Mercurial > repos > iuc > biapy
comparison create_yaml.py @ 0:356d58ae85fa draft default tip
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/main/tools/biapy commit 63860b5c6c21e0b76b1c55a5e71cafcb77d6cc84
| author | iuc |
|---|---|
| date | Fri, 06 Feb 2026 17:50:32 +0000 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| -1:000000000000 | 0:356d58ae85fa |
|---|---|
| 1 import argparse | |
| 2 import sys | |
| 3 | |
| 4 import requests | |
| 5 import yaml | |
| 6 | |
| 7 | |
| 8 def download_yaml_template(workflow, dims, biapy_version=""): | |
| 9 template_dir_map = { | |
| 10 "SEMANTIC_SEG": "semantic_segmentation", | |
| 11 "INSTANCE_SEG": "instance_segmentation", | |
| 12 "DETECTION": "detection", | |
| 13 "DENOISING": "denoising", | |
| 14 "SUPER_RESOLUTION": "super-resolution", | |
| 15 "CLASSIFICATION": "classification", | |
| 16 "SELF_SUPERVISED": "self-supervised", | |
| 17 "IMAGE_TO_IMAGE": "image-to-image", | |
| 18 } | |
| 19 | |
| 20 # Use .get() to avoid KeyError if workflow is unexpected | |
| 21 dir_name = template_dir_map.get(workflow) | |
| 22 if not dir_name: | |
| 23 raise ValueError(f"Unknown workflow: {workflow}") | |
| 24 | |
| 25 template_name = f"{dir_name}/{dims.lower()}_{dir_name}.yaml" | |
| 26 url = f"https://raw.githubusercontent.com/BiaPyX/BiaPy/refs/tags/v{biapy_version}/templates/{template_name}" | |
| 27 | |
| 28 print(f"Downloading YAML template from {url}") | |
| 29 try: | |
| 30 response = requests.get(url, timeout=10) # Added timeout | |
| 31 response.raise_for_status() # Automatically raises HTTPError for 4xx/5xx | |
| 32 return yaml.safe_load(response.text) or {} | |
| 33 except requests.exceptions.RequestException as e: | |
| 34 print(f"Error: Could not download template. {e}") | |
| 35 sys.exit(1) # Exit gracefully rather than crashing with a stack trace | |
| 36 | |
| 37 | |
| 38 def tuple_to_list(obj): | |
| 39 """Convert tuples to lists recursively.""" | |
| 40 if isinstance(obj, tuple): | |
| 41 return list(obj) | |
| 42 if isinstance(obj, dict): | |
| 43 return {k: tuple_to_list(v) for k, v in obj.items()} | |
| 44 if isinstance(obj, list): | |
| 45 return [tuple_to_list(v) for v in obj] | |
| 46 return obj | |
| 47 | |
| 48 | |
| 49 def main(): | |
| 50 parser = argparse.ArgumentParser( | |
| 51 description="Generate a YAML configuration from given arguments." | |
| 52 ) | |
| 53 parser.add_argument( | |
| 54 '--input_config_path', default='', type=str, | |
| 55 help="Input configuration file to reuse" | |
| 56 ) | |
| 57 parser.add_argument( | |
| 58 '--new_config', action='store_true', | |
| 59 help="Whether to create a new config or reuse an existing one." | |
| 60 ) | |
| 61 parser.add_argument( | |
| 62 '--out_config_path', required=True, type=str, | |
| 63 help="Path to save the generated YAML configuration." | |
| 64 ) | |
| 65 parser.add_argument( | |
| 66 '--workflow', default='semantic', type=str, | |
| 67 choices=['semantic', 'instance', 'detection', 'denoising', | |
| 68 'sr', 'cls', 'sr2', 'i2i'], | |
| 69 ) | |
| 70 parser.add_argument( | |
| 71 '--dims', default='2d', type=str, | |
| 72 choices=['2d_stack', '2d', '3d'], | |
| 73 help="Number of dimensions for the problem" | |
| 74 ) | |
| 75 parser.add_argument( | |
| 76 '--obj_slices', default='', type=str, | |
| 77 choices=['', '1-5', '5-10', '10-20', '20-60', '60+'], | |
| 78 help="Number of slices for the objects in the images" | |
| 79 ) | |
| 80 parser.add_argument( | |
| 81 '--obj_size', default='0-25', type=str, | |
| 82 choices=['0-25', '25-100', '100-200', '200-500', '500+'], | |
| 83 help="Size of the objects in the images" | |
| 84 ) | |
| 85 parser.add_argument( | |
| 86 '--img_channel', default=1, type=int, | |
| 87 help="Number of channels in the input images" | |
| 88 ) | |
| 89 parser.add_argument( | |
| 90 '--model_source', default='biapy', | |
| 91 choices=['biapy', 'bmz', 'torchvision'], | |
| 92 help="Source of the model." | |
| 93 ) | |
| 94 parser.add_argument( | |
| 95 '--model', default='', type=str, | |
| 96 help=("Path to the model file if using a pre-trained model " | |
| 97 "from BiaPy or name of the model within BioImage " | |
| 98 "Model Zoo or TorchVision.") | |
| 99 ) | |
| 100 parser.add_argument( | |
| 101 '--raw_train', default='', type=str, | |
| 102 help="Path to the training raw data." | |
| 103 ) | |
| 104 parser.add_argument( | |
| 105 '--gt_train', default='', type=str, | |
| 106 help="Path to the training ground truth data." | |
| 107 ) | |
| 108 parser.add_argument( | |
| 109 '--test_raw_path', default='', type=str, | |
| 110 help="Path to the testing raw data." | |
| 111 ) | |
| 112 parser.add_argument( | |
| 113 '--test_gt_path', default='', type=str, | |
| 114 help="Path to the testing ground truth data." | |
| 115 ) | |
| 116 parser.add_argument( | |
| 117 '--biapy_version', default='', type=str, | |
| 118 help="BiaPy version to use." | |
| 119 ) | |
| 120 parser.add_argument( | |
| 121 '--num_cpus', default="1", type=str, | |
| 122 help="Number of CPUs to allocate." | |
| 123 ) | |
| 124 args = parser.parse_args() | |
| 125 | |
| 126 if args.new_config: | |
| 127 workflow_map = { | |
| 128 "semantic": "SEMANTIC_SEG", | |
| 129 "instance": "INSTANCE_SEG", | |
| 130 "detection": "DETECTION", | |
| 131 "denoising": "DENOISING", | |
| 132 "sr": "SUPER_RESOLUTION", | |
| 133 "cls": "CLASSIFICATION", | |
| 134 "sr2": "SELF_SUPERVISED", | |
| 135 "i2i": "IMAGE_TO_IMAGE", | |
| 136 } | |
| 137 workflow_type = workflow_map[args.workflow] | |
| 138 | |
| 139 ndim = "3D" if args.dims == "3d" else "2D" | |
| 140 as_stack = args.dims in ["2d_stack", "2d"] | |
| 141 | |
| 142 config = download_yaml_template(workflow_type, ndim, biapy_version=args.biapy_version) | |
| 143 | |
| 144 # Initialization using setdefault to prevent KeyErrors | |
| 145 config.setdefault("PROBLEM", {}) | |
| 146 config["PROBLEM"].update({"TYPE": workflow_type, "NDIM": ndim}) | |
| 147 | |
| 148 config.setdefault("TEST", {})["ANALIZE_2D_IMGS_AS_3D_STACK"] = as_stack | |
| 149 | |
| 150 # Handle MODEL and PATHS | |
| 151 model_cfg = config.setdefault("MODEL", {}) | |
| 152 if args.model_source == "biapy": | |
| 153 model_cfg["SOURCE"] = "biapy" | |
| 154 is_loading = bool(args.model) | |
| 155 model_cfg["LOAD_CHECKPOINT"] = is_loading | |
| 156 model_cfg["LOAD_MODEL_FROM_CHECKPOINT"] = is_loading | |
| 157 if is_loading: | |
| 158 config.setdefault("PATHS", {})["CHECKPOINT_FILE"] = args.model | |
| 159 elif args.model_source == "bmz": | |
| 160 model_cfg["SOURCE"] = "bmz" | |
| 161 model_cfg.setdefault("BMZ", {})["SOURCE_MODEL_ID"] = args.model | |
| 162 elif args.model_source == "torchvision": | |
| 163 model_cfg["SOURCE"] = "torchvision" | |
| 164 model_cfg["TORCHVISION_MODEL_NAME"] = args.model | |
| 165 | |
| 166 # PATCH_SIZE Logic | |
| 167 obj_size_map = { | |
| 168 "0-25": (256, 256), "25-100": (256, 256), | |
| 169 "100-200": (512, 512), "200-500": (512, 512), "500+": (1024, 1024), | |
| 170 } | |
| 171 obj_size = obj_size_map[args.obj_size] | |
| 172 | |
| 173 obj_slices_map = {"": -1, "1-5": 5, "5-10": 10, "10-20": 20, "20-60": 40, "60+": 80} | |
| 174 obj_slices = obj_slices_map.get(args.obj_slices, -1) | |
| 175 | |
| 176 if ndim == "2D": | |
| 177 patch_size = obj_size + (args.img_channel,) | |
| 178 else: | |
| 179 if obj_slices == -1: | |
| 180 print("Error: For 3D problems, obj_slices must be specified.") | |
| 181 sys.exit(1) | |
| 182 patch_size = (obj_slices,) + obj_size + (args.img_channel,) | |
| 183 | |
| 184 config.setdefault("DATA", {})["PATCH_SIZE"] = str(patch_size) | |
| 185 config["DATA"]["REFLECT_TO_COMPLETE_SHAPE"] = True | |
| 186 | |
| 187 else: | |
| 188 if not args.input_config_path: | |
| 189 print("Error: Input configuration path must be specified.") | |
| 190 sys.exit(1) | |
| 191 try: | |
| 192 with open(args.input_config_path, 'r', encoding='utf-8') as f: | |
| 193 config = yaml.safe_load(f) or {} | |
| 194 except FileNotFoundError: | |
| 195 print(f"Error: File {args.input_config_path} not found.") | |
| 196 sys.exit(1) | |
| 197 | |
| 198 # Always set NUM_CPUS | |
| 199 config.setdefault("SYSTEM", {}) | |
| 200 try: | |
| 201 num_cpus = max(int(args.num_cpus), 1) | |
| 202 except BaseException: | |
| 203 num_cpus = 1 | |
| 204 config["SYSTEM"].update({"NUM_CPUS": num_cpus}) | |
| 205 | |
| 206 # Global overrides (Train/Test) | |
| 207 config.setdefault("TRAIN", {}) | |
| 208 config.setdefault("DATA", {}) | |
| 209 | |
| 210 if args.raw_train: | |
| 211 config["TRAIN"]["ENABLE"] = True | |
| 212 config["DATA"].setdefault("TRAIN", {}).update({ | |
| 213 "PATH": args.raw_train, | |
| 214 "GT_PATH": args.gt_train | |
| 215 }) | |
| 216 else: | |
| 217 config["TRAIN"]["ENABLE"] = False | |
| 218 | |
| 219 test_cfg = config.setdefault("TEST", {}) | |
| 220 if args.test_raw_path: | |
| 221 test_cfg["ENABLE"] = True | |
| 222 data_test = config["DATA"].setdefault("TEST", {}) | |
| 223 data_test["PATH"] = args.test_raw_path | |
| 224 data_test["LOAD_GT"] = bool(args.test_gt_path) | |
| 225 if args.test_gt_path: | |
| 226 data_test["GT_PATH"] = args.test_gt_path | |
| 227 else: | |
| 228 test_cfg["ENABLE"] = False | |
| 229 | |
| 230 config.setdefault("MODEL", {})["OUT_CHECKPOINT_FORMAT"] = "safetensors" | |
| 231 | |
| 232 # Final cleanup and save | |
| 233 config = tuple_to_list(config) | |
| 234 with open(args.out_config_path, 'w', encoding='utf-8') as f: | |
| 235 yaml.dump(config, f, default_flow_style=False) | |
| 236 | |
| 237 print(f"Success: YAML configuration written to {args.out_config_path}") | |
| 238 | |
| 239 | |
| 240 if __name__ == "__main__": | |
| 241 main() |
