Scipy Suggestion: Improved Redirects

[pjhampton]

Description

This IA is pretty popular, and it is really well developed. The problem with it is that people are not searching for the full path, such as scipy fftpack fft2, they are searching for the likes scipy fft2. This is problematic if there is function naming collisions. I don't think there is, so it should be ok to take the last section of the path and create a redirect.

Triggering

Not Triggering

Get Started

• 1) Claim this issue by commenting below
• 2) Review our Contributing Guide
• 3) Set up your development environment, and fork this repository
• 4) Create a Pull Request

Resources

• Join DuckDuckHack Slack to ask questions
• Join the DuckDuckHack Forum to discuss project planning and Instant Answer metrics
• Read the DuckDuckHack Documentation for technical help

---

Instant Answer Page: https://duck.co/ia/view/scipy

[ashishbhate]

I'll take a look at this. I don't think there are collisions either, but if there are I can just skip those redirects.

[ashishbhate]

I was wrong, there are lot of collisions. I'm not sure I should be skipping collisions.

E.g. add_points
['scipy', 'spatial', 'ConvexHull', 'add_points']
['scipy', 'spatial', 'Delaunay', 'add_points']
['scipy', 'spatial', 'Voronoi', 'add_points']

antiderivative:
['scipy', 'interpolate', 'Akima1DInterpolator', 'antiderivative']
['scipy', 'interpolate', 'BPoly', 'antiderivative']
['scipy', 'interpolate', 'CubicSpline', 'antiderivative']
['scipy', 'interpolate', 'InterpolatedUnivariateSpline', 'antiderivative']
['scipy', 'interpolate', 'LSQUnivariateSpline', 'antiderivative']
['scipy', 'interpolate', 'PPoly', 'antiderivative']
['scipy', 'interpolate', 'PchipInterpolator', 'antiderivative']
['scipy', 'interpolate', 'UnivariateSpline', 'antiderivative']
['scipy', 'interpolate', 'pchip', 'antiderivative']
['scipy', 'interpolate', 'ppform', 'antiderivative']

This is just a small sample; there are a lot more collisions. I dont have the data on the search terms that are not triggering, but if it's a small list we can just manually add them.

[pjhampton]

Hmm, thanks for reporting back @ashishbhate. I think for now we should skip any collisions.

Maybe we should take note of them though in the cover/ dir in a file called uncovered.txt or noredirect.txt - this could give us a starting point to a new issue once we decide on how to go forward. What do you think?

cc. @moollaza

[ashishbhate]

Generating a list of collisions is quite easy so I don't think we need to take save them specifically. There are also a lot of them.

Based on the search data analytics we could generate redirects for the most popular search terms. E.g. if we know that a lot of people search for scipy fft2 instead of scipy fftpack fft2 we could just manually add this redirect to a file. The parser, while parsing the scipy docs, could read this file, check that the redirect is valid (i.e. scipy fftpack fft2 article exists) and include that redirect.