Add methods to print information about the evaluations #175
Conversation
for more information, see https://pre-commit.ci
…nto feature/print_history
AngelFP
left a comment
AngelFP
left a comment
There was a problem hiding this comment.
Thanks for this! Could you implement the changes in the comments and add these two function calls to test_exploration_diagnostics.py?
| self.history_file = os.path.join(path, output_files[-1]) | ||
| self.params_file = os.path.join( | ||
| path, "exploration_parameters.json" | ||
| ) | ||
| elif path.endswith(".npy"): | ||
| exploration_dir_path = pathlib.Path(path).parent | ||
| output_file = path | ||
| params_file = os.path.join( | ||
| self.history_file = path | ||
| self.params_file = os.path.join( | ||
| exploration_dir_path, "exploration_parameters.json" | ||
| ) | ||
| else: | ||
| raise RuntimeError( | ||
| "The path should either point to a folder or a `.npy` file." | ||
| ) | ||
| exploration = self._create_exploration( | ||
| exploration_dir_path, params_file, output_file | ||
| exploration_dir_path, self.params_file, self.history_file |
There was a problem hiding this comment.
Is there a need to store the paths to history_file and params_file in self? They are not used anywhere else.
There was a problem hiding this comment.
You are right, they are not being used anywhere else yet.
Nevertheless, I think that it is important that when one initializes an instance of ExplorationDiagnostics from files, it should be known which are those files.
So the user knows exactly where this data is coming from.
There was a problem hiding this comment.
Right, but the path to the data is already in exploration_dir_path. Is there any reason to have the path to the json and npy files? In principle I don't see a use case.
There was a problem hiding this comment.
In previous versions there could be more than one history file in the exploration_dir_path.
When this is the case, and just the exploration_dir_path is given, the most recent history file is automatically selected.
Keeping the path to the selected history file seems to me a good practice so it can be known the origin of the data store in ExplorationDiagnostics.
But I see that in newer versions of optimas there is just one history file per exploration_dir_path.
So it should be sufficiently clear which one applies.
| self, | ||
| top: Optional[int] = 3, | ||
| objective: Optional[Union[str, Objective]] = None, | ||
| ) -> str: |
There was a problem hiding this comment.
| ) -> str: | |
| ) -> None: |
|
|
||
| print() | ||
|
|
||
| def print_top_evaluations( |
There was a problem hiding this comment.
To keek it consistent with the other methods, I would change the name to
| def print_top_evaluations( | |
| def print_best_evaluations( |
| sim_path = None | ||
| print("%20s = None" % ("dir_path")) | ||
|
|
||
| print("objective functions:") |
There was a problem hiding this comment.
| print("objective functions:") | |
| print("objectives:") |
2 participants
print_best_evaluationsshows a table with the top scoring evaluations in the given objective.example:
print_evaluationshows the relevant parameters of a given evaluation.example: