Initial commit of generating random data #60
Conversation
|
If you want the commit to be tied to your GitHub username, you should follow these instructions: https://help.github.com/articles/why-are-my-commits-linked-to-the-wrong-user/ Not mandatory, but might be nice to give you proper credit. Once you do that, you can just |
|
Also, if you have time, there are style/linting issues which you can see here: https://travis-ci.org/frewsxcv/python-geojson/jobs/69054890 If you don't get around to them, I can fix them myself later |
|
Hi @frewsxcv Anyway let me know how you're getting on, no rush of course |
|
Before: def generate_random(featureType, numberFeatures=1, numberVertices=3, boundingBox=[-180.0, -90.0, 180.0, 90.0]): After: +def generate_random(featureType, numberFeatures=1, numberVertices=3,
boundingBox=[-180.0, -90.0, 180.0, 90.0]): |
|
More info about indentation from the official style guide |
|
Hi @frewsxcv Sorry for the delay in progress, just a quick update, I've been a bit stuck on the random polygon generation, my original idea of simply sorting the coords failed (given hindsight that was kind of obvious!). The geojson-random library from mapbox contains code to generate random polys although Im having a hard time understanding the maths, I think I need to sit down with a piece of paper to fully understand whats going on. Cheers |
|
If you want, we can merge in a basic implementation of the random geometry algorithm you have now, then when you understand the math more, open a new PR improving it |
|
Hi @frewsxcv well I think I've pretty got things working now at a basic level, hooray! For the polygon creation I took some code from here although this code wasn't designed for dealing with lat lons so its not working with the bounding box properly yet, I still need to finalise the maths around it but it's at least working for the timebeing. This linting business took a while I get my head around, I've think I've now got a decent linter installed so fingers crossed things are looking better now! Anyway let me know what you think, I hope it forms a useful part of the library. Cheers |
|
I haven't forgotten about this! I'll be returning home from my month long traveling in a couple days. Hopefully I'll look at this later this week |
| a linestring or polygon will have | ||
| :type int: defaults to 3 | ||
| :param boundingBox: A bounding box in which features will be restricted to | ||
| :type string: defaults to the world - [-180.0, -90.0, 180.0, 90.0] |
There was a problem hiding this comment.
Should this be :type list:?
There was a problem hiding this comment.
Not sure @frewsxcv :)
are you referring to a particular line?
Your comment is a bit cryptic to this novice python-er
There was a problem hiding this comment.
With GitHub Pull Request, you can make comments on specific lines. The comment I made above is referring to line number 74. You can see this here. Let me know if you don't understand.
There was a problem hiding this comment.
Ahh for some reason lines 59-74 came through with the comment which is why I was confused.
Re your question I'll take it your suggesting that it should be a list so i've committed again :)
|
@rowanwins Looks good! You're welcome to add yourself under the Credits in the README.md if you want before I merge this |
|
Hi @frewsxcv , happy not to be credited I do have one final question which it would be good to get your opinion on. I've been debating removing the multi-feature support. The current approach locks people into a geometrycollection and doest allow properties. It may be easier to let people write the loop themselves if they want multi-features. What do you think? FYI I I just used this to generate 100,000 random points with properties in a feature collection using this approach, took ~5s to run. Very handy for test data! |
|
I think it's okay to leave out the multi-feature support; writing a loop should be simple enough, though it's up to you whether you want to keep it or not
:D |
…rating their own loops
|
@frewsxcv have removed the multi-feature support and re-committed, this is it I reckon... |
|
Looks great @rowanwins, thanks! |
Initial commit of generating random data
|
Just released 1.3.0 that includes this change https://pypi.python.org/pypi/geojson/1.3.0 Thanks again! |
1.3.0 (2015-08-11) ------------------ - Add utility to generate geometries with random data - jazzband/geojson#60
Hi @frewsxcv
As suggested here is a pull request for work initial work on generating random data.
Cheers
Rowan