Recently I was wondering whether it was possible to automatically export plots from plotly. Unfortunately plotly's documentation is a joke and the module wants you to pay for a feature as basic as this. In the free version only automatic exports in non-vector formats (png) is possible. Here I want to present a workaround relying on phantomjs.
Asked
Active
Viewed 1,131 times
1 Answers
2
Step 1: Generate html file
import plotly.offline as py
import plotly.graph_objs as go
data = go.Scatter(x=[i for i in range(10)],
y=[i**2 for i in range(10)],
mode='lines')
layout = go.Layout(title="First Plot",
xaxis={'title':'x - axis'},
yaxis={'title':'y - axis'})
fig = dict(data=[data], layout=layout)
py.plot(fig,
filename='temp.html',
include_plotlyjs=True,
output_type='file'
)
Note: This includes a great part of plotly.js into the html file. This alone takes up some 2 MB. If need be, change include_plotlyjs
to False
and reference plotly.js inside the html file:
<script src="path/to/plotly.js"></script>
Step 2: Script for PhantomJS
Write a javascript file for phantomjs for exporting the PDF file:
var page = require('webpage').create();
var system = require('system');
var args = system.args;
page.viewportSize = { width: args[3], height: args[4] };
page.clipRect = { top: 0, left: 0, width: args[3], height: args[4] };
page.open(args[1], function() {
page.render(args[2]);
phantom.exit();
});
this expects 4 parameters: The path to the html file, the output file path, width and height of the output file.
Step3: Call PhantomJS script from python
import subprocess
xSize = 1280
ySize = 1024
subprocess.check_output(
['phantomjs', 'phantom_html_to_PDF.js', 'temp.html', out_path, str(x_Size), str(y_Size)])
#remove temporary html file
subprocess.check_output(['rm', 'temp.html])
This calls phantomjs and passes as arguments the path to the javascript from above, the path to the html file, the output file path and the width and height
I hope this is of use for someone

Suppenkasper
- 845
- 4
- 10
- 29
-
Thanks a lot for this! I can't get it to work with 3D images, though. Running phantomjs yields: `TypeError: undefined is not an object (evaluating 't.glplot.canvas')` and the generated pdf only shows title and legend and tells me that my browser does not support Webgl. Any ideas? – mattmilten May 28 '18 at 06:44