DM-34117: Update stats to improve readability #37
Conversation
sr525
changed the title
pipelines/coaddQAPlots_test.yaml
Outdated
| @@ -0,0 +1,46 @@ | |||
| description: Make coadd plots for QA | |||
There was a problem hiding this comment.
Did you mean to push this file?
There was a problem hiding this comment.
I did not, sorry, I lost track of my test pipeline.
| lowStats[sourceType] = lowStatsStr | ||
|
|
||
| if np.sum(highSn) > 0: | ||
| highMags[sourceType] = f"{np.nanmax(catPlot.loc[highSn, magCol]):.2f}" | ||
| sortedMags = np.sort(catPlot.loc[highSn, magCol]) | ||
| x = int((len(sortedMags)/100)*10) |
There was a problem hiding this comment.
Did you want to do an integer division with // here? Otherwise, this is going to be identical to int(len(sortedMags)/10)
There was a problem hiding this comment.
I left it like this to make it more obvious that I was going for 10%, I can tidy it up if you would prefer.
There was a problem hiding this comment.
To me, it looked like you were trying something else here. It might be cleaner to just have this line as int(len(sortedMags)/10) or int(0.1*len(sortedMags)) and have a comment that it's 10%, if you prefer to make it more evident.
| if len(xs) < 2: | ||
| medLine, = ax.plot(xs, np.nanmedian(ys), color, | ||
| label="Median: {:0.2f}".format(np.nanmedian(ys)), lw=0.8) | ||
| label="Median: {:0.3g}".format(np.nanmedian(ys)), lw=0.8) |
There was a problem hiding this comment.
| label="Median: {:0.3g}".format(np.nanmedian(ys)), lw=0.8) | |
| label=f"Median: {np.nanmedian(ys):0.3g}", lw=0.8) |
sr525
force-pushed
the
tickets/DM-34117
branch
2 times, most recently
from
87aec6a
to
f1554a2
Compare
|
Rebasing has picked up old commits, it looks like |
There was a problem hiding this comment.
Many commits are doing much more than what the commit message says - for e.g., "Fix bugs so that the pipelines runs on dc2" in addition to fixing bugs, does string formatting and converting to scientific notation when there's another commit that does the conversion in other places. Some commits are also negative changes from previous commits - for e.g., addition and removal of print("END OF SOURCE TYPE LIST"). It might be easier to get the branch to a working state, squash all commits and split them into smaller pieces.
| lowStats[sourceType] = lowStatsStr | ||
|
|
||
| if np.sum(highSn) > 0: | ||
| highMags[sourceType] = f"{np.nanmax(catPlot.loc[highSn, magCol]):.2f}" | ||
| sortedMags = np.sort(catPlot.loc[highSn, magCol]) | ||
| x = int((len(sortedMags)/100)*10) |
There was a problem hiding this comment.
To me, it looked like you were trying something else here. It might be cleaner to just have this line as int(len(sortedMags)/10) or int(0.1*len(sortedMags)) and have a comment that it's 10%, if you prefer to make it more evident.
arunkannawadi
changed the title
|
Could you use {.3e} instead of {.3g}. The latter seems to switch to scientific notation only in some cases. For example, In [1]: a = 0.000123
In [2]: f"a = {a:.3g}"
Out[2]: 'a = 0.000123'
In [3]: f"a = {a:.3e}"
Out[3]: 'a = 1.230e-04'
In [4]: a = 0.0000123
In [5]: f"a = {a:.3g}"
Out[5]: 'a = 1.23e-05'
In [6]: f"a = {a:.3e}"
Out[6]: 'a = 1.230e-05'The distinction between the two formatting becomes significant if In [7]: a = 456.789
In [8]: f"a = {a:.3g}"
Out[8]: 'a = 457'
In [9]: f"a = {a:.3e}"
Out[9]: 'a = 4.568e+02' |
We actually get integers and decimals with not that many 0s quite often and I find the scientific notation for them very awkward. |
|
Fair enough. This should be good to merge after a rebase. |
No description provided.