## http://heliopropa.irap.omp.eu

# What

## Web Server

Build and serve heliopropa's visualizations.
It's a `flask` webserver, serving `d3.js` plots.
It also gathers NetCDF data from AMDA, and serves it as CSV to the plotter.


### Rationale

- Reading NetCDF from javascript is doable, but still *very* hacky.
- The bridge can handle pagination and collecting multiple NetCDF into one CSV.


# How

## Source

https://gitlab.irap.omp.eu/CDPP/SPACEWEATHERONLINE


## Overview

- `config.yml` : the main configuration file.
- `web/run.py` : the front controller, holding most of the code.
- `web/view/home.html.jinja2` : the HTML template.
- `web/static/js/swapp.ls` : most of the javascript client-side.


## Install

``` bash
virtualenv venv
source venv/bin/activate
pip install matplotlib
pip install numpy==1.8.*
pip install -r requirements.txt
```

### CentOs 6 Shenanigans

``` bash
yum install https://centos6.iuscommunity.org/ius-release.rpm
yum install python27*
```

But then `virtualenv` yields

```
File "/var/www/html/HELIOPROPA/web/run.py", line 24, in <module>
[Fri Nov 30 14:30:38 2018] [error] [client 10.10.135.119]     from netCDF4 import Dataset, date2num
[Fri Nov 30 14:30:38 2018] [error] [client 10.10.135.119]   File "/var/www/html/HELIOPROPA/venv/lib/python2.7/site-packages/netCDF4/__init__.py", line 3, in <module>
[Fri Nov 30 14:30:38 2018] [error] [client 10.10.135.119]     from ._netCDF4 import *
[Fri Nov 30 14:30:38 2018] [error] [client 10.10.135.119] ImportError: /var/www/html/HELIOPROPA/venv/lib/python2.7/site-packages/netCDF4/_netCDF4.so: undefined symbol: PyUnicodeUCS2_FromStringAndSize
```

_Solution:_
Configure Apache to use `web/run_ssa.wsgi` and use pip2.7 install system wide.


### Install libcdf.

``` bash
wget https://spdf.sci.gsfc.nasa.gov/pub/software/cdf/dist/cdf37_0/linux/cdf37_0-dist-all.tar.gz
tar -xzvf cdf37_0-dist-all.tar.gz
cd cdf37_0-dist
make all OS=linux ENV=gnu
make INSTALLDIR=/usr/local/lib/libcdf install
```

## Develop

``` bash
source venv/bin/activate
python web/run.py
```

Then, browse [localhost:5000](http://localhost:5000).


## Troubleshoot

### Installing scipy

Getting `MemoryErrors` ?
Try with the `--no-cache-dir` option.

`pip install scipy --no-cache-dir`

### Fix permissions problems with the cache

```
sudo setfacl  -R -m u:apache:rwx -m u:`whoami`:rwx cache
sudo setfacl -dR -m u:apache:rwx -m u:`whoami`:rwx cache
```