ipython-books / cookbook-2nd Goto Github PK

Hi,
nice work, however I believe that the KDE example in 7.6 is not quite correct, with the errors apparent if you choose any other projection than Plate Carree (x=longitude, y=latitude).

Specifically the first call to scatter:

ax.scatter(x, y, color='r', s=.5, alpha=.25)

should be something like:

ax.scatter(x, y, color='r', s=.5, alpha=.25, transform=ccrs.Geodetic())

as your source data is in lon/lat.

If you were to replace the whole cell with:

mcrs = ccrs.Mercator()
ax = plt.axes(projection=mcrs)
ax.stock_img()
ax.scatter(x, y, color='r', s=.5, alpha=.25, transform=ccrs.Geodetic())

then the data does not display correctly without the transform argument.

In the cell with:

geo = ccrs.Geodetic()
h = geo.transform_points(crs, x, y)[:, :2].T

you are actually transforming the data from Plate Carree to Geodetic, not the other way around. This, again, is only working because x=lon, y=lat.

Hope this make sense,
Regards,
Chris.

Hi Cyrille,
in https://ipython-books.github.io/47-implementing-an-efficient-rolling-average-algorithm-with-stride-tricks/ the bullet point no.8, benchmark of copying method, has 15 μs for copying only and 10 μs for copying and taking the mean.
Regards ፨ gb

Hello, the code does not function for me when run in Jupyter Notebook. I replaced to_datetime with to_pydatetime. That fixed one error but now I am getting an error with temp_fft = sp.fftpack.fft(temp) it is throwing the error AttributeError: 'Series' object has no attribute 'flags'

I have updated Conda and scipy to the latest available and the error persists.
Thank you
-A

Hello, I have tried the synthetizer notebook, but it failed to work, both on my own computer and with the binder link provided. More precisely, while a call to "synth(440)" in an individual cell works as expected, clicking on the buttons, however, does not seem to trigger anything.
I made a workaround as follows.
First I created
out=widgets.Output()
Then instead of with widgets.Output() I wrote with out.
FInally in an another cell I displayed the output:
display(out)
This, however, creates a new widget each time a button is clicked. It can be circumvented by using clear_output(wait=True) after the call to 'synth', but this works only because here the duration of the sound is very short.
I'm interested in this, so if you have a better way of playing a sound (or several at the same time) without displaying unnecessary widgets, I'm all ear ;).

Nathaniel

Hallo some1 or two weeks ago , januari 2019I download all and am very content what you (?) did.
I found a small error in 6.1. Using matplotlib styles
matplotlib is imported as mpl but a few sections later
matplotlib was used in place of mpl ==>missing package
with a consequence of a missing () a bit later

Greets (a new beginner but very enthousiast about it)
PKHG = Peter K.H. Gragert

Hi there, it seems the Binder button for the notebooks isn't working:

Step 41/51 : RUN conda env update -p ${NB_PYTHON_PREFIX} -f "environment.yml" && conda clean --all -f -y && conda list -p ${NB_PYTHON_PREFIX}
---> Running in 9349de55aa95
Collecting package metadata (repodata.json): ...working... done
Solving environment: ...working... failed

ResolvePackageNotFound:

• pyproj==1.9.5.1=py36_0
• holoviews==1.9.0=py36h01658e4_0
• cudatoolkit==8.0=3
• util-linux==2.21=0
• proj4==4.9.3=5
• libiconv==1.15=0
• geos==3.6.2=1
• libdap4==3.19.2=1
• libpq==9.5.4=0
• libgdal==2.0.0=1

Removing intermediate container 9349de55aa95
The command '/bin/sh -c conda env update -p ${NB_PYTHON_PREFIX} -f "environment.yml" && conda clean --all -f -y && conda list -p ${NB_PYTHON_PREFIX}' returned a non-zero code: 1