Propose Python HIPS as GSoC 2017 idea

_projects/2017/astropy/hips.md

@@ -0,0 +1,51 @@
+---
+name: HIPS to Py
+desc: Write a Python client for Hierarchical Progressive Surveys (HIPS)
+requirements:
+ - Have Python / Numpy / Astropy skills
+ - Know how to write tests with pytest and docs with Sphinx
+ - Some open source experience
+ - Some image processing experience
+ - Interest in spherical geometry and HEALPIX / HIPS
+difficulty: Intermediate to high
+mentors: 
+ - cdeil
+ - tboch
+initiatives:
+ - GSOC
+tags:
+ - python
+ - regions
+ - web
+ - math
+collaborating_projects:
+ - astropy
+---
+
+#### Description
+
+The goal of the [HIPS to Py](https://www.youtube.com/watch?v=DUT5rEU6pqM) GSoC project is to create a
+new Python package to work with [Hierarchical Progressive Surveys (HIPS)](http://aladin.u-strasbg.fr/hips/).
+
+HIPS has become a standard to share large astronomical survey images, using multi-resolution image tiles
+that can be fetched by a client, similar to Google maps.
+So far, a Java client (Aladin) and a JavaScript client ([AladinLite](http://aladin.u-strasbg.fr/AladinLite/)) exist,
+but there's no easy way to fetch HIPS data and make images from Python.
+
+This project would build on [Astropy coordinates](http://docs.astropy.org/en/stable/coordinates/index.html),
+and probably on [healpy](https://healpy.readthedocs.io/)
+and [reproject](https://reproject.readthedocs.io/en/stable/healpix.html),
+and maybe on [regions](http://astropy-regions.readthedocs.io),
+i.e. we would use the functionionality implemented there to implement a HIPS client.
+
+The main feature / deliverable would be a way to take an arbitrary sky image geometry, given by a
+[astropy.wcs.WCS](http://docs.astropy.org/en/stable/wcs/index.html), and to compute a 2-dim numpy array
+where the pixel values are filled from HIPS data. In the background, there should be lower-level functionality
+(also exposed by the package) to fetch and manage HIPS tiles, and to reproject and stitch them together into an image.
+Then one could analyse the data from Python, or save it locally to a FITS or PNG image file.
+
+If you're interested in this project, please have a look at the
+[HIPS paper](http://cdsads.u-strasbg.fr/abs/2015A%26A...578A.114F)
+and the packages mentioned above (they have tutorial docs pages),
+and then contact Christoph if you have any questions or would be interested
+to start working on an application for this project.