Add helper script to get list of most frequent values for a key from taginfo #192
Conversation
…aginfo The idea is that this script would help to define the values that need to be rendered, finding the most frequent ones and avoiding explicit negatives and misspellings. This commit just adds the script, but doesn't make actual use of it anywhere yet.
|
I think this is a useful script, but I'd prefer to see it developed independently (and therefore stored in an independent repo) from these stylesheets. It'll be useful to other people for other stylesheets too, and I prefer to keep projects separated. |
|
Thanks for the review. Keeping the script in the style's repo makes sense if we also store the script's parameters in the repo, but I understand that we don't want to do that yet. Puting the script in some other "bag of tricks" repo would be good too, but I have no idea of which repo would fit that description ? A repo for a lone script seems silly. |
|
I wouldn't worry too much - most single scripts generally evolve into bigger things as time goes on! I don't have any suggestions for existing repos where it might be at home - perhaps you could ask on the [email protected] mailing list for suggestions. |
|
you could use osmlab. I can set up a repo there if you want. |
The idea is that this script would help to define the values that need to be rendered,
finding the most frequent ones and avoiding explicit negatives and misspellings.
This commit just adds the script, but doesn't make actual use of it anywhere yet.
The idea follows discussions on #116 and #182.
@gravitystorm : when I said that 'rendering "all" values of a tag is often desirable', the quotes were important : We clearly do not want to render misspelings and "no" values. So we use a whitelist and avoid catch-alls.
But for some keys like 'shop' or 'tunnel', that whitelist can become large and unwieldy. This script just makes the creation and maintenance of those whitelists easyer, based on the observations that cutting off the long tail and some frequent-but-unwanted values makes a pretty nice list.
Of course, the script's output must be reviewed. But eventually, instead of storing the actual output, we could just store the parameters (key, threshold, blacklist). That's to be done on a case by case basis.