Unit testing tolerances + more consistent definition of SNR #79
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Dynamic scaling adjusted. Also boundaries adjusted Remove normalisation as signal isnormalized. Remove algorithm-specific tolerance boundaries such that default boundaries are used Take medians of parameters as means give some machine precission error...
oliverchampion
requested review from
etpeterson,
IvanARashid and
oscarjalnefjord
etpeterson
requested changes
Aug 29, 2024
| 'D': np.mean(Dim[selector], axis=0), | ||
| 'f': np.mean(fim[selector], axis=0), | ||
| 'Dp': np.mean(Dpim[selector], axis=0), | ||
| 'D': np.median(Dim[selector], axis=0), |
| scale = dyn_rtol["offset"] + dyn_rtol["ratio"]*ratio + dyn_rtol["noise"]*noise + dyn_rtol["noiseCrossRatio"]*ratio*noise | ||
| tolerances["rtol"] = {"f": scale*dyn_rtol["f"], "D": scale*dyn_rtol["D"], "Dp": scale*dyn_rtol["Dp"]} | ||
| scale = dyn_rtol["offset"] + dyn_rtol["noise"]*data["noise"] | ||
| if dyn_rtol["scale_f"]: |
There was a problem hiding this comment.
I think you want to also check if this key exists.
| scale = dyn_atol["offset"] + dyn_atol["ratio"]*ratio + dyn_atol["noise"]*noise + dyn_atol["noiseCrossRatio"]*ratio*noise | ||
| tolerances["atol"] = {"f": scale*dyn_atol["f"], "D": scale*dyn_atol["D"], "Dp": scale*dyn_atol["Dp"]} | ||
| scale = dyn_atol["offset"] + dyn_atol["noise"]*data["noise"] | ||
| if dyn_atol["scale_f"]: |
| scale = dyn_rtol["offset"] + dyn_rtol["ratio"]*ratio + dyn_rtol["noise"]*noise + dyn_rtol["noiseCrossRatio"]*ratio*noise | ||
| tolerances["rtol"] = {"f": scale*dyn_rtol["f"], "D": scale*dyn_rtol["D"], "Dp": scale*dyn_rtol["Dp"]} | ||
| scale = dyn_rtol["offset"] + dyn_rtol["noise"]*data["noise"] | ||
| if dyn_rtol["scale_f"]: |
There was a problem hiding this comment.
I think this is a misnomer and should be renamed. I thought something called scale_f would be scaling the f parameter. Instead it scales the others by f.
| if dyn_rtol["scale_f"]: | ||
| tolerances["rtol"] = {"f": scale*dyn_rtol["f"], "D": scale*dyn_rtol["D"]/(1-data['f']+epsilon), "Dp": scale*dyn_rtol["Dp"]/(data['f']+epsilon)} | ||
| else: | ||
| tolerances["rtol"] = {"f": scale * dyn_rtol["f"], "D": scale * dyn_rtol["D"]/0.9,"Dp": scale * dyn_rtol["Dp"]/0.1} |
There was a problem hiding this comment.
Why 0.9 and 0.1? Also why would we not want to scale by f?
| signal /= signal[0] | ||
| ratio = 1 / signal[0] | ||
| return signal, ratio | ||
| signal = np.abs(signal) #Oliver: do we need this for unit testing? It will bias the fits by forcing positive signals. |
There was a problem hiding this comment.
Probably not necessary.
| ratio = 1 / signal[0] | ||
| return signal, ratio | ||
| signal = np.abs(signal) #Oliver: do we need this for unit testing? It will bias the fits by forcing positive signals. | ||
| signal /= signal[0] # this scales noise differenly. I have now made sure signal is produced with amplitude 1 if right flags are selected. Then SNR is fixed. Then, this sentence may be redundant |
There was a problem hiding this comment.
Should probably be removed, at least eventually each algorithm should have their own scaling built in. For now maybe it needs to stay though?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Describe the changes you have made in this PR
1: I want to have a more consistent definition of SNR in the phantom. By putting T1T2 to False, T1 and T2 effects are ignored and all signal starts at 1. This means that SNR = 1/noise.
2: I have tried to make tolerances scale with SNR and f. --> This is also meant to get the discussion going for Friday and can be adapted accordingly
Link this PR to an issue [optional]
1 Fixes _#76
2 Fixes _#75
Checklist