qgear

extract_qft_from_filename

 extract_qft_from_filename (filename)

Check if filename contains ‘qft1’ and return it if found

source

extract_date_from_path

 extract_date_from_path (file_path)

Extract date from a path containing ‘dataCudaQ_’

source

readOne

 readOne (inpF, dataD, verb=1)

Read one YAML metrics file (non-QFT)

source

readOneQFT

 readOneQFT (inpF, dataD, qft, verb=1)

Read one YAML metrics file for QFT runs

source

find_yaml_files

 find_yaml_files (directory_path, vetoL=None)

Find all .yaml files in a directory, excluding files with veto strings

source

sort_end_lists

 sort_end_lists (d, sort_key='nq', val_key='runt', parent_key='')

Recursively sort lists in nested dicts by a sort_key

source

MetricsPlotter

 MetricsPlotter (prjName='metrics', shift=True, outPath='out',
                 noXterm=True, verb=1)

Plotter for concatenated metrics

source

metrics_plot

 metrics_plot (measPath:str=None, pathL=None, vetoL=None, showPlots='b',
               shift=True, outPath='out', noXterm=True, verb=1)

*Concatenate and plot metrics from multiple job YAML files.

Args: corePath: Base path to data folders. Defaults to the current working directory. pathL: List of date folder suffixes (e.g., [‘Nov15’]). If None, auto-detects all dataCudaQ_* folders. vetoL: List of substrings to exclude from filenames. showPlots: String of plots to show (‘a’=cpu, ‘b’=par-gpu, ‘c’=adj-gpu, ‘d’=qft). shift: Whether to randomly shift CPU points to avoid overlap. outPath: Output directory for plots. noXterm: Disable X-term (for headless plotting). verb: Verbosity level.*

source

postproc_qcrank

 postproc_qcrank (expD, md, verb=1)

*Post-process QCrank experiment results to reconstruct user data.

Args: expD: Experiment data dictionary from HDF5. md: Metadata dictionary from HDF5. verb: Verbosity level.*

source

restore_canned_image

 restore_canned_image (expD, md)

*Restore reconstructed image from QCrank output.

Args: expD: Experiment data dictionary. md: Metadata dictionary.*

source

residual_analysis

 residual_analysis (expD, md)

*Compute residual statistics between reconstructed and true data.

Args: expD: Experiment data dictionary. md: Metadata dictionary.*

source

process_qcrank_experiment

 process_qcrank_experiment (exp_name, inp_path=None, out_path='out',
                            show_plots='a', no_xterm=True, verb=1,
                            save_plots=True)

*Load, post-process, and optionally plot QCrank experiment results.

Args: exp_name: Experiment name (without .h5 extension). inp_path: Path to raw outputs from experiment. Defaults to current working directory. out_path: Path to save post-processed results. show_plots: String of plots to show (‘a’=accuracy, ‘b’=image, ‘c’=dynamic range). no_xterm: Disable X-term for headless plotting. verb: Verbosity level. save_plots: If True, save plots to file.

Returns: (expD, expMD) - processed data and metadata.*