Experiment 5.4#

To run this experiment, the following steps are required:

For every combination of algorithm (signal fitting and signal masking), data set, and binning factor, run the experiment using the scripts in HSR4HCI_SCRIPTS_DIR/experiments/hypothesis-based (to use the hypothesis-based version of the HSR):
```
python HSR4HCI_SCRIPTS_DIR/experiments/hypothesis-based/01_train_based_on_hypothesis.py \
  --experiment-dir $HSR4HCI_EXPERIMENTS_DIR/main/5.4_photometry-real-planets/<algorithm>/<dataset>/<binning-factor>
```
If you have access to a cluster, you can parallelize this using the --n-roi-splits and --roi-split options. This step will create a directory:
```
$HSR4HCI_EXPERIMENTS_DIR/main/5.4_photometry-real-planets/<algorithm>/<dataset>/<binning-factor>/fits/partial
```
which contains the partial result files. If you do not parallelize the experiment using --n-roi-splits and --roi-split, this folder will only contain a single FITS file.
Merge all partial result files using the following command:
```
python $HSR4HCI_SCRIPTS_DIR/experiments/hypothesis-based/02_merge_residuals_and_get_signal_estimate.py \
  --experiment-dir $HSR4HCI_EXPERIMENTS_DIR/main/5.4_photometry-real-planets/<algorithm>/<dataset>/<binning-factor>
```
This will create a residuals.fits and a hypotheses.fits file at the same level as the partial folder.

Attention

This step is necessary even if you did not parallelize the experiment using --n-roi-splits and --roi-split.
Once you have run all 18 combinations of algorithm, data set and binning factor, you can generate the final result table from Table 2 as follows:
```
python $HSR4HCI_EXPERIMENTS_DIR/main/5.4_photometry-real-planets/get_contrast_estimates.py
```
This will create a TSV file with the estimated contrasts and a LaTeX file with the table from the paper. Both files will be placed in:
```
$HSR4HCI_EXPERIMENTS_DIR/main/5.4_photometry-real-planets
```