diff --git a/qurator/eynollah/eynollah.py b/qurator/eynollah/eynollah.py index 2bb09a1..c9e6674 100644 --- a/qurator/eynollah/eynollah.py +++ b/qurator/eynollah/eynollah.py @@ -229,6 +229,8 @@ class Eynollah: self.model_textline_dir = dir_models + "/eynollah-textline_20210425.h5" self.model_tables = dir_models + "/eynollah-tables_20210319.h5" + self.models = {} + if dir_in and light_version: config = tf.compat.v1.ConfigProto() config.gpu_options.allow_growth = True @@ -391,10 +393,6 @@ class Eynollah: prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - margin, :] = seg prediction_true = prediction_true.astype(int) - session_enhancement.close() - del model_enhancement - del session_enhancement - gc.collect() return prediction_true @@ -500,13 +498,6 @@ class Eynollah: num_col = np.argmax(label_p_pred[0]) + 1 self.logger.info("Found %s columns (%s)", num_col, label_p_pred) - if not self.dir_in: - session_col_classifier.close() - - del model_num_classifier - del session_col_classifier - K.clear_session() - gc.collect() @@ -537,12 +528,6 @@ class Eynollah: prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) - if not self.dir_in: - session_bin.close() - del model_bin - del session_bin - gc.collect() - prediction_bin = prediction_bin.astype(np.uint8) img= np.copy(prediction_bin) img_bin = np.copy(prediction_bin) @@ -579,10 +564,7 @@ class Eynollah: label_p_pred = model_num_classifier.predict(img_in, verbose=0) num_col = np.argmax(label_p_pred[0]) + 1 - self.logger.info("Found %s columns (%s)", num_col, label_p_pred) - if not self.dir_in: - session_col_classifier.close() - K.clear_session() + self.logger.info("Found %d columns (%s)", num_col, np.around(label_p_pred, decimals=5)) if dpi < DPI_THRESHOLD: img_new, num_column_is_classified = self.calculate_width_height_by_columns(img, num_col, width_early, label_p_pred) @@ -595,8 +577,6 @@ class Eynollah: num_column_is_classified = True image_res = np.copy(img) is_image_enhanced = False - if not self.dir_in: - session_col_classifier.close() self.logger.debug("exit resize_and_enhance_image_with_column_classifier") @@ -665,9 +645,14 @@ class Eynollah: def start_new_session_and_model(self, model_dir): self.logger.debug("enter start_new_session_and_model (model_dir=%s)", model_dir) - gpu_options = tf.compat.v1.GPUOptions(allow_growth=True) + #gpu_options = tf.compat.v1.GPUOptions(allow_growth=True) #gpu_options = tf.compat.v1.GPUOptions(per_process_gpu_memory_fraction=7.7, allow_growth=True) - session = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(gpu_options=gpu_options)) + #session = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(gpu_options=gpu_options)) + physical_devices = tf.config.list_physical_devices('GPU') + try: + tf.config.experimental.set_memory_growth(physical_devices[0], True) + except: + self.logger.warning("no GPU device available") # try: # model = load_model(model_dir, compile=False) @@ -676,9 +661,13 @@ class Eynollah: if model_dir.endswith('.h5') and Path(model_dir[:-3]).exists(): # prefer SavedModel over HDF5 format if it exists model_dir = model_dir[:-3] - model = load_model(model_dir, compile=False) + if model_dir in self.models: + model = self.models[model_dir] + else: + model = load_model(model_dir, compile=False) + self.models[model_dir] = model - return model, session + return model, None def do_prediction(self, patches, img, model, marginal_of_patch_percent=0.1): self.logger.debug("enter do_prediction") @@ -797,8 +786,8 @@ class Eynollah: prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - margin, :] = seg_color prediction_true = prediction_true.astype(np.uint8) - del model - gc.collect() + #del model + #gc.collect() return prediction_true def do_prediction_new_concept(self, patches, img, model, marginal_of_patch_percent=0.1): self.logger.debug("enter do_prediction") @@ -963,17 +952,19 @@ class Eynollah: prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - margin, :] = seg_color prediction_true = prediction_true.astype(np.uint8) - del model - gc.collect() + ##del model + ##gc.collect() return prediction_true def extract_page(self): self.logger.debug("enter extract_page") cont_page = [] if not self.ignore_page_extraction: + img = cv2.GaussianBlur(self.image, (5, 5), 0) + if not self.dir_in: model_page, session_page = self.start_new_session_and_model(self.model_page_dir) - img = cv2.GaussianBlur(self.image, (5, 5), 0) + if not self.dir_in: img_page_prediction = self.do_prediction(False, img, model_page) else: @@ -1003,12 +994,7 @@ class Eynollah: box = [0, 0, img.shape[1], img.shape[0]] croped_page, page_coord = crop_image_inside_box(box, self.image) cont_page.append(np.array([[page_coord[2], page_coord[0]], [page_coord[3], page_coord[0]], [page_coord[3], page_coord[1]], [page_coord[2], page_coord[1]]])) - if not self.dir_in: - session_page.close() - del model_page - del session_page - K.clear_session() - gc.collect() + self.logger.debug("exit extract_page") else: box = [0, 0, self.image.shape[1], self.image.shape[0]] @@ -1046,14 +1032,6 @@ class Eynollah: box = [0, 0, img.shape[1], img.shape[0]] croped_page, page_coord = crop_image_inside_box(box, img) - if not self.dir_in: - session_page.close() - del model_page - del session_page - K.clear_session() - - gc.collect() - self.logger.debug("exit early_page_for_num_of_column_classification") else: img = self.imread() @@ -1156,12 +1134,6 @@ class Eynollah: prediction_regions = self.do_prediction(patches, img, model_region, marginal_of_patch_percent) prediction_regions = resize_image(prediction_regions, img_height_h, img_width_h) - if not self.dir_in: - session_region.close() - del model_region - del session_region - gc.collect() - self.logger.debug("exit extract_text_regions") return prediction_regions, prediction_regions2 @@ -1558,8 +1530,6 @@ class Eynollah: prediction_textline_longshot = self.do_prediction(False, img, self.model_textline) prediction_textline_longshot_true_size = resize_image(prediction_textline_longshot, img_h, img_w) - if not self.dir_in: - session_textline.close() if self.textline_light: return (prediction_textline[:, :, 0]==1)*1, (prediction_textline_longshot_true_size[:, :, 0]==1)*1 @@ -1631,8 +1601,6 @@ class Eynollah: else: img_w_new = 4000 img_h_new = int(img_org.shape[0] / float(img_org.shape[1]) * img_w_new) - gc.collect() - ##img_resized = resize_image(img_bin,img_height_h, img_width_h ) img_resized = resize_image(img,img_h_new, img_w_new ) if not self.dir_in: @@ -1645,11 +1613,6 @@ class Eynollah: prediction_bin = prediction_bin*255 prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) - if not self.dir_in: - session_bin.close() - del model_bin - del session_bin - gc.collect() prediction_bin = prediction_bin.astype(np.uint16) #img= np.copy(prediction_bin) @@ -1695,9 +1658,6 @@ class Eynollah: text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_texts, color=(1,1,1)) - #erosion_hurts = True - if not self.dir_in: - K.clear_session() return text_regions_p_true, erosion_hurts, polygons_lines_xml, textline_mask_tot_ea def get_regions_from_xy_2models(self,img,is_image_enhanced, num_col_classifier): @@ -1742,16 +1702,9 @@ class Eynollah: prediction_regions_org = self.do_prediction(True, img, model_region) prediction_regions_org = resize_image(prediction_regions_org, img_height_h, img_width_h ) - ##plt.imshow(prediction_regions_org[:,:,0]) - ##plt.show() prediction_regions_org=prediction_regions_org[:,:,0] prediction_regions_org[(prediction_regions_org[:,:]==1) & (mask_zeros_y[:,:]==1)]=0 - if not self.dir_in: - session_region.close() - del model_region - del session_region - gc.collect() if not self.dir_in: model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p2) @@ -1763,11 +1716,6 @@ class Eynollah: prediction_regions_org2 = self.do_prediction(True, img, model_region, 0.2) prediction_regions_org2=resize_image(prediction_regions_org2, img_height_h, img_width_h ) - if not self.dir_in: - session_region.close() - del model_region - del session_region - gc.collect() mask_zeros2 = (prediction_regions_org2[:,:,0] == 0) mask_lines2 = (prediction_regions_org2[:,:,0] == 3) @@ -1788,8 +1736,6 @@ class Eynollah: mask_lines_only=(prediction_regions_org[:,:]==3)*1 prediction_regions_org = cv2.erode(prediction_regions_org[:,:], KERNEL, iterations=2) - #plt.imshow(text_region2_1st_channel) - #plt.show() prediction_regions_org = cv2.dilate(prediction_regions_org[:,:], KERNEL, iterations=2) @@ -1811,11 +1757,6 @@ class Eynollah: prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) - if not self.dir_in: - session_bin.close() - del model_bin - del session_bin - gc.collect() if not self.dir_in: @@ -1834,11 +1775,6 @@ class Eynollah: prediction_regions_org=prediction_regions_org[:,:,0] mask_lines_only=(prediction_regions_org[:,:]==3)*1 - if not self.dir_in: - session_region.close() - del model_region - del session_region - gc.collect() mask_texts_only=(prediction_regions_org[:,:]==1)*1 @@ -1859,19 +1795,11 @@ class Eynollah: text_regions_p_true=cv2.fillPoly(text_regions_p_true,pts=polygons_of_only_texts, color=(1,1,1)) - if not self.dir_in: - K.clear_session() return text_regions_p_true, erosion_hurts, polygons_lines_xml except: if self.input_binary: prediction_bin = np.copy(img_org) - else: - if not self.dir_in: - session_region.close() - del model_region - del session_region - gc.collect() if not self.dir_in: model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization) @@ -1887,12 +1815,6 @@ class Eynollah: prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) - - if not self.dir_in: - session_bin.close() - del model_bin - del session_bin - gc.collect() if not self.dir_in: @@ -1910,11 +1832,6 @@ class Eynollah: prediction_regions_org=prediction_regions_org[:,:,0] #mask_lines_only=(prediction_regions_org[:,:]==3)*1 - if not self.dir_in: - session_region.close() - del model_region - del session_region - gc.collect() #img = resize_image(img_org, int(img_org.shape[0]*1), int(img_org.shape[1]*1)) @@ -1925,12 +1842,6 @@ class Eynollah: #prediction_regions_org = prediction_regions_org[:,:,0] #prediction_regions_org[(prediction_regions_org[:,:] == 1) & (mask_zeros_y[:,:] == 1)]=0 - #session_region.close() - #del model_region - #del session_region - #gc.collect() - - mask_lines_only = (prediction_regions_org[:,:] ==3)*1 @@ -1957,8 +1868,6 @@ class Eynollah: text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_texts, color=(1,1,1)) erosion_hurts = True - if not self.dir_in: - K.clear_session() return text_regions_p_true, erosion_hurts, polygons_lines_xml def do_order_of_regions_full_layout(self, contours_only_text_parent, contours_only_text_parent_h, boxes, textline_mask_tot): @@ -2515,10 +2424,6 @@ class Eynollah: prediction_table_erode = cv2.erode(prediction_table[:,:,0], KERNEL, iterations=20) prediction_table_erode = cv2.dilate(prediction_table_erode, KERNEL, iterations=20) - - del model_region - del session_region - gc.collect() return prediction_table_erode.astype(np.int16) @@ -2619,8 +2524,7 @@ class Eynollah: self.logger.info("Resizing and enhancing image...") is_image_enhanced, img_org, img_res, num_col_classifier, num_column_is_classified, img_bin = self.resize_and_enhance_image_with_column_classifier(light_version) self.logger.info("Image was %senhanced.", '' if is_image_enhanced else 'not ') - if not self.dir_in: - K.clear_session() + scale = 1 if is_image_enhanced: if self.allow_enhancement: @@ -2646,8 +2550,6 @@ class Eynollah: textline_mask_tot_ea, _ = self.textline_contours(image_page, True, scaler_h_textline, scaler_w_textline) if self.textline_light: textline_mask_tot_ea = textline_mask_tot_ea.astype(np.int16) - if not self.dir_in: - K.clear_session() if self.plotter: self.plotter.save_plot_of_textlines(textline_mask_tot_ea, image_page) return textline_mask_tot_ea @@ -2660,7 +2562,7 @@ class Eynollah: if self.plotter: self.plotter.save_deskewed_image(slope_deskew) - self.logger.info("slope_deskew: %s", slope_deskew) + self.logger.info("slope_deskew: %.2f°", slope_deskew) return slope_deskew, slope_first def run_marginals(self, image_page, textline_mask_tot_ea, mask_images, mask_lines, num_col_classifier, slope_deskew, text_regions_p_1, table_prediction): @@ -2709,8 +2611,6 @@ class Eynollah: if np.abs(slope_deskew) >= SLOPE_THRESHOLD: _, _, matrix_of_lines_ch_d, splitter_y_new_d, _ = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2), num_col_classifier, self.tables, pixel_lines) - if not self.dir_in: - K.clear_session() self.logger.info("num_col_classifier: %s", num_col_classifier) @@ -2775,8 +2675,6 @@ class Eynollah: pixel_img = 10 contours_tables = return_contours_of_interested_region(text_regions_p, pixel_img, min_area_mar) - if not self.dir_in: - K.clear_session() self.logger.debug('exit run_boxes_no_full_layout') return polygons_of_images, img_revised_tab, text_regions_p_1_n, textline_mask_tot_d, regions_without_separators_d, boxes, boxes_d, polygons_of_marginals, contours_tables @@ -2807,9 +2705,6 @@ class Eynollah: if np.abs(slope_deskew) >= SLOPE_THRESHOLD: num_col_d, peaks_neg_fin_d, matrix_of_lines_ch_d, splitter_y_new_d, seperators_closeup_n_d = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2),num_col_classifier, self.tables, pixel_lines) - if not self.dir_in: - K.clear_session() - gc.collect() if num_col_classifier>=3: if np.abs(slope_deskew) < SLOPE_THRESHOLD: @@ -2875,38 +2770,22 @@ class Eynollah: text_regions_p[:, :][text_regions_p[:, :] == 2] = 5 text_regions_p[:, :][text_regions_p[:, :] == 3] = 6 text_regions_p[:, :][text_regions_p[:, :] == 4] = 8 - if not self.dir_in: - K.clear_session() + image_page = image_page.astype(np.uint8) regions_fully, regions_fully_only_drop = self.extract_text_regions(image_page, True, cols=num_col_classifier) text_regions_p[:,:][regions_fully[:,:,0]==6]=6 regions_fully_only_drop = put_drop_out_from_only_drop_model(regions_fully_only_drop, text_regions_p) regions_fully[:, :, 0][regions_fully_only_drop[:, :, 0] == 4] = 4 - if not self.dir_in: - K.clear_session() - # plt.imshow(regions_fully[:,:,0]) - # plt.show() regions_fully = putt_bb_of_drop_capitals_of_model_in_patches_in_layout(regions_fully) - # plt.imshow(regions_fully[:,:,0]) - # plt.show() - if not self.dir_in: - K.clear_session() + regions_fully_np, _ = self.extract_text_regions(image_page, False, cols=num_col_classifier) - # plt.imshow(regions_fully_np[:,:,0]) - # plt.show() if num_col_classifier > 2: regions_fully_np[:, :, 0][regions_fully_np[:, :, 0] == 4] = 0 else: regions_fully_np = filter_small_drop_capitals_from_no_patch_layout(regions_fully_np, text_regions_p) - # plt.imshow(regions_fully_np[:,:,0]) - # plt.show() - if not self.dir_in: - K.clear_session() - # plt.imshow(regions_fully[:,:,0]) - # plt.show() regions_fully = boosting_headers_by_longshot_region_segmentation(regions_fully, regions_fully_np, img_only_regions) # plt.imshow(regions_fully[:,:,0]) # plt.show() @@ -2929,8 +2808,6 @@ class Eynollah: regions_without_separators_d = None if not self.tables: regions_without_separators = (text_regions_p[:, :] == 1) * 1 - if not self.dir_in: - K.clear_session() img_revised_tab = np.copy(text_regions_p[:, :]) polygons_of_images = return_contours_of_interested_region(img_revised_tab, 5) self.logger.debug('exit run_boxes_full_layout') @@ -3025,13 +2902,12 @@ class Eynollah: contours_only_text_parent = return_parent_contours(contours_only_text, hir_on_text) if len(contours_only_text_parent) > 0: - areas_cnt_text = np.array([cv2.contourArea(contours_only_text_parent[j]) for j in range(len(contours_only_text_parent))]) + areas_cnt_text = np.array([cv2.contourArea(c) for c in contours_only_text_parent]) areas_cnt_text = areas_cnt_text / float(text_only.shape[0] * text_only.shape[1]) #self.logger.info('areas_cnt_text %s', areas_cnt_text) contours_biggest = contours_only_text_parent[np.argmax(areas_cnt_text)] - contours_only_text_parent = [contours_only_text_parent[jz] for jz in range(len(contours_only_text_parent)) if areas_cnt_text[jz] > min_con_area] - areas_cnt_text_parent = [areas_cnt_text[jz] for jz in range(len(areas_cnt_text)) if areas_cnt_text[jz] > min_con_area] - + contours_only_text_parent = [c for jz, c in enumerate(contours_only_text_parent) if areas_cnt_text[jz] > min_con_area] + areas_cnt_text_parent = [area for area in areas_cnt_text if area > min_con_area] index_con_parents = np.argsort(areas_cnt_text_parent) contours_only_text_parent = list(np.array(contours_only_text_parent)[index_con_parents]) areas_cnt_text_parent = list(np.array(areas_cnt_text_parent)[index_con_parents]) @@ -3042,14 +2918,14 @@ class Eynollah: contours_only_text_d, hir_on_text_d = return_contours_of_image(text_only_d) contours_only_text_parent_d = return_parent_contours(contours_only_text_d, hir_on_text_d) - areas_cnt_text_d = np.array([cv2.contourArea(contours_only_text_parent_d[j]) for j in range(len(contours_only_text_parent_d))]) + areas_cnt_text_d = np.array([cv2.contourArea(c) for c in contours_only_text_parent_d]) areas_cnt_text_d = areas_cnt_text_d / float(text_only_d.shape[0] * text_only_d.shape[1]) if len(areas_cnt_text_d)>0: contours_biggest_d = contours_only_text_parent_d[np.argmax(areas_cnt_text_d)] - index_con_parents_d=np.argsort(areas_cnt_text_d) - contours_only_text_parent_d=list(np.array(contours_only_text_parent_d)[index_con_parents_d] ) - areas_cnt_text_d=list(np.array(areas_cnt_text_d)[index_con_parents_d] ) + index_con_parents_d = np.argsort(areas_cnt_text_d) + contours_only_text_parent_d = list(np.array(contours_only_text_parent_d)[index_con_parents_d]) + areas_cnt_text_d = list(np.array(areas_cnt_text_d)[index_con_parents_d]) cx_bigest_d_big, cy_biggest_d_big, _, _, _, _, _ = find_new_features_of_contours([contours_biggest_d]) cx_bigest_d, cy_biggest_d, _, _, _, _, _ = find_new_features_of_contours(contours_only_text_parent_d) @@ -3103,12 +2979,12 @@ class Eynollah: contours_only_text_parent = return_parent_contours(contours_only_text, hir_on_text) if len(contours_only_text_parent) > 0: - areas_cnt_text = np.array([cv2.contourArea(contours_only_text_parent[j]) for j in range(len(contours_only_text_parent))]) + areas_cnt_text = np.array([cv2.contourArea(c) for c in contours_only_text_parent]) areas_cnt_text = areas_cnt_text / float(text_only.shape[0] * text_only.shape[1]) contours_biggest = contours_only_text_parent[np.argmax(areas_cnt_text)] - contours_only_text_parent = [contours_only_text_parent[jz] for jz in range(len(contours_only_text_parent)) if areas_cnt_text[jz] > min_con_area] - areas_cnt_text_parent = [areas_cnt_text[jz] for jz in range(len(areas_cnt_text)) if areas_cnt_text[jz] > min_con_area] + contours_only_text_parent = [c for jz, c in enumerate(contours_only_text_parent) if areas_cnt_text[jz] > min_con_area] + areas_cnt_text_parent = [area for area in areas_cnt_text if area > min_con_area] index_con_parents = np.argsort(areas_cnt_text_parent) contours_only_text_parent = list(np.array(contours_only_text_parent)[index_con_parents]) @@ -3146,8 +3022,6 @@ class Eynollah: all_found_texline_polygons = small_textlines_to_parent_adherence2(all_found_texline_polygons, textline_mask_tot_ea, num_col_classifier) all_found_texline_polygons_marginals, boxes_marginals, _, polygons_of_marginals, all_box_coord_marginals, _, slopes_marginals = self.get_slopes_and_deskew_new_curved(polygons_of_marginals, polygons_of_marginals, cv2.erode(textline_mask_tot_ea, kernel=KERNEL, iterations=1), image_page_rotated, boxes_marginals, text_only, num_col_classifier, scale_param, slope_deskew) all_found_texline_polygons_marginals = small_textlines_to_parent_adherence2(all_found_texline_polygons_marginals, textline_mask_tot_ea, num_col_classifier) - if not self.dir_in: - K.clear_session() if self.full_layout: if np.abs(slope_deskew) >= SLOPE_THRESHOLD: @@ -3167,8 +3041,6 @@ class Eynollah: if self.plotter: self.plotter.save_plot_of_layout(text_regions_p, image_page) self.plotter.save_plot_of_layout_all(text_regions_p, image_page) - if not self.dir_in: - K.clear_session() pixel_img = 4 polygons_of_drop_capitals = return_contours_of_interested_region_by_min_size(text_regions_p, pixel_img)