data, linestring, shape, "{} (IMCT)".format(obj_type), image_filename_func(target_directory_stub + "IMCT_{}/".format(obj_file_type), filename))392 statistic_generation_block(partition_func, [("_density", density)], imct_data, meta_data_value, [heatmap_filename_func(target_directory_stub + "IMCT_{}/".format(obj_file_type), filename)])393 num_files += 1394 assert(num_files == expected_num_files)395def get_images_with_interactions_condensed(tb_files, lym_files, expected_num_files, meta_data_files,396 target_directory_stub, heatmap_transform, statistics, image_filename_func,397 heatmap_filename_func, margin_files, d):398 tb_meta_pairs = load_meta("TB", meta_data_files, code_extract_func)399 tb_meta_codes = [code for (code, val) in tb_meta_pairs]400 tb_meta_values = [val for (code, val) in tb_meta_pairs]401 cd3_meta_pairs = load_meta("CD3", meta_data_files, code_extract_func)402 cd8_meta_pairs = load_meta("CD8", meta_data_files, code_extract_func)403 cd3cd8_meta_pairs = load_meta("CD3&CD8", meta_data_files, code_extract_func)404 cd3_meta_codes = [code for (code, val) in cd3_meta_pairs]405 cd8_meta_codes = [code for (code, val) in cd8_meta_pairs]406 cd3cd8_meta_codes = [code for (code, val) in cd3cd8_meta_pairs]407 assert(cd3_meta_codes == cd8_meta_codes and cd8_meta_codes == cd3cd8_meta_codes)408 cd3_meta_values = [val for (code, val) in cd3_meta_pairs]409 cd8_meta_values = [val for (code, val) in cd8_meta_pairs]410 cd3cd8_meta_values = [val for (code, val) in cd3cd8_meta_pairs]411 num_files = 0412 for (tb_filename, cluster_raw_data), (lym_filename, lym_raw_data), (margin_filename, linestring, shape) in zip(load_cluster_list(tb_files),413 load_lymphocyte_list(lym_files),414 load_margin(margin_files)):415 tb_meta_value = 0416 cd3_meta_value, cd8_meta_value, cd3cd8_meta_value = 0, 0, 0417 try:418 tb_meta_value = tb_meta_values[tb_meta_codes.index(code_extract_func(tb_filename))]419 except:420 raise Exception("Meta-data does not have a value for file " + tb_filename + "\nwith obj_type TB")421 try:422 cd3_meta_value = cd3_meta_values[cd3_meta_codes.index(code_extract_func(lym_filename))]423 cd8_meta_value = cd8_meta_values[cd8_meta_codes.index(code_extract_func(lym_filename))]424 cd3cd8_meta_value = cd3cd8_meta_values[cd3cd8_meta_codes.index(code_extract_func(lym_filename))]425 except:426 raise Exception("Meta-data does not have a value for file " + lym_filename)427 all_tb_data = convert_cluster_to_tb(cluster_raw_data, tb_meta_value)428 all_lym_data = convert_lymphocyte_to_cd3cd8(lym_raw_data, cd3cd8_meta_value)429 assert type(all_tb_data) == np.ndarray and type(all_lym_data) == np.ndarray430 #####################431 ### CD3&CD8 BLOCK ###432 #####################433 combined_tb_lym_data = merge_tb_lym(all_tb_data, all_lym_data)434 ##Â ALL - TB AND CD3/CD8435 title = "All TB and CD3/CD8"436 print(code_extract_func(tb_filename) + ": " + title)437 visualisation_interaction_block(combined_tb_lym_data, linestring, shape, title, image_filename_func(target_directory_stub + "ALL_TB_CD3CD8_IMAGES/", tb_filename))438 statistic_generation_block(heatmap_transform, statistics, combined_tb_lym_data, tb_meta_value + cd3cd8_meta_value, 50, heatmap_filename_func, target_directory_stub + "ALL_", tb_filename)439 statistic_generation_block(heatmap_transform, statistics, combined_tb_lym_data, tb_meta_value + cd3cd8_meta_value, 100, heatmap_filename_func, target_directory_stub + "ALL_", tb_filename)440 start_time = timeit.default_timer()441 tb_lym_margin_data, tb_lym_core_data = get_margin_core_cells(combined_tb_lym_data, shape, linestring, 500)442 elapsed = timeit.default_timer() - start_time443 print("get_margin_core_cells() took:", elapsed)444 del(combined_tb_lym_data)445 assert tb_lym_margin_data.shape[0] > 0 and tb_lym_margin_data.shape[1] >= 4, "Something wrong with TB-CD3CD8 margin data"446 assert tb_lym_core_data.shape[0] > 0 and tb_lym_core_data.shape[1] >= 4, "Something wrong with TB-CD3CD8 core data"447 ## MARGIN - TB AND CD3/CD8448 title = "TB and CD3/CD8 (IM)"449 print(code_extract_func(tb_filename) + ": " + title)450 visualisation_interaction_block(tb_lym_margin_data, linestring, shape, title, image_filename_func(target_directory_stub + "MARGIN_TB_CD3CD8_IMAGES/", tb_filename))451 statistic_generation_block(heatmap_transform, statistics, tb_lym_margin_data, len(tb_lym_margin_data), 50, heatmap_filename_func, target_directory_stub + "MARGIN_", tb_filename)452 statistic_generation_block(heatmap_transform, statistics, tb_lym_margin_data, len(tb_lym_margin_data), 100, heatmap_filename_func, target_directory_stub + "MARGIN_", tb_filename)453 ## CORE - TB AND CD3/CD8454 title = "TB and CD3/CD8 (CT)"455 print(code_extract_func(tb_filename) + ": " + title)456 visualisation_interaction_block(tb_lym_core_data, linestring, shape, title, image_filename_func(target_directory_stub + "CORE_TB_CD3CD8_IMAGES/", tb_filename))457 statistic_generation_block(heatmap_transform, statistics, tb_lym_core_data, len(tb_lym_core_data), 50, heatmap_filename_func, target_directory_stub + "CORE_", tb_filename)458 statistic_generation_block(heatmap_transform, statistics, tb_lym_core_data, len(tb_lym_core_data), 100, heatmap_filename_func, target_directory_stub + "CORE_", tb_filename)459 ## IMCT - TB AND CD3/CD8460 title = "TB and CD3/CD8 (IMCT)"461 print(code_extract_func(tb_filename) + ": " + title)462 tb_lym_imct_data = np.concatenate((tb_lym_margin_data, tb_lym_core_data), axis=0)463 assert tb_lym_imct_data.shape[0] == (tb_lym_margin_data.shape[0] + tb_lym_core_data.shape[0]), "Margin and core cells not added properly"464 visualisation_interaction_block(tb_lym_imct_data, linestring, shape, title, image_filename_func(target_directory_stub + "IMCT_TB_CD3CD8_IMAGES/", tb_filename))465 statistic_generation_block(heatmap_transform, statistics, tb_lym_imct_data, len(tb_lym_imct_data), 50, heatmap_filename_func, target_directory_stub + "IMCT_", tb_filename)466 statistic_generation_block(heatmap_transform, statistics, tb_lym_imct_data, len(tb_lym_imct_data), 100, heatmap_filename_func, target_directory_stub + "IMCT_", tb_filename)467 num_files += 1468 assert(num_files == expected_num_files)469# def get_images_with_interactions(tb_files, tb_convert, lym_files, lym_convert, tb_lym_interaction_funcs, expected_num_files, meta_data_files, target_directory_stub, visualise_funcs, title_tree, output_filename_func, margin_files, d):470 """471 TB and do CD3, CD8, CD3&CD8 all at once.472 Four(!) levels of looping:473 for all files:474 meta vals...475 for tb_select, lym_select (must be the same region!):476 clean_tb, clean_lym = select_tb(), select_lym()477 for all interactions between current tb and lym selection (target_directories):478 create visualisation (pointwise-scatter with margin)479 for all statistics of interactions between tb and lym:480 ...481 num_files++482 Titles, target directories either need stubs or need to be hierarichial. Visualisation funcs are fine.483 """484 tb_meta_pairs = load_meta("TB", meta_data_files, code_extract_func)485 tb_meta_codes = [code for (code, val) in tb_meta_pairs]486 tb_meta_values = [val for (code, val) in tb_meta_pairs]487 cd3_meta_pairs = load_meta("CD3", meta_data_files, code_extract_func)488 cd8_meta_pairs = load_meta("CD8", meta_data_files, code_extract_func)489 cd3cd8_meta_pairs = load_meta("CD3&CD8", meta_data_files, code_extract_func)490 cd3_meta_codes = [code for (code, val) in cd3_meta_pairs]491 cd8_meta_codes = [code for (code, val) in cd8_meta_pairs]492 cd3cd8_meta_codes = [code for (code, val) in cd3cd8_meta_pairs]493 assert(cd3_meta_codes == cd8_meta_codes and cd8_meta_codes == cd3cd8_meta_codes)494 cd3_meta_values = [val for (code, val) in cd3_meta_pairs]495 cd8_meta_values = [val for (code, val) in cd8_meta_pairs]496 cd3cd8_meta_values = [val for (code, val) in cd3cd8_meta_pairs]497 num_files = 0498 for (tb_file, tb_raw_data), (lym_file, lym_raw_data) in zip(load_cluster_list(tb_files), load_lymphocyte_list(lym_files)):499 tb_meta_value, lym_meta_value = 0, 0500 try:501 tb_meta_value = tb_meta_values[tb_meta_codes.index(code_extract_func(tb_file))]502 cd3_meta_value = cd3_meta_values[cd3_meta_codes.index(code_extract_func(lym_file))]503 cd8_meta_value = cd8_meta_values[cd8_meta_codes.index(code_extract_func(lym_file))]504 cd3cd8_meta_value = cd3cd8_meta_values[cd3cd8_meta_codes.index(code_extract_func(lym_file))]505 except:506 raise Exception("Meta-data does not have a value for file-pair (" + tb_file + ", " + lym_file + ")")507 tb_data = tb_convert(tb_raw_data, tb_meta_value)508 lym_data = lym_convert(tb_raw_data, lym_meta_value)509 for selection in region_selections:510 selected_tb = selection(tb_data, linestring, shape, d)511 selected_lym = selection(lym_data, linestring, shape, d)512 for interaction_func in tb_lym_interaction_funcs:513 tb_lym_data = interaction_func(selected_tb, selected_lym)514 for feature in feature_funcs:515 feature_map = feature(tb_lym_data)516 tb_data = feature_map["TB"]517 cd3_data = feature_map["CD3"]518 cd8_data = feature_map["CD8"]519 cd3cd8_data = feature_map["CD3&CD8"]520 # for beta_transform, title_list, target_directory_list in zip(beta_transforms, title_tree, target_directory_tree):521 # clean_beta_data = beta_transform(beta_raw_data, beta_meta_value)522 # for alpha_beta_transform, visualise_func, title, target_directory in zip(alpha_beta_transforms, visualise_funcs, title_list, target_directory_list):523 # clean_data = alpha_beta_transform(clean_alpha_data, clean_beta_data)524 # fig, ax = plt.subplots(1, 1, figsize=(6, 6))525 # ax = visualise_func(clean_data, ax, title=title)526 # output_filename = output_filename_func(target_directory, alpha_file) # doesn't matter whether alpha-file or beta-file for the name527 # fig.savefig(output_filename, dpi=100, bbox_inches="tight")528 num_files += 1529 assert(num_files == expected_num_files)530def convert_cluster_to_tb(cluster_data, expected_size):531 data = np.array([[int(row[2]), int(row[1]), int(row[4]), int(row[3])] for row in cluster_data if int(row[0]) > 0 and int(row[0]) < 5])532 assert(len(data) == expected_size or expected_size == -1) # Allow a by-pass in the case of results we can't tell ahead of time.533 return data534def convert_lymphocyte_to_cd3(lymphocyte_data, expected_size):535 data = np.array([[int(row[0]), int(row[1]), int(row[2]), int(row[3]), int(row[4]), int(row[5])] for row in lymphocyte_data if int(row[4]) > 0])536 assert(len(data) == expected_size or expected_size == -1)537 return data538def convert_lymphocyte_to_cd8(lymphocyte_data, expected_size):539 data = np.array([[int(row[0]), int(row[1]), int(row[2]), int(row[3]), int(row[4]), int(row[5])] for row in lymphocyte_data if int(row[5]) > 0])540 assert(len(data) == expected_size or expected_size == -1)541 return data542def convert_lymphocyte_to_cd3cd8(lymphocyte_data, expected_size):543 data = np.array([[int(row[0]), int(row[1]), int(row[2]), int(row[3]), int(row[4]), int(row[5])] for row in lymphocyte_data if int(row[4]) > 0 or int(row[5]) > 0])544 assert(len(data) == expected_size or expected_size == -1)545 return data546def density(heatmap):547 """548 Assumes a non-interacting heatmap of lists.549 """550 densities = np.vectorize(len)(heatmap)551 return densities552def density_values(heatmap):553 return heatmap # it already is the answer554def ratio_values(heatmap):555 w, h = heatmap.shape[0], heatmap.shape[1]556 ratios = np.zeros(heatmap.shape)557 for i in range(w):558 for j in range(h):559 tb_val = heatmap[i][j][0]560 cd3_val = heatmap[i][j][1]561 cd8_val = heatmap[i][j][2]562 cd3cd8_val = heatmap[i][j][3]563 ratios[i][j][0] = 0 # don't care about tb, only the other 3564 ratios[i][j][1] = cd3_val / tb_val if tb_val != 0 else -1565 ratios[i][j][2] = cd8_val / tb_val if tb_val != 0 else -1566 ratios[i][j][3] = cd3cd8_val / tb_val if tb_val != 0 else -1567 return ratios568# def get_num_hotspots(heatmap, p_value_threshold=0.05, Z_score_threshold=1.96):569# hotspots = np.zeros((heatmap.shape[0], heatmap.shape[1], 2)) # layer 0 corresponds to coldspots, layer 1 to hotspots570# for i in range(array_of_heatmaps.shape[0]):571# all_count = 0572# for j in range(array_of_heatmaps.shape[1]):573# num_hot, num_cold = 0, 0574# (n1, n2) = heatmap.shape575# points = [[(x, y) for y in range(n2)] for x in range(n1)]576# dist = math.sqrt(2)577# with sb.axes_style("white"):578# msk = heatmap != 0579# points = [points[r][c] for r in range(n1) for c in range(n2) if msk[r][c]]580# w = DistanceBand(points, threshold=dist)581# lg_star = G_Local(heatmap[msk], w, transform='B', star=True)582# Z_scores = np.zeros(heatmap.shape, dtype=np.float32)583# p_values = np.zeros(heatmap.shape, dtype=np.float32)584# ind = 0585# for r in range(n1):586# for c in range(n2):587# if msk[r][c]:588# Z_scores[r, c] = lg_star.Zs[ind]589# p_values[r, c] = lg_star.p_sim[ind]590# ind += 1591# else:592# Z_scores[r][c] = -100.0 # why -100?593# for ind in range(lg_star.Zs.shape[0]):594# if lg_star.p_sim[ind] < p_value_threshold:595# if lg_star.Zs[ind] > Z_score_threshold:596# num_hot += 1597# elif lg_star.Zs[ind] < -(Z_score_threshold):598# num_cold += 1599# array_of_hotspots[i, all_count] = num_cold600# array_of_hotspots[i, all_count+1] = num_hot601# return heatmap602def get_tb_images(cluster_files, expected_num_files, meta_data_files, target_directory_stub, margin_files, d):603 # All, IM, CT, IM&CT604 obj_type = "TB"605 load_obj_func = load_cluster_list606 convert_func = lambda data, exp_size: convert_cluster_to_tb(data, exp_size)607 visualise_func = lambda data, ax, title: Visualiser.visualise_clusters_on_ax(data, ax, size_included=False, title=title)608 partition_func = lambda entire_list, exp_size: partition.partition(entire_list, tile_size=886, to_list=True, input_type="clean", by_metric=True, scale=1)[0]609 image_filename_func = lambda target_directory, filename: target_directory + code_extract_func(filename) + ".png"610 heatmap_filename_func = lambda target_directory, filename, statistic: target_directory + code_extract_func(filename) + statistic + ".p"611 args = (obj_type, cluster_files, expected_num_files, meta_data_files, load_obj_func,612 target_directory_stub, convert_func, visualise_func, image_filename_func, partition_func,613 heatmap_filename_func, margin_files, d)614 get_images_with_margins(*args)615def get_lym_images(lym_type, lymphocyte_files, expected_num_files, meta_data_files, target_directory_stub, margin_files, d):616 # All, IM, CT, IM&CT, for CD3, CD8, CD3&CD8617 load_obj_func = load_lymphocyte_list618 if lym_type == "CD3":619 convert_func = lambda data, exp_size: convert_lymphocyte_to_cd3(data, exp_size)620 elif lym_type == "CD8":621 convert_func = lambda data, exp_size: convert_lymphocyte_to_cd8(data, exp_size)622 else: # CD3&CD8623 convert_func = lambda data, exp_size: convert_lymphocyte_to_cd3cd8(data, exp_size)624 if lym_type == "CD3&CD8":625 visualise_func = lambda data, ax, title: Visualiser.visualise_lymphocytes_on_ax(data, ax, title=title)626 else:627 visualise_func = lambda data, ax, title: Visualiser.visualise_one_cd_on_ax(data, ax, "CD3", title=title, size=0.1, colour='r')628 partition_func = lambda entire_list, exp_size: partition.partition(entire_list, tile_size=886, to_list=True, input_type="clean", by_metric=True, scale=1)[0]629 image_filename_func = lambda target_directory, filename: target_directory + code_extract_func(filename) + ".png"630 heatmap_filename_func = lambda target_directory, filename, statistic: target_directory + code_extract_func(filename) + statistic + ".p"631 args = (lym_type, lymphocyte_files, expected_num_files, meta_data_files, load_obj_func,632 target_directory_stub, convert_func, visualise_func, image_filename_func,633 partition_func, heatmap_filename_func, margin_files, d)634 get_images_with_margins(*args)635def partition_and_transform_tb_interactions(items, meta_data_value, d):636 """637 Mixed partition, produce heatmap of interaction given some function of interaction.638 """639 init_num_objs = len(items)640 partitioned_items, tiles, \641 lymph_w, lymph_h, \642 clust_w, clust_h, \643 x_step, y_step = partition.partition(items, tile_size=886, to_list=True, input_type="mixed",644 by_metric=True, scale=1)645 num_objs = 0646 for i in range(partitioned_items.shape[0]):647 for j in range(partitioned_items.shape[1]):648 (lymphocytes, cancer_clusters) = partitioned_items[i][j]649 num_objs += (len(lymphocytes) + len(cancer_clusters))650 assert num_objs == init_num_objs and init_num_objs == meta_data_value, "num_objs: {}, init_num_objs: {}, and meta_data_value: {}".format(num_objs, init_num_objs, meta_data_value)651 heatmap = Cluster.get_interacting_object_counts(partitioned_items, d, partitioned_items.shape, tiles,652 lymph_w, lymph_h, clust_w, clust_h, x_step, y_step)653 assert heatmap.shape[0] == partitioned_items.shape[0] and heatmap.shape[1] == partitioned_items.shape[1]654 heatmap = np.flip(np.rot90(heatmap), 0)655 return heatmap656def merge_tb_lym(tb_data, lym_data):657 """658 Assumes tb_data to be np array of form:659 +------+------+------+------+660 | xMin | xMax | yMin | yMax |661 +------+------+------+------+662 | .... | .... | .... | .... |663 +------+------+------+------+664 lym data to be np array of form:665 +------+------+------+------+-----+-----+666 | xMin | xMax | yMin | yMax | CD3 | CD8 | <-- Note that CD3 and CD8 columns are (effectively) binary classifications.667 +------+------+------+------+-----+-----+668 | .... | .... | .... | .... | ... | ... |669 +------+------+------+------+-----+-----+670 With the output to be of form:671 +------+------+------+------+-----+-----+----+672 | xMin | xMax | yMin | yMax | CD3 | CD8 | TB | <-- As with CD3/CD8 above, TB column is for binary classification.673 +------+------+------+------+-----+-----+----+674 | .... | .... | .... | .... | ... | ... | .. |675 +------+------+------+------+-----+-----+----+676 """677 tb_data = np.concatenate((tb_data, np.zeros((tb_data.shape[0], 2)), np.ones((tb_data.shape[0], 1))), axis=1)678 lym_data = np.concatenate((lym_data, np.zeros((lym_data.shape[0], 1))), axis=1)679 mixed_data = np.concatenate((lym_data, tb_data))680 assert len(mixed_data) == len(tb_data) + len(lym_data)681 return mixed_data682# Functions for selections of data683def select_all(data, linestring, shape, d):684 return data685def select_CT(data, linestring, shape, d):686 return np.asarray(get_core_cells(data, shape, linestring, d))687def select_IM(data, linestring, shape, d):688 return np.asarray(get_margin_cells(data, linestring, d))689def select_IMCT(data, linestring, shape, d):690 margin_cells, core_cells = get_margin_core_cells(data, shape, linestring, d)691 margin_cells.extend(core_cells)692 return np.asarray(margin_cells)693def get_tb_cd3_cd8_cd3cd8_images(root_dir, meta_data_dir, target_dir_root, expected_num_files):694 """695 All input dirs should have a terminating slash696 This function assumes that select(data, linestring, shape, d) => return convert(data), i.e. ALL-DATA, no interaction with margin697 """698 # TODO: confirm that d = 500um is the correct distance from margin to allow699 d = 500700 tb_target_directory_stub = target_dir_root+"IMAGES/"701 cd3_target_directory_stub = tb_target_directory_stub702 cd8_target_directory_stub = tb_target_directory_stub703 cd3cd8_target_directory_stub = tb_target_directory_stub704 tb_options = (root_dir+"TB_pickle_files/*", expected_num_files, meta_data_dir+"*", tb_target_directory_stub, root_dir+"margin_files/*", d)705 cd3_options = ("CD3", root_dir+"lymphocyte_csv_files/*", expected_num_files, meta_data_dir+"*", cd3_target_directory_stub, root_dir+"margin_files/*", d)706 cd8_options = ("CD8", root_dir+"lymphocyte_csv_files/*", expected_num_files, meta_data_dir+"*", cd8_target_directory_stub, root_dir+"margin_files/*", d)707 cd3cd8_options = ("CD3&CD8", root_dir+"lymphocyte_csv_files/*", expected_num_files, meta_data_dir+"*", cd3cd8_target_directory_stub, root_dir+"margin_files/*", d)708 get_tb_images(*tb_options) # cluster_files, expected_num_files, meta_data_files, target_directories, margin_files, d709 get_lym_images(*cd3_options) # (lym_type, lymphocyte_files, expected_num_files, meta_data_files, target_directories, margin_files, d)710 get_lym_images(*cd8_options)711 get_lym_images(*cd3cd8_options)712def get_tb_cd3_cd8_cd3cd8_interaction_images(root_dir, target_dir_root, expected_num_files):713 """714 All input dirs should have a terminating slash715 This function assumes that select(data, linestring, shape, d) returns:716 > return convert(data) // ALL-DATA, but display with margin717 > return get_core(convert(data), linestring, shape, d) // CORE TUMOUR (CT)718 > return get_margin(convert(data), linestring, shape, d) // INVASIVE MARGIN (IM)719 > return get_margin_and_core(convert(data), linestring, shape, d) // IM & CT720 """721 # d is the maximum distance between nearby lymphocytes and any TB722 d = 50723 target_directory_stub = target_dir_root+"IMAGES/"724 tb_files = root_dir+"TB_pickle_files/*"725 lym_files = root_dir+"lymphocyte_csv_files/*"726 margin_files = root_dir + "margin_files/*"727 meta_data_files = root_dir+"meta_data/*"728 partition_func = partition_and_transform_tb_interactions729 statistics = [("_density", density_values, lambda heatmap: heatmap == 0), ("_ratio", ratio_values, lambda heatmap: heatmap == -1)]730 image_filename_func = lambda target_directory, filename: target_directory + code_extract_func(filename) + ".png"731 heatmap_filename_func = lambda target_directory, filename, statistic: target_directory + code_extract_func(filename) + statistic + ".p"732 args = (tb_files, lym_files, expected_num_files, meta_data_files, target_directory_stub, partition_func, statistics, image_filename_func, heatmap_filename_func, margin_files, d)733 get_images_with_interactions_condensed(*args)734 # ax = fig.gca(projection='3d')735 # Add margin736 # ax = plot_line_and_shape(ax, linestring, shape)737 # X_3D = np.arange(0, heatmap_tb.shape[1], 1)738 # Y_3D = np.arange(0, heatmap_tb.shape[0], 1)739 # X_3D, Y_3D = np.meshgrid(X_3D, Y_3D)740 # Z = heatmap_tb741 # # print(X_3D.shape, Y_3D.shape, Z.shape)742 # # Plot the surface.743 # surf = ax.plot_surface(X_3D, Y_3D, Z, cmap=cm.hot, linewidth=0, antialiased=False, alpha=1)744 # # # Add a color bar which maps values to colors.745 # fig.colorbar(surf, shrink=0.5, aspect=5)746 # plt.show()...
stats.py
Source:stats.py
1'''2Created on Feb 14, 20173@author: nicolas4'''5from __future__ import print_function6from functools import reduce7from lemoncheesecake.helpers.console import print_table, bold8from lemoncheesecake.cli.command import Command9from lemoncheesecake.cli.utils import load_suites_from_project10from lemoncheesecake.filter import add_test_filter_cli_args, make_test_filter11from lemoncheesecake.testtree import flatten_suites12from lemoncheesecake.project import load_project13class Stats:14 def __init__(self):15 self.tests_nb = 016 self.disabled_tests_nb = 017 self.suites_nb = 018 self.non_empty_suites_nb = 019 self.tags = {}20 self.properties = {}21 self.links = {}22def compute_stats(suites):23 stats = Stats()24 for suite in flatten_suites(suites):25 stats.suites_nb += 126 for test in suite.get_tests():27 stats.tests_nb += 128 if test.is_disabled():29 stats.disabled_tests_nb += 130 for tag in test.hierarchy_tags:31 stats.tags[tag] = stats.tags.get(tag, 0) + 132 for prop, value in test.hierarchy_properties.items():33 if prop not in stats.properties:34 stats.properties[prop] = {}35 if value not in stats.properties[prop]:36 stats.properties[prop][value] = 037 stats.properties[prop][value] += 138 for link in test.hierarchy_links:39 stats.links[link] = stats.links.get(link, 0) + 140 else:41 stats.non_empty_suites_nb += 142 return stats43class StatsCommand(Command):44 def get_name(self):45 return "stats"46 def get_description(self):47 return "Display statistics about the project's tests"48 def add_cli_args(self, cli_parser):49 add_test_filter_cli_args(cli_parser)50 def run_cmd(self, cli_args):51 project = load_project()52 suites = load_suites_from_project(project, make_test_filter(cli_args))53 stats = compute_stats(suites)54 def percent_of_tests(val):55 return "%2d%%" % (float(val) / stats.tests_nb * 100)56 # Show tags57 lines = []58 for tag in sorted(stats.tags.keys(), key=lambda k: stats.tags[k], reverse=True):59 lines.append([bold(tag), stats.tags[tag], percent_of_tests(stats.tags[tag])])60 print_table(bold("Tags"), ["Tag", "Tests", "In %"], lines)61 # Show properties62 lines = []63 prop_names = sorted(64 stats.properties.keys(),65 key=lambda k: reduce(lambda x, y: x + y, stats.properties[k].values(), 0),66 reverse=True67 )68 for prop_name in prop_names:69 prop_values = sorted(70 stats.properties[prop_name].keys(),71 key=lambda k: stats.properties[prop_name][k],72 reverse=True73 )74 for prop_value in prop_values:75 lines.append([76 bold(prop_name), bold(prop_value),77 stats.properties[prop_name][prop_value],78 percent_of_tests(stats.properties[prop_name][prop_value])79 ])80 print_table(bold("Properties"), ["Property", "Value", "Tests", "In %"], lines)81 # Show links82 lines = []83 for link in sorted(stats.links.keys(), key=lambda k: stats.links[k], reverse=True):84 lines.append([85 bold(link[1] or "-"), link[0], stats.links[link], percent_of_tests(stats.links[link])86 ])87 print_table(bold("Links"), ["Name", "URL", "Tests", "In %"], lines)88 tests_info = bold("%d tests" % stats.tests_nb)89 if stats.disabled_tests_nb > 0:90 tests_info += " (among which %s disabled tests)" % stats.disabled_tests_nb91 suites_info = bold("%d suites" % stats.non_empty_suites_nb)92 if stats.suites_nb > stats.non_empty_suites_nb:93 suites_info += " (+ %d empty suites)" % (stats.suites_nb - stats.non_empty_suites_nb)94 print("Total: %s in %s" % (tests_info, suites_info))95 print()...
margin.py
Source:margin.py
1import re2import matplotlib.pyplot as plt3from shapely.geometry import Point, Polygon, MultiPolygon4from descartes import PolygonPatch5# 2016_12_17__0690.annotations6# name = "2016_12_17__0765(Ext).annotations"7name = "2016_12_17__0690.annotations"8hierarchy = open("./Daniel line/" + name, "r").read()9p = re.compile("<*")10hierarchy_tags = p.split(hierarchy)11print(len(hierarchy_tags))12x_tags = []13y_tags = []14for tag in hierarchy_tags:15 if "X=" in tag and "Y=" in tag:16 x = tag.split("\"")[1]17 y = tag.split("\"")[3]18 x_tags.append(int(x))19 y_tags.append(int(y))20tags = []21# We have 4 points22#Â For each of 4 points:23# - is within core tumour? (core_tumour.contains(point)) // inner line24# - if no, then check: tumour_front.contains(point) or invasive_front.contains(point) // second or third lines25#26for i, j in zip(x_tags, y_tags):27 tags.append((i, j))28fig = plt.figure(1)29# 130ax = fig.add_subplot(111)31poly = Polygon(tags)32def plot_line(ax, ob):33 x, y = ob.xy34def construct_shape(poly, distance, delta, update):35 difference = 036 if type(poly.buffer(distance)) is MultiPolygon:37 difference = update(difference, delta)38 while type(poly.buffer(distance + difference)) is MultiPolygon:39 difference = update(difference, delta)40 new_poly = poly.buffer(distance + difference)41 return new_poly.buffer(-difference)42 else:43 return poly.buffer(distance)44def construct_shape_add_on_new_shape(poly, distance, delta, update):45 padding = update(0, delta)46 difference = 047 if type(poly.buffer(distance)) is MultiPolygon:48 print("Boo!")49 shape = poly50 difference = update(difference, delta)51 while type(shape.buffer(distance + padding)) is MultiPolygon:52 shape = shape.buffer(distance + padding)53 difference = update(difference, delta)54 return shape.buffer(-difference)55 else:56 return poly.buffer(distance)57# Iterative increasing approach58dilated = construct_shape(poly, 500, 1, lambda x, y: x + y)59contracted = construct_shape(dilated, -1000, 1, lambda x, y: x - y)60print(type(contracted))61print(MultiPolygon)62if type(contracted) is MultiPolygon:63 print("Uh oh. Contracted for " + name + " invalid.")64 int_x = [i[0] for po in contracted for i in po.exterior.coords]65 int_y = [i[1] for po in contracted for i in po.exterior.coords]66else:67 int_x = [i[0] for i in contracted.exterior.coords]68 int_y = [i[1] for i in contracted.exterior.coords]69if type(dilated) is MultiPolygon:70 print("Uh oh. Dilated for " + name + " invalid.")71 ext_x = [i[0] for po in dilated for i in po.exterior.coords]72 ext_y = [i[1] for po in dilated for i in po.exterior.coords]73else:74 ext_x = [i[0] for i in dilated.exterior.coords]75 ext_y = [i[1] for i in dilated.exterior.coords]76plt.plot(ext_x, ext_y, color="r")77plt.plot(int_x, int_y, color="b")78patch = PolygonPatch(poly, fc="red", alpha=0.5, zorder=2)79ax.add_patch(patch)80#Â plt.plot(x_tags, y_tags)81plt.show()...
Automation Testing Tutorials
Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.
LambdaTest Learning Hubs:
- JUnit Tutorial
- TestNG Tutorial
- Webdriver Tutorial
- WebDriverIO Tutorial
- Protractor Tutorial
- Selenium 4 Tutorial
- Jenkins Tutorial
- NUnit Tutorial
- Jest Tutorial
- Playwright Tutorial
- Cypress Tutorial
- PyTest Tutorial
YouTube
You could also refer to video tutorials over LambdaTest YouTube channel to get step by step demonstration from industry experts.
Run Lemoncheesecake automation tests on LambdaTest cloud grid
Perform automation testing on 3000+ real desktop and mobile devices online.
Try LambdaTest Now !!
Get 100 minutes of automation test minutes FREE!!
Was this article helpful?
