fitbenchmarking.core.results_output module

Functions that create the tables, support pages, figures, and indexes.

fitbenchmarking.core.results_output.check_max_solvers(opts, solvers, max_solvers)

Check number of solvers and update the options to display in the dropdown.

Parameters:

  • opts (dict[str]) – The options for the dropdown to be updated

  • solvers (list[str]) – The solvers to be plotted

  • max_solvers (int) – Maximum number of solvers that can be plotted

Returns:

Options to display in the dropdown

Return type:

dict[str]

fitbenchmarking.core.results_output.create_directories(options, group_name)

Create the directory structure ready to store the results

Parameters:
Returns:

paths to the top level group results, support pages, and figures directories

Return type:

(str, str, str)

fitbenchmarking.core.results_output.create_index_page(options: Options, groups: list[str], result_directories: list[str]) → str

Creates the results index page for the benchmark, and copies the fonts and js directories to the correct location.

Parameters:

  • options (fitbenchmarking.utils.options.Options) – The user options for the benchmark.

  • groups (A list of strings.) – Names for each of the problem set groups.

  • result_directories – Result directory paths for each

problem set group. :type result_directories: A list of strings. :return: The filepath of the results_index.html file. :rtype: str

fitbenchmarking.core.results_output.create_plots(options, results, best_results, figures_dir)

Create a plot for each result and store in the figures directory

Parameters:
fitbenchmarking.core.results_output.create_problem_level_index(options, table_names, group_name, group_dir, table_descriptions)

Generates problem level index page.

Parameters:

  • options (fitbenchmarking.utils.options.Options) – The options used in the fitting problem and plotting

  • table_names (list) – list of table names

  • group_name (str) – name of the problem group

  • group_dir (str) – Path to the directory where the index should be stored

  • table_descriptions (dict) – dictionary containing descriptions of the tables and the comparison mode

fitbenchmarking.core.results_output.display_page(pathname, profile_instances_all_groups, layout, max_solvers)

Update the layout of the dash app.

Parameters:

  • pathname (str) – The link to the page with the Dash plot

  • profile_instances_all_groups (dict[str[dict]]) – The data to be plotted

  • layout (list of dcc or html components) – Layout of the Dash app

  • max_solvers (int) – Maximum number of solvers that can be plotted

Returns:

The updated layout

Return type:

html.Div

fitbenchmarking.core.results_output.open_browser(output_file: str, options, pp_dfs_all_prob_sets) → None

Opens a browser window to show the results of a fit benchmark.

Parameters:

  • output_file (str) – The absolute path to the results index file.

  • options (fitbenchmarking.utils.options.Options) – The user options for the benchmark.

  • pp_dfs_all_prob_sets (dict[str, dict[str, pandas.DataFrame]]) – For each problem set, data to create dash plots.

fitbenchmarking.core.results_output.preprocess_data(results: list[FittingResult])

Generate a dictionary of results lists sorted into the correct order with rows and columns as the key and list elements respectively.

This is used to create HTML and txt tables. This is stored in self.sorted_results

Parameters:

results (list[fitbenchmarking.utils.fitbm_result.FittingResult]) – The results to process

Returns:

The best result grouped by row and category (cost function), The sorted results grouped by row and category

Return type:

dict[str, dict[str, utils.fitbm_result.FittingResult]], dict[str, dict[str, list[utils.fitbm_result.FittingResult]]]

fitbenchmarking.core.results_output.run_dash_app(options, pp_dfs_all_prob_sets) → None

Runs the Dash app to produce the interactive performance profile plots.

Parameters:

  • options (fitbenchmarking.utils.options.Options) – The user options for the benchmark.

  • pp_dfs_all_prob_sets (dict[str, dict[str, pandas.DataFrame]]) – For each problem set, data to create dash plots.

fitbenchmarking.core.results_output.save_results(options, results, group_name, failed_problems, unselected_minimizers)

Create all results files and store them. Result files are plots, support pages, tables, and index pages.

Parameters:
Params unselected_minimizers:

Dictionary containing unselected minimizers based on the algorithm_type option

Returns:

Path to directory of group results, data for building the performance profile plots

Return type:

str, dict[str, pandas.DataFrame]

fitbenchmarking.core.results_output.update_warning(solvers, max_solvers)

Give a warning if the number of solvers is above the maximum that can be displayed.

Parameters:

  • solvers (list[str]) – The solvers to be plotted

  • max_solvers (int) – Maximum number of solvers that can be plotted

Returns:

The warning

Return type:

str