Merge pull request #1 from CAAI/versioning

Version control

Author: RaphaelSura

README.md

@@ -7,4 +7,22 @@ Install only to your user. Go to your virtual environment. Run:
 ```
 git clone https://github.com/CAAI/rh-torch.git && cd rh-torch
 pip install .
-```
+```
+
+## HOW TO CONTRIBUTE
+
+Create your edits in a different branch, decicated to a few specific things. We prefer many minor edits over one huge. Once everything is well-documented and tested, perform a pull request for your edits to be made available in the main branch. See steps here:
+```
+git branch awesome-addition
+git checkout awesome-addition
+# do your changes
+git commit -a -m 'your changes'
+git push
+gh pr create --title "your title" --body "longer description of what you did"
+```
+If you wish, you can add ```--assignee <github_name>``` to ping specific persons for looking at your pull request.
+
+One someone accepted your pull request (after reviewing the changes), it will be part of the main branch.
+
+### Important before accepting pull requests
+Before you accept a pull request, please update the version.py with an incremented number and a description of what has changed. This number is logged as part of the training config file that is autogenerated.

rhtorch/config_utils.py

@@ -4,6 +4,8 @@
 from datetime import datetime
 from pathlib import Path
 import torch
+from rhtorch.version import __version__
+import socket
 
 loss_map = {'MeanAbsoluteError': 'mae',
             'MeanSquaredError': 'mse',
@@ -24,10 +26,9 @@ def load_model_config(rootdir, arguments):
     with open(config_file) as file:
         config = yaml.load(file, Loader=yaml.RoundTripLoader)
 
-    batch_size = config['batch_size'] * torch.cuda.device_count()
     data_shape = 'x'.join(map(str, config['data_shape']))
     base_name = f"{config['module']}_{config['version_name']}_{config['data_generator']}"
-    dat_name = f"bz{batch_size}_{data_shape}"
+    dat_name = f"bz{config['batch_size']}_{data_shape}"
     full_name = f"{base_name}_{dat_name}_k{arguments.kfold}_e{config['epoch']}"
 
     # check for data folder
@@ -39,22 +40,31 @@ def load_model_config(rootdir, arguments):
         raise FileNotFoundError("Data path not found. Define relative to the project directory or as absolute path in config file")
 
     # additional info from args and miscellaneous to save in config
-    config['build date'] = datetime.now().strftime("%Y-%m-%d %H.%M.%S")
+    config['build_date'] = datetime.now().strftime("%Y-%m-%d %H.%M.%S")
     config['model_name'] = full_name
     config['project_dir'] = str(rootdir)
     config['data_folder'] = str(data_folder)
     config['config_file'] = str(config_file)
     config['k_fold'] = arguments.kfold
-    config['precision'] = arguments.precision
+    if 'precision' not in config:
+        config['precision'] = 32
+    config['GPUs'] = torch.cuda.device_count()
+    config['global_batch_size'] = config['batch_size'] * config['GPUs']
+    config['rhtorch_version'] = __version__
+    config['hostname'] = socket.gethostname()
     if 'acc_grad_batches' not in config:
         config['acc_grad_batches'] = 1
 
     return config
 
 
-def copy_model_config(path, config):
+def copy_model_config(path, config, append_timestamp=False):
     model_name = config['model_name']
-    config_file = path.joinpath(f"config_{model_name}.yaml")
+    if append_timestamp:
+        timestamp = config['build_date'].replace(' ','_')
+        config_file = path.joinpath(f"config_{model_name}_{timestamp}.yaml")
+    else:
+        config_file = path.joinpath(f"config_{model_name}.yaml") 
     config.yaml_set_start_comment(f'Config file for {model_name}')
     with open(config_file, 'w') as file:
         yaml.dump(config, file, Dumper=yaml.RoundTripDumper)

rhtorch/torch_training.py

@@ -14,7 +14,6 @@
 from rhtorch.callbacks import plotting
 from rhtorch.config_utils import load_model_config, copy_model_config
 
-
 def main():
     import argparse
 
@@ -25,7 +24,6 @@ def main():
     parser.add_argument("-c", "--config", help="Config file else than 'config.yaml' in project directory (input dir)", type=str, default='config.yaml')
     parser.add_argument("-k", "--kfold", help="K-value for selecting train/test split subset. Default k=0", type=int, default=0)
     parser.add_argument("-t", "--test", help="Test run for 1 patient", action="store_true", default=False)
-    # parser.add_argument("-p", "--precision", help="Torch precision. Default 32", type=int, default=32)
 
     args = parser.parse_args()
     project_dir = Path(args.input)
@@ -117,6 +115,11 @@ def main():
     )
     callbacks.append(checkpoint_callback)
 
+    # Save the config prior to training the model - one for each time the script is started
+    if not is_test:        
+        copy_model_config(model_path, configs, append_timestamp=True)
+        print("Saved config prior to model training")
+    
     # set the trainer and fit
     trainer = pl.Trainer(max_epochs=configs['epoch'], 
                          logger=wandb_logger, 
@@ -140,7 +143,7 @@ def main():
     torch.save(model.state_dict(), output_file)
     copy_model_config(model_path, configs)
     print("Saved model and config file to disk")
-
+    
 
 if __name__ == "__main__":
     main()

rhtorch/version.py

@@ -0,0 +1,18 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+Created on Fri May 21 08:58:01 2021
+
+@author: claes
+"""
+
+__version__ = '0.0.3'
+
+"""
+
+VERSIONING (UPDATED WHEN PR ARE MERGED INTO MASTER BRANCH)
+ 0.0.1 # Added repository (CL 18-05-2021)
+ 0.0.2 # Cleaned up main, moved to torchmetrics in modules (RD 20-05-2021)
+ 0.0.3 # Added version control to config-logfiles
+
+"""