Download APOGEE data files and StarNet model

In order to go through the notebooks you will need a few files. Depending on the intended usage of the following notebooks, not all of these files are necessary. Below is a description of each file and which notebooks it is necessary for.

file	size	content
apStar_visits_main.h5	35.9GB	apStar individual visit spectra for training StarNet, created from the APOGEE DR13. Used in 2_Preprocessing_of_Training_Data.ipynb
apStar_combined_main.h5	9.6GB	apStar combined spectra for preprocessing StarNet, created by pulling apStar combined spectra from the APOGEE DR13. Used in 3_Preprocessing_of_Test_Data.ipynb
training_data.h5	1.2GB	apStar individual visit spectra for training, useful to skip preprocessing. Used in 3_Preprocessing_of_Test_Data.ipynb and 4_Train_Model.ipynb
mean_and_std.npy	104B	mean and standard deviations for the stellar labels used during preprocessing. Used in 4_Train_Model.ipynb and 5_Test_Model.ipynb
test_data.h5	1.13GB	apStar combined spectra test set. Used in 5_Test_Model.ipynb
starnet_cnn.h5	85MB	pretrained StarNet model with keras (tensorflow as the backend) on APOGEE DR13. Used in 5_Test_Model.ipynb

You can download the data from the StarNet public VOSpace at CADC. You can either:

browse and download from this URL
or download the files from the python VOSpace command line client, installed with pip:
```
pip install vos
```
```
getcert
```
You can choose a directory with enough space (~50GB) and download all files into a directory:
```
vcp vos:starnet/public /path/to/my/starnet/directory
```
Or you can copy each file within an IPython session with a function such as the one below:



In [1]:

    
import vos
datadir=""

def starnet_download_file(filename):
    vclient = vos.Client()
    vclient.copy('vos:starnet/public/'+filename, datadir+filename)
    print(filename+' downloaded')



In [2]:

    
# for example:
starnet_download_file('starnet_cnn.h5')









    



starnet_cnn.h5 downloaded