import io import logging import math import os import re import xml.dom import xml.dom.minidom from collections import defaultdict from copy import deepcopy from datetime import datetime from enum import Enum from glob import glob from shutil import copy2 from typing import Optional, List, Tuple import ijson import ujson as json from PIL import Image from label_studio_sdk.converter import brush from label_studio_sdk.converter.audio import convert_to_asr_json_manifest from label_studio_sdk.converter.keypoints import process_keypoints_for_coco, build_kp_order, update_categories_for_keypoints, keypoints_in_label_config, get_yolo_categories_for_keypoints from label_studio_sdk.converter.exports import csv2 from label_studio_sdk.converter.utils import ( parse_config, create_tokens_and_tags, download, get_image_size_and_channels, ensure_dir, get_polygon_area, get_polygon_bounding_box, get_annotator, get_json_root_type, prettify_result, convert_annotation_to_yolo, convert_annotation_to_yolo_obb, ) from label_studio_sdk._extensions.label_studio_tools.core.utils.io import get_local_path from label_studio_sdk.converter.exports.yolo import process_and_save_yolo_annotations logger = logging.getLogger(__name__) class FormatNotSupportedError(NotImplementedError): pass class Format(Enum): JSON = 1 JSON_MIN = 2 CSV = 3 TSV = 4 CONLL2003 = 5 COCO = 6 VOC = 7 BRUSH_TO_NUMPY = 8 BRUSH_TO_PNG = 9 ASR_MANIFEST = 10 YOLO = 11 YOLO_OBB = 12 CSV_OLD = 13 YOLO_WITH_IMAGES = 14 COCO_WITH_IMAGES = 15 YOLO_OBB_WITH_IMAGES = 16 BRUSH_TO_COCO = 17 def __str__(self): return self.name @classmethod def from_string(cls, s): try: return Format[s] except KeyError: raise ValueError() class Converter(object): _FORMAT_INFO = { Format.JSON: { "title": "JSON", "description": "List of items in raw JSON format stored in one JSON file. Use to export both the data " "and the annotations for a dataset. It's Label Studio Common Format", "link": "https://labelstud.io/guide/export.html#JSON", }, Format.JSON_MIN: { "title": "JSON-MIN", "description": 'List of items where only "from_name", "to_name" values from the raw JSON format are ' "exported. Use to export only the annotations for a dataset.", "link": "https://labelstud.io/guide/export.html#JSON-MIN", }, Format.CSV: { "title": "CSV", "description": "Results are stored as comma-separated values with the column names specified by the " 'values of the "from_name" and "to_name" fields.', "link": "https://labelstud.io/guide/export.html#CSV", }, Format.TSV: { "title": "TSV", "description": "Results are stored in tab-separated tabular file with column names specified by " '"from_name" "to_name" values', "link": "https://labelstud.io/guide/export.html#TSV", }, Format.CONLL2003: { "title": "CONLL2003", "description": "Popular format used for the CoNLL-2003 named entity recognition challenge.", "link": "https://labelstud.io/guide/export.html#CONLL2003", "tags": ["sequence labeling", "text tagging", "named entity recognition"], }, Format.COCO: { "title": "COCO", "description": "Popular machine learning format used by the COCO dataset for object detection and image " "segmentation tasks with polygons and rectangles.", "link": "https://labelstud.io/guide/export.html#COCO", "tags": ["image segmentation", "object detection", "keypoints"], }, Format.COCO_WITH_IMAGES: { "title": "COCO with Images", "description": "COCO format with images downloaded.", "link": "https://labelstud.io/guide/export.html#COCO", "tags": ["image segmentation", "object detection", "keypoints"], }, Format.VOC: { "title": "Pascal VOC XML", "description": "Popular XML format used for object detection and polygon image segmentation tasks.", "link": "https://labelstud.io/guide/export.html#Pascal-VOC-XML", "tags": ["image segmentation", "object detection"], }, Format.YOLO: { "title": "YOLO", "description": "Popular TXT format is created for each image file. Each txt file contains annotations for " "the corresponding image file, that is object class, object coordinates, height & width.", "link": "https://labelstud.io/guide/export.html#YOLO", "tags": ["image segmentation", "object detection", "keypoints"], }, Format.YOLO_WITH_IMAGES: { "title": "YOLO with Images", "description": "YOLO format with images downloaded.", "link": "https://labelstud.io/guide/export.html#YOLO", "tags": ["image segmentation", "object detection", "keypoints"], }, Format.YOLO_OBB: { "title": "YOLOv8 OBB", "description": "Popular TXT format is created for each image file. Each txt file contains annotations for " "the corresponding image file. The YOLO OBB format designates bounding boxes by their four corner points " "with coordinates normalized between 0 and 1, so it is possible to export rotated objects.", "link": "https://labelstud.io/guide/export.html#YOLO", "tags": ["image segmentation", "object detection"], }, Format.YOLO_OBB_WITH_IMAGES: { "title": "YOLOv8 OBB with Images", "description": "YOLOv8 OBB format with images downloaded.", "link": "https://labelstud.io/guide/export.html#YOLO", "tags": ["image segmentation", "object detection"], }, Format.BRUSH_TO_NUMPY: { "title": "Brush labels to NumPy", "description": "Export your brush labels as NumPy 2d arrays. Each label outputs as one image.", "link": "https://labelstud.io/guide/export.html#Brush-labels-to-NumPy-amp-PNG", "tags": ["image segmentation"], }, Format.BRUSH_TO_PNG: { "title": "Brush labels to PNG", "description": "Export your brush labels as PNG images. Each label outputs as one image.", "link": "https://labelstud.io/guide/export.html#Brush-labels-to-NumPy-amp-PNG", "tags": ["image segmentation"], }, Format.ASR_MANIFEST: { "title": "ASR Manifest", "description": "Export audio transcription labels for automatic speech recognition as the JSON manifest " "format expected by NVIDIA NeMo models.", "link": "https://labelstud.io/guide/export.html#ASR-MANIFEST", "tags": ["speech recognition"], }, Format.BRUSH_TO_COCO: { "title": "Brush labels to COCO", "description": "Export your brush labels as COCO format for segmentation tasks. Converts RLE encoded masks to COCO polygons.", "link": "https://labelstud.io/guide/export.html#COCO", "tags": ["image segmentation", "brush annotations"], }, } def all_formats(self): return self._FORMAT_INFO def __init__( self, config, project_dir, output_tags=None, upload_dir=None, download_resources=True, access_token=None, hostname=None, ): """Initialize Label Studio Converter for Exports :param config: string or dict: XML string with Label studio labeling config or path to this file or parsed_config :param project_dir: upload root directory for images, audio and other labeling files :param output_tags: it will be calculated automatically, contains label names :param upload_dir: upload root directory with files that were imported using LS GUI :param download_resources: if True, LS will try to download images, audio, etc and include them to export """ self.project_dir = project_dir self.upload_dir = upload_dir self.download_resources = download_resources self._schema = None self.access_token = access_token self.hostname = hostname self.is_keypoints = None if isinstance(config, dict): self._schema = config elif isinstance(config, str): if os.path.isfile(config): with io.open(config) as f: config_string = f.read() else: config_string = config self._schema = parse_config(config_string) if self._schema is None: logger.warning( "Label config or schema for Converter is not provided, " "it might be critical for some export formats, now set schema to empty dict" ) self._schema = {} self._data_keys, self._output_tags = self._get_data_keys_and_output_tags( output_tags ) self._supported_formats = self._get_supported_formats() def convert(self, input_data, output_data, format, is_dir=True, **kwargs): if isinstance(format, str): format = Format.from_string(format) if format == Format.JSON: self.convert_to_json(input_data, output_data, is_dir=is_dir) elif format == Format.JSON_MIN: self.convert_to_json_min(input_data, output_data, is_dir=is_dir) elif format == Format.CSV: header = kwargs.get("csv_header", True) sep = kwargs.get("csv_separator", ",") self.convert_to_csv( input_data, output_data, sep=sep, header=header, is_dir=is_dir ) elif format == Format.TSV: header = kwargs.get("csv_header", True) sep = kwargs.get("csv_separator", "\t") self.convert_to_csv( input_data, output_data, sep=sep, header=header, is_dir=is_dir ) elif format == Format.CONLL2003: self.convert_to_conll2003(input_data, output_data, is_dir=is_dir) elif format in [Format.COCO, Format.COCO_WITH_IMAGES]: image_dir = kwargs.get("image_dir") self.download_resources = format == Format.COCO_WITH_IMAGES self.convert_to_coco( input_data, output_data, output_image_dir=image_dir, is_dir=is_dir ) elif format in [Format.YOLO, Format.YOLO_OBB, Format.YOLO_OBB_WITH_IMAGES, Format.YOLO_WITH_IMAGES]: image_dir = kwargs.get("image_dir") label_dir = kwargs.get("label_dir") self.download_resources = format in [Format.YOLO_WITH_IMAGES, Format.YOLO_OBB_WITH_IMAGES] self.convert_to_yolo( input_data, output_data, output_image_dir=image_dir, output_label_dir=label_dir, is_dir=is_dir, is_obb=(format in [Format.YOLO_OBB, Format.YOLO_OBB_WITH_IMAGES]), ) elif format == Format.VOC: image_dir = kwargs.get("image_dir") self.convert_to_voc( input_data, output_data, output_image_dir=image_dir, is_dir=is_dir ) elif format == Format.BRUSH_TO_NUMPY: items = ( self.iter_from_dir(input_data) if is_dir else self.iter_from_json_file(input_data) ) brush.convert_task_dir(items, output_data, out_format="numpy") elif format == Format.BRUSH_TO_PNG: items = ( self.iter_from_dir(input_data) if is_dir else self.iter_from_json_file(input_data) ) brush.convert_task_dir(items, output_data, out_format="png") elif format == Format.ASR_MANIFEST: items = ( self.iter_from_dir(input_data) if is_dir else self.iter_from_json_file(input_data) ) convert_to_asr_json_manifest( items, output_data, data_key=self._data_keys[0], project_dir=self.project_dir, upload_dir=self.upload_dir, download_resources=self.download_resources, ) elif format == Format.BRUSH_TO_COCO: items = ( self.iter_from_dir(input_data) if is_dir else self.iter_from_json_file(input_data) ) from label_studio_sdk.converter.exports.brush_to_coco import convert_to_coco image_dir = kwargs.get("image_dir") convert_to_coco( items, output_data, output_image_dir=image_dir ) def _get_data_keys_and_output_tags(self, output_tags=None): data_keys = set() output_tag_names = [] if output_tags is not None: for tag in output_tags: if tag not in self._schema: logger.debug( 'Specified tag "{tag}" not found in config schema: ' "available options are {schema_keys}".format( tag=tag, schema_keys=str(list(self._schema.keys())) ) ) for name, info in self._schema.items(): if output_tags is not None and name not in output_tags: continue for input_tag in info["inputs"]: for value_key_name in ["value", "valueList"]: if value_key_name in input_tag: data_keys.add(input_tag[value_key_name]) output_tag_names.append(name) return list(data_keys), output_tag_names def _get_supported_formats(self): is_mig = False if len(self._data_keys) > 1: return [ Format.JSON.name, Format.JSON_MIN.name, Format.CSV.name, Format.TSV.name, ] output_tag_types = set() input_tag_types = set() for info in self._schema.values(): output_tag_types.add(info["type"]) for input_tag in info["inputs"]: if input_tag.get("valueList"): is_mig = True if input_tag["type"] == "Text" and input_tag.get("valueType") == "url": logger.error('valueType="url" are not supported for text inputs') continue input_tag_types.add(input_tag["type"]) all_formats = [f.name for f in Format] # Check if KeyPointLabels exists without RectangleLabels has_keypoint_labels = "KeyPointLabels" in output_tag_types has_rectangle_labels = "RectangleLabels" in output_tag_types # If config has KeyPointLabels but no RectangleLabels, exclude COCO and YOLO formats # RectangleLabels are required for COCO and YOLO formats for category mapping if has_keypoint_labels and not has_rectangle_labels: if Format.COCO.name in all_formats: all_formats.remove(Format.COCO.name) if Format.COCO_WITH_IMAGES.name in all_formats: all_formats.remove(Format.COCO_WITH_IMAGES.name) if Format.YOLO.name in all_formats: all_formats.remove(Format.YOLO.name) if Format.YOLO_WITH_IMAGES.name in all_formats: all_formats.remove(Format.YOLO_WITH_IMAGES.name) if not ("Text" in input_tag_types and "Labels" in output_tag_types): all_formats.remove(Format.CONLL2003.name) if is_mig or not ( "Image" in input_tag_types and ( "RectangleLabels" in output_tag_types or "Rectangle" in output_tag_types and "Labels" in output_tag_types ) ): all_formats.remove(Format.VOC.name) if is_mig or not ( "Image" in input_tag_types and ( "RectangleLabels" in output_tag_types or "PolygonLabels" in output_tag_types or "KeyPointLabels" in output_tag_types ) or "Rectangle" in output_tag_types and "Labels" in output_tag_types or "PolygonLabels" in output_tag_types and "Labels" in output_tag_types or "KeyPointLabels" in output_tag_types and "Labels" in output_tag_types ): all_formats.remove(Format.COCO.name) all_formats.remove(Format.COCO_WITH_IMAGES.name) all_formats.remove(Format.YOLO.name) all_formats.remove(Format.YOLO_WITH_IMAGES.name) if Format.YOLO_OBB.name in all_formats: all_formats.remove(Format.YOLO_OBB.name) if Format.YOLO_OBB_WITH_IMAGES.name in all_formats: all_formats.remove(Format.YOLO_OBB_WITH_IMAGES.name) if not ( "Image" in input_tag_types and ( "BrushLabels" in output_tag_types or "brushlabels" in output_tag_types or "Brush" in output_tag_types and "Labels" in output_tag_types ) ): all_formats.remove(Format.BRUSH_TO_NUMPY.name) all_formats.remove(Format.BRUSH_TO_PNG.name) all_formats.remove(Format.BRUSH_TO_COCO.name) if not ( ("Audio" in input_tag_types or "AudioPlus" in input_tag_types) and "TextArea" in output_tag_types ): all_formats.remove(Format.ASR_MANIFEST.name) if is_mig or ('Video' in input_tag_types and 'TimelineLabels' in output_tag_types): if Format.YOLO_OBB.name in all_formats: all_formats.remove(Format.YOLO_OBB.name) if Format.YOLO_OBB_WITH_IMAGES.name in all_formats: all_formats.remove(Format.YOLO_OBB_WITH_IMAGES.name) return all_formats @property def supported_formats(self): return self._supported_formats def iter_from_dir(self, input_dir): if not os.path.exists(input_dir): raise FileNotFoundError( "{input_dir} doesn't exist".format(input_dir=input_dir) ) for json_file in glob(os.path.join(input_dir, "*.json")): for item in self.iter_from_json_file(json_file): if item: yield item def iter_from_json_file(self, json_file): """Extract annotation results from json file param json_file: path to task list or dict with annotations """ data_type = get_json_root_type(json_file) # one task if data_type == "dict": with open(json_file, "r") as json_file: data = json.load(json_file) for item in self.annotation_result_from_task(data): yield item # many tasks elif data_type == "list": with io.open(json_file, "rb") as f: logger.debug(f"ijson backend in use: {ijson.backend}") data = ijson.items( f, "item", use_float=True ) # 'item' means to read array of dicts for task in data: for item in self.annotation_result_from_task(task): if item is not None: yield item def _maybe_matching_tag_from_schema(self, from_name: str) -> Optional[str]: """If the from name exactly matches an output tag from the schema, return that tag. Otherwise, certain tags (like those from Repeater) contain placeholders like {{idx}}. Such placeholders are mapped to a regex in self._schema. For example, if "my_output_tag_{{idx}}" is a tag in the schema, then the from_name "my_output_tag_0" should match it, and we should return "my_output_tag_{{idx}}". """ best_match = None best_match_len = 0 for tag_name, tag_info in self._schema.items(): if tag_name == from_name: return tag_name if not tag_info.get("regex"): continue tag_name_pattern = tag_name for variable, regex in tag_info["regex"].items(): tag_name_pattern = tag_name_pattern.replace(variable, regex) # In some cases there are tags with same prefix - we need to find the best or longest matching pattern if r := re.compile(tag_name_pattern).match(from_name): if match_len := len(tag_name_pattern) > best_match_len: best_match = tag_name best_match_len = len(tag_name_pattern) return best_match if best_match else None def annotation_result_from_task(self, task): has_annotations = "completions" in task or "annotations" in task if not has_annotations: logger.warning( 'Each task dict item should contain "annotations" or "completions" [deprecated], ' "where value is list of dicts" ) return None # get last not skipped completion and make result from it annotations = ( task["annotations"] if "annotations" in task else task["completions"] ) # return task with empty annotations if not annotations: data = Converter.get_data(task, {}, {}) yield data # skip cancelled annotations cancelled = lambda x: not ( x.get("skipped", False) or x.get("was_cancelled", False) ) annotations = list(filter(cancelled, annotations)) if not annotations: return None # sort by creation time annotations = sorted( annotations, key=lambda x: x.get("created_at", 0), reverse=True ) for annotation in annotations: result = annotation.get("result") or [] outputs = defaultdict(list) # get results only as output for r in result: from_name = r.get("from_name") tag_name = ( self._maybe_matching_tag_from_schema(from_name) if from_name is not None else None ) if from_name and tag_name: v = deepcopy(r["value"]) v["type"] = self._schema[tag_name]["type"] if "original_width" in r: v["original_width"] = r["original_width"] if "original_height" in r: v["original_height"] = r["original_height"] outputs[from_name].append(v) if self.is_keypoints: v['id'] = r.get('id') v['parentID'] = r.get('parentID') elif from_name and r.get("type") == "chatmessage": v = deepcopy(r.get("value", {})) v["type"] = "chatmessage" outputs[from_name].append(v) else: pass data = Converter.get_data(task, outputs, annotation) if "agreement" in task: data["agreement"] = task["agreement"] yield data @staticmethod def get_data(task, outputs, annotation): return { "id": task["id"], "input": task["data"], "output": outputs or {}, "completed_by": annotation.get("completed_by", {}), "annotation_id": annotation.get("id"), "created_at": annotation.get("created_at"), "updated_at": annotation.get("updated_at"), "lead_time": annotation.get("lead_time"), "history": annotation.get("history"), "was_cancelled": annotation.get("was_cancelled"), } def _check_format(self, fmt): pass def convert_to_json(self, input_data, output_dir, is_dir=True): self._check_format(Format.JSON) ensure_dir(output_dir) output_file = os.path.join(output_dir, "result.json") if is_dir: # Memory-optimized: stream JSON writing instead of accumulating in memory with io.open(output_file, mode="w", encoding="utf8") as fout: fout.write("[\n") first_record = True for json_file in glob(os.path.join(input_data, "*.json")): with io.open(json_file, encoding="utf8") as f: record = json.load(f) if not first_record: fout.write(",\n") json.dump(record, fout, indent=2, ensure_ascii=False) first_record = False # Free memory immediately del record fout.write("\n]") else: copy2(input_data, output_file) def convert_to_json_min(self, input_data, output_dir, is_dir=True): self._check_format(Format.JSON_MIN) ensure_dir(output_dir) output_file = os.path.join(output_dir, "result.json") item_iterator = self.iter_from_dir if is_dir else self.iter_from_json_file with io.open(output_file, mode="w", encoding="utf8") as fout: fout.write("[\n") first_record = True for item in item_iterator(input_data): # SAFE memory optimization: use json serialization/deserialization # This avoids deepcopy but ensures complete isolation of objects record = json.loads(json.dumps(item["input"])) if item.get("id") is not None: record["id"] = item["id"] for name, value in item["output"].items(): record[name] = prettify_result(value) record["annotator"] = get_annotator(item, int_id=True) record["annotation_id"] = item["annotation_id"] record["created_at"] = item["created_at"] record["updated_at"] = item["updated_at"] record["lead_time"] = item["lead_time"] if "agreement" in item: record["agreement"] = item["agreement"] # Write record to file immediately if not first_record: fout.write(",\n") json.dump(record, fout, indent=2, ensure_ascii=False) first_record = False # Explicitly delete record to free memory del record fout.write("\n]") def convert_to_csv(self, input_data, output_dir, is_dir=True, **kwargs): self._check_format(Format.CSV) item_iterator = self.iter_from_dir if is_dir else self.iter_from_json_file return csv2.convert(item_iterator, input_data, output_dir, **kwargs) def convert_to_conll2003(self, input_data, output_dir, is_dir=True): self._check_format(Format.CONLL2003) ensure_dir(output_dir) output_file = os.path.join(output_dir, "result.conll") data_key = self._data_keys[0] with io.open(output_file, "w", encoding="utf8") as fout: fout.write("-DOCSTART- -X- O\n") item_iterator = self.iter_from_dir if is_dir else self.iter_from_json_file for item in item_iterator(input_data): filtered_output = list( filter( lambda x: x[0]["type"].lower() == "labels", item["output"].values(), ) ) tokens, tags = create_tokens_and_tags( text=item["input"][data_key], spans=next(iter(filtered_output), None), ) for token, tag in zip(tokens, tags): fout.write("{token} -X- _ {tag}\n".format(token=token, tag=tag)) fout.write("\n") def convert_to_coco( self, input_data, output_dir, output_image_dir=None, is_dir=True ): def add_image(images, width, height, image_id, image_path): images.append( { "width": width, "height": height, "id": image_id, "file_name": image_path, } ) return images self._check_format(Format.COCO) ensure_dir(output_dir) output_file = os.path.join(output_dir, "result.json") if output_image_dir is not None: ensure_dir(output_image_dir) else: output_image_dir = os.path.join(output_dir, "images") os.makedirs(output_image_dir, exist_ok=True) images, categories, annotations = [], [], [] categories, category_name_to_id = self._get_labels() categories, category_name_to_id = update_categories_for_keypoints(categories, category_name_to_id, self._schema) data_key = self._data_keys[0] item_iterator = ( self.iter_from_dir(input_data) if is_dir else self.iter_from_json_file(input_data) ) for item_idx, item in enumerate(item_iterator): image_path = item["input"][data_key] task_id = item["id"] image_id = len(images) width = None height = None # download all images of the dataset, including the ones without annotations if not os.path.exists(image_path): try: image_path = get_local_path( url=image_path, hostname=self.hostname, project_dir=self.project_dir, image_dir=self.upload_dir, cache_dir=output_image_dir, download_resources=self.download_resources, access_token=self.access_token, task_id=task_id, ) # make path relative to output_image_dir image_path = os.path.relpath(image_path, output_dir) except: logger.info( "Unable to download {image_path}. The image of {item} will be skipped".format( image_path=image_path, item=item ), exc_info=True, ) # add image to final images list try: with Image.open(os.path.join(output_dir, image_path)) as img: width, height = img.size images = add_image(images, width, height, image_id, image_path) except: logger.info( "Unable to open {image_path}, can't extract width and height for COCO export".format( image_path=image_path, item=item ), exc_info=True, ) # skip tasks without annotations if not item["output"]: # image wasn't load and there are no labels if not width: images = add_image(images, width, height, image_id, image_path) logger.warning(f"No annotations found for item {task_id=}") continue # concatenate results over all tag names labels = [] for key in item["output"]: labels += item["output"][key] if len(labels) == 0: logger.debug(f'Empty bboxes for {item["output"]} {task_id=}') continue keypoint_labels = [] for label in labels: category_name = None for key in ["rectanglelabels", "polygonlabels", "keypointlabels", "labels"]: if key in label and len(label[key]) > 0: category_name = label[key][0] break if category_name is None: logger.warning("Unknown label type or labels are empty {task_id=}") continue if not height or not width: if "original_width" not in label or "original_height" not in label: logger.debug( f"original_width or original_height not found in {image_path} {task_id=}" ) continue width, height = label["original_width"], label["original_height"] images = add_image(images, width, height, image_id, image_path) if category_name not in category_name_to_id: category_id = len(categories) category_name_to_id[category_name] = category_id categories.append({"id": category_id, "name": category_name}) category_id = category_name_to_id[category_name] annotation_id = len(annotations) if "rectanglelabels" in label or "labels" in label: xywh = self.rotated_rectangle(label) if xywh is None: continue x, y, w, h = xywh x = x * label["original_width"] / 100 y = y * label["original_height"] / 100 w = w * label["original_width"] / 100 h = h * label["original_height"] / 100 annotations.append( { "id": annotation_id, "image_id": image_id, "category_id": category_id, "segmentation": [], "bbox": [x, y, w, h], "ignore": 0, "iscrowd": 0, "area": w * h, } ) elif "polygonlabels" in label: points_abs = [ (x / 100 * width, y / 100 * height) for x, y in label["points"] ] x, y = zip(*points_abs) annotations.append( { "id": annotation_id, "image_id": image_id, "category_id": category_id, "segmentation": [ [coord for point in points_abs for coord in point] ], "bbox": get_polygon_bounding_box(x, y), "ignore": 0, "iscrowd": 0, "area": get_polygon_area(x, y), } ) elif "keypointlabels" in label: keypoint_labels.append(label) else: raise ValueError("Unknown label type") if os.getenv("LABEL_STUDIO_FORCE_ANNOTATOR_EXPORT"): annotations[-1].update({"annotator": get_annotator(item)}) if keypoint_labels: kp_order = build_kp_order(self._schema) annotations.append(process_keypoints_for_coco( keypoint_labels, kp_order, annotation_id=len(annotations), image_id=image_id, category_name_to_id=category_name_to_id, )) with io.open(output_file, mode="w", encoding="utf8") as fout: json.dump( { "images": images, "categories": categories, "annotations": annotations, "info": { "year": datetime.now().year, "version": "1.0", "description": "", "contributor": "Label Studio", "url": "", "date_created": str(datetime.now()), }, }, fout, indent=2, ) def convert_to_yolo( self, input_data, output_dir, output_image_dir=None, output_label_dir=None, is_dir=True, split_labelers=False, is_obb=False, ): """Convert data in a specific format to the YOLO format. Parameters ---------- input_data : str The input data, either a directory or a JSON file. output_dir : str The directory to store the output files in. output_image_dir : str, optional The directory to store the image files in. If not provided, it will default to a subdirectory called 'images' in output_dir. output_label_dir : str, optional The directory to store the label files in. If not provided, it will default to a subdirectory called 'labels' in output_dir. is_dir : bool, optional A boolean indicating whether `input_data` is a directory (True) or a JSON file (False). split_labelers : bool, optional A boolean indicating whether to create a dedicated subfolder for each labeler in the output label directory. obb : bool, optional A boolean indicating whether to convert to Oriented Bounding Box (OBB) format. """ if is_obb: self._check_format(Format.YOLO_OBB) else: self._check_format(Format.YOLO) ensure_dir(output_dir) notes_file = os.path.join(output_dir, "notes.json") class_file = os.path.join(output_dir, "classes.txt") if output_image_dir is not None: ensure_dir(output_image_dir) else: output_image_dir = os.path.join(output_dir, "images") os.makedirs(output_image_dir, exist_ok=True) if output_label_dir is not None: ensure_dir(output_label_dir) else: output_label_dir = os.path.join(output_dir, "labels") os.makedirs(output_label_dir, exist_ok=True) is_keypoints = keypoints_in_label_config(self._schema) if is_keypoints: # we use this attribute to add id and parentID to annotation data self.is_keypoints = True categories, category_name_to_id = get_yolo_categories_for_keypoints(self._schema) else: categories, category_name_to_id = self._get_labels() data_key = self._data_keys[0] item_iterator = ( self.iter_from_dir(input_data) if is_dir else self.iter_from_json_file(input_data) ) for item_idx, item in enumerate(item_iterator): # get image path(s) and label file path image_paths = item["input"][data_key] image_paths = [image_paths] if isinstance(image_paths, str) else image_paths # download image(s) image_path = None task_id = item["id"] # TODO: for multi-page annotation, this code won't produce correct relationships between page and annotated shapes # fixing the issue in RND-84 for image_path in reversed(image_paths): if not os.path.exists(image_path): try: image_path = get_local_path( url=image_path, hostname=self.hostname, project_dir=self.project_dir, image_dir=self.upload_dir, cache_dir=output_image_dir, download_resources=self.download_resources, access_token=self.access_token, task_id=task_id, ) # make path relative to output_image_dir image_path = os.path.relpath(image_path, output_dir) except: logger.info( "Unable to download {image_path}. The item {item} will be skipped".format( image_path=image_path, item=item ), exc_info=True, ) if not image_path: logger.error(f"No image path found for {task_id=}") continue # create dedicated subfolder for each labeler if split_labelers=True labeler_subfolder = str(item["completed_by"]) if split_labelers else "" os.makedirs( os.path.join(output_label_dir, labeler_subfolder), exist_ok=True ) # identify label file path filename = os.path.splitext(os.path.basename(image_path))[0] filename = filename[ 0 : 255 - 4 ] # urls might be too long, use 255 bytes (-4 for .txt) limit for filenames label_path = os.path.join( output_label_dir, labeler_subfolder, filename + ".txt" ) # Skip tasks without annotations if not item["output"]: logger.warning(f"No completions found for {task_id=}") if not os.path.exists(label_path): with open(label_path, "x"): pass continue # concatenate results over all tag names labels = [] for key in item["output"]: labels += item["output"][key] if len(labels) == 0: logger.warning(f'Empty bboxes for {item["output"]} in {task_id=}') if not os.path.exists(label_path): with open(label_path, "x"): pass continue categories, category_name_to_id = process_and_save_yolo_annotations(labels, label_path, category_name_to_id, categories, is_obb, is_keypoints, self._schema) with open(class_file, "w", encoding="utf8") as f: for c in categories: f.write(c["name"] + "\n") with io.open(notes_file, mode="w", encoding="utf8") as fout: json.dump( { "categories": categories, "info": { "year": datetime.now().year, "version": "1.0", "contributor": "Label Studio", }, }, fout, indent=2, ) @staticmethod def rotated_rectangle(label): if not ( "x" in label and "y" in label and "width" in label and "height" in label ): return None label_x, label_y, label_w, label_h, label_r = ( label["x"], label["y"], label["width"], label["height"], label["rotation"] if "rotation" in label else 0.0, ) if abs(label_r) > 0: alpha = math.atan(label_h / label_w) beta = math.pi * ( label_r / 180 ) # Label studio defines the angle towards the vertical axis radius = math.sqrt((label_w / 2) ** 2 + (label_h / 2) ** 2) # Label studio saves the position of top left corner after rotation x_0 = ( label_x - radius * (math.cos(math.pi - alpha - beta) - math.cos(math.pi - alpha)) + label_w / 2 ) y_0 = ( label_y + radius * (math.sin(math.pi - alpha - beta) - math.sin(math.pi - alpha)) + label_h / 2 ) theta_1 = alpha + beta theta_2 = math.pi - alpha + beta theta_3 = math.pi + alpha + beta theta_4 = 2 * math.pi - alpha + beta x_coord = [ x_0 + radius * math.cos(theta_1), x_0 + radius * math.cos(theta_2), x_0 + radius * math.cos(theta_3), x_0 + radius * math.cos(theta_4), ] y_coord = [ y_0 + radius * math.sin(theta_1), y_0 + radius * math.sin(theta_2), y_0 + radius * math.sin(theta_3), y_0 + radius * math.sin(theta_4), ] label_x = min(x_coord) label_y = min(y_coord) label_w = max(x_coord) - label_x label_h = max(y_coord) - label_y return label_x, label_y, label_w, label_h def convert_to_voc( self, input_data, output_dir, output_image_dir=None, is_dir=True ): ensure_dir(output_dir) if output_image_dir is not None: ensure_dir(output_image_dir) output_image_dir_rel = output_image_dir else: output_image_dir = os.path.join(output_dir, "images") os.makedirs(output_image_dir, exist_ok=True) output_image_dir_rel = "images" def create_child_node(doc, tag, attr, parent_node): child_node = doc.createElement(tag) text_node = doc.createTextNode(attr) child_node.appendChild(text_node) parent_node.appendChild(child_node) data_key = self._data_keys[0] item_iterator = ( self.iter_from_dir(input_data) if is_dir else self.iter_from_json_file(input_data) ) for item_idx, item in enumerate(item_iterator): image_path = item["input"][data_key] annotations_dir = os.path.join(output_dir, "Annotations") if not os.path.exists(annotations_dir): os.makedirs(annotations_dir) # Download image channels = 3 task_id = item["id"] if not os.path.exists(image_path): try: image_path = download( image_path, output_image_dir, project_dir=self.project_dir, upload_dir=self.upload_dir, return_relative_path=True, download_resources=self.download_resources, ) except: logger.info( "Unable to download {image_path}. The item {item} will be skipped".format( image_path=image_path, item=item ), exc_info=True, ) else: full_image_path = os.path.join( output_image_dir, os.path.basename(image_path) ) # retrieve number of channels from downloaded image try: _, _, channels = get_image_size_and_channels(full_image_path) except: logger.warning(f"Can't read channels from image {task_id=}") # skip tasks without annotations if not item["output"]: logger.warning(f"No annotations found for {task_id=}") continue image_name = os.path.basename(image_path) xml_name = os.path.splitext(image_name)[0] + ".xml" # concatenate results over all tag names bboxes = [] for key in item["output"]: bboxes += item["output"][key] if len(bboxes) == 0: logger.debug(f'Empty bboxes for {item["output"]} {task_id=}') continue if "original_width" not in bboxes[0] or "original_height" not in bboxes[0]: logger.debug( f"original_width or original_height not found in {image_name} {task_id=}" ) continue width, height = bboxes[0]["original_width"], bboxes[0]["original_height"] xml_filepath = os.path.join(annotations_dir, xml_name) my_dom = xml.dom.getDOMImplementation() doc = my_dom.createDocument(None, "annotation", None) root_node = doc.documentElement create_child_node(doc, "folder", output_image_dir_rel, root_node) create_child_node(doc, "filename", image_name, root_node) source_node = doc.createElement("source") create_child_node(doc, "database", "MyDatabase", source_node) create_child_node(doc, "annotation", "COCO2017", source_node) create_child_node(doc, "image", "flickr", source_node) create_child_node(doc, "flickrid", "NULL", source_node) create_child_node(doc, "annotator", get_annotator(item, ""), source_node) root_node.appendChild(source_node) owner_node = doc.createElement("owner") create_child_node(doc, "flickrid", "NULL", owner_node) create_child_node(doc, "name", "Label Studio", owner_node) root_node.appendChild(owner_node) size_node = doc.createElement("size") create_child_node(doc, "width", str(width), size_node) create_child_node(doc, "height", str(height), size_node) create_child_node(doc, "depth", str(channels), size_node) root_node.appendChild(size_node) create_child_node(doc, "segmented", "0", root_node) for bbox in bboxes: key = ( "rectanglelabels" if "rectanglelabels" in bbox else ("labels" if "labels" in bbox else None) ) if key is None or len(bbox[key]) == 0: continue name = bbox[key][0] x = int(bbox["x"] / 100 * width) y = int(bbox["y"] / 100 * height) w = int(bbox["width"] / 100 * width) h = int(bbox["height"] / 100 * height) object_node = doc.createElement("object") create_child_node(doc, "name", name, object_node) create_child_node(doc, "pose", "Unspecified", object_node) create_child_node(doc, "truncated", "0", object_node) create_child_node(doc, "difficult", "0", object_node) bndbox_node = doc.createElement("bndbox") create_child_node(doc, "xmin", str(x), bndbox_node) create_child_node(doc, "ymin", str(y), bndbox_node) create_child_node(doc, "xmax", str(x + w), bndbox_node) create_child_node(doc, "ymax", str(y + h), bndbox_node) object_node.appendChild(bndbox_node) root_node.appendChild(object_node) with io.open(xml_filepath, mode="w", encoding="utf8") as fout: doc.writexml(fout, addindent="" * 4, newl="\n", encoding="utf-8") def _get_labels(self): labels = set() categories = list() category_name_to_id = dict() for name, info in self._schema.items(): labels |= set(info["labels"]) attrs = info["labels_attrs"] for label in attrs: if attrs[label].get("category"): categories.append( {"id": attrs[label].get("category"), "name": label} ) category_name_to_id[label] = attrs[label].get("category") labels_to_add = set(labels) - set(list(category_name_to_id.keys())) labels_to_add = sorted(list(labels_to_add)) idx = 0 while idx in list(category_name_to_id.values()): idx += 1 for label in labels_to_add: categories.append({"id": idx, "name": label}) category_name_to_id[label] = idx idx += 1 while idx in list(category_name_to_id.values()): idx += 1 return categories, category_name_to_id