All notable changes to this project will be documented in this file. This project adheres to Semantic Versioning.

• Previously, using plotly offline required a paid license. No more: plotly.js is now shipped inside this package to allow unlimited free use of plotly inside the ipython notebook environment. The plotly.js library that is included in this package is free, open source, and maintained independently on GitHub at https://github.com/plotly/plotly.js.
• The plotly.js bundle that is required for offline use is no longer downloaded and installed independently from this package: plotly.offline.download_plotlyjs is deprecated.
• New versions of plotly.js will be tested and incorporated into this package as new versioned pip releases; plotly.js is not automatically kept in sync with this package.

• Big data warning mentions plotly.graph_objs.Scattergl as possible solution.

• If you're behind a proxy, you can make requests by setting the environmental variable HTTP_PROXY and HTTPS_PROXY (http://docs.python-requests.org/en/v1.0.4/user/advanced/#proxies). This didn't work for streaming, but now it does.

• Sometimes creating a graph with a private share-key doesn't work - the graph is private, but not accessible with the share key. Now we check to see if it didn't work, and re-try a few times until it does.

• The FigureFactory can now create dendrogram plots with .create_dendrogram.

• Saving "world_readable" to your config file via plotly.tools.set_config actually works.

• You can also save auto_open and sharing to the config file so that you can forget these keyword argument in py.iplot and py.plot.

• Fixed validation errors (validate=False workaround no longer required)

• Auto-sync API request on import to get the latest schema from Plotly
• .-access for nested attributes in plotly graph objects
• General .help() method for plotly graph objects
• Specific attribute .help(<attribute>) also included

• No more is streamable, streaming validation.

• Fixed typos in plot and iplot documentations

• CHANGELOG
• sharing keyword argument for plotly.plotly.plot and plotly.plotly.iplot with options 'public' | 'private' | 'secret' to control the privacy of the charts. Depreciates world_readable

• If the response from plot or iplot contains an error message, raise an exception

• height and width are no longer accepted in iplot. Just stick them into your figure's layout instead, it'll be more consistent when you view it outside of the IPython notebook environment. So, instead of this:

   py.iplot([{'x': [1, 2, 3], 'y': [3, 1, 5]}], height=800)
  

  do this:

   py.iplot({
   	'data': [{'x': [1, 2, 3], 'y': [3, 1, 5]}],
   	'layout': {'height': 800}
   })
  

• The height of the graph in iplot respects the figure's height in layout