feat: Combine plots #24
Conversation
|
This pull request introduces 1 alert when merging f6919de into 11fc554 - view on LGTM.com new alerts:
|
This code can remove the warning information on std console. If we don't use it, I will get warning information like this on my PC @matthewfeickert Do we need to reserve this code? |
No. This doesn't matter and is just normal for TensorFlow. |
Codecov Report
@@ Coverage Diff @@
## master #24 +/- ##
==========================================
- Coverage 65.52% 61.91% -3.61%
==========================================
Files 9 10 +1
Lines 467 470 +3
Branches 57 66 +9
==========================================
- Hits 306 291 -15
- Misses 143 160 +17
- Partials 18 19 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
| def subplot(y_label, column, output, directory, filename): | ||
|
|
||
| x_value = output["_runtime"] | ||
| if y_label == "Network Traffic (bytes)": | ||
| y_value1 = output.get(column[0], [0] * len(x_value)) | ||
| y_value2 = output.get(column[1], [0] * len(x_value)) | ||
| plt.plot(x_value, y_value1, ls="--", label="send") | ||
| plt.plot(x_value, y_value2, label="recv") | ||
| plt.legend(loc="upper left") |
There was a problem hiding this comment.
All the plotting code should get switched over to using the matplotlib subplots API that I sent you earlier.
There was a problem hiding this comment.
@matthewfeickert So all the graph plots will be in one picture?
There was a problem hiding this comment.
No. There are lots of examples on using matplotlib.pyplot.subplots. Start with the docs https://matplotlib.org/3.3.0/api/_as_gen/matplotlib.pyplot.subplots.html and then there are tons of results on google.
There was a problem hiding this comment.
@coolalexzb Okay, following up here finally (sorry). What I mean here is that using the plot.plot API can be problematic as it is global and so you have to worry about dealing with a universal canvas. It is more robust to use the plt.subplots API. So you could simply create a function with a more descriptive name that takes similar arguments but then generates a figure and axis pair from
fig, axis = plt.subplots()
ax.set_xlabel("Time (seconds)")
ax.set_ylabel(y_label)
# ...
return fig, axisand soforth. This is a more robust solution and then actually allows you to return the fig and axis objects
fig, axis = your_function(y_label, column, output)
# Do other things with fig and axis if you need
fig.savefig(directory / filename)
|
@coolalexzb There were things in here that didn't get properly rebased and now need to get removed again. Let's start having me review PRs from now on. |
This PR aims to provide solution to issue#21.
User can use the command
to run severalcomputations in one-time run and get the combination plots of running results and separation plots of running result like the following:
