whosonfirst.mapzen.com – it's a website.
whosonfirst.mapzen.com – it's a website.
What is going on in here?
This is one of those static website generators, with lowest-common-denominator dependencies. The very short version is that a
Makefile turns concatinates many smaller bits of HTML into a site structure that we can upload to the Internet.
Makefile- a whole lotta
sedthat makes it all go.
pages/- each section of the site has its own sub-makefile (e.g.,
components/- things like headers and footers, navs and subnavs, that sandwich the content.
content/- what it sounds like, these HTML files define what you see in the main part of the page.
www/- where everything gets built to.
Editing the site
- Some files are downloaded from GitHub, so you should edit them there.
- You can build all the pages with
make build-pagesor a section with
make [section]or a specific page
- All the site content, including GitHub-derived pages, is stored in
- The navigation and other template HTML lives in
- Nothing is automatic, you have to edit the navigation files yourself.
You may notice the
Makefile is full of
@ symbols. This hides the command from the output, which makes it easier to observe the site-build progress. Feel free to un-
@ a specific line if you want to see more details.
Example: editing a page from GitHub
Let's say you want to edit the Known Knowns page. This is one of the pages pulled down from GitHub. You can tell because
content/data/knownknowns.html has an HTML comment that links to the Markdown document it's built from.
There are known knowns...
You'll notice that GitHub-derived pages are modified so that all the HTML is on a single line, so editing them directly would be a pain anyway.
- Edit README.KNOWN.KNOWNS.md on GitHub using the pencil button (you could also clone/commit/push your edits, it's up to you).
- Download the updated content HTML:
- Build the page:
Example: editing a page that lives in whosonfirst-www
Most of that same process applies for pages that don't originate from GitHub. For example, to update the main docs page:
content/docs/docs.htmldirectly, then save your file.
- Build the page:
Adding a new page
Look in the
pages folder and open the
.mk file for the section where your new page will live. In this example we'll create a new page "foo" in the "docs" section.
Downloading HTML from GitHub
Note: If your page doesn't "live" on GitHub, you can skip this part.
pages/docs.mk and add an additional line to the
docs-download-content: \ docs-download-properties \ docs-download-brooklynintegers \ ... docs-download-foo
Define your new download target, using a variable for the repo base URL, defined at the top of
```SOME_REPO := https://github.com/whosonfirst/whosonfirst-some-repo/blob/master/
docs-download-foo: @make URL=$(SOME_REPO)pages/foo.md \ OUT=docs/foo.html \ download-content```
This will download HTML from
https://github.com/whosonfirst/whosonfirst-some-repo/blob/master/pages/foo.md and save it to
Test it out by running your new target at the command line:
www page HTML
There are different page "levels" that determine how the templates work.
build-page-level0- everything in
home.mk, no top nav section
build-page-level1- section home pages (e.g., docs)
build-page-level2- sub-pages inside a section (e.g., properties)
build-page-level3- sub-sub-pages (e.g., wof properties)
Each section has its own set of
[section]-build-page-level[n] targets that define common section-wide values.
In the case of our example,
content/docs/foo.html, we will use
docs-build-page-level2 since it's a sub-page inside docs.
docs-foo: @make CONTENT=docs/foo.html \ OUT=docs/foo/index.html \ PAGE_TITLE='foo page title' \ SUBNAV_LINK=foo \ docs-build-page-level2
The variables getting passed are:
CONTENT- path to the file to use inside the
OUT- path to the
www/output file (use the
PAGE_TITLE- ends up in the page
SUBNAV_LINK- the link to highlight in the sidebar subnav
SUBSUBNAV- optional, for linking to level 3 pages (e.g., set to
propertiesin the properties pages)
Test it out at the command line, and see if you can load up your new page:
You should see your new page here:
Adding a blog post
To add a blog post:
- Open content/blog/blog.html file.
- Scroll to the bottom of that file.
- If it is a new year, add a h6 element with updated content and id.
- If you added a h6 element, also add a new unordered list.
If you added a new unordered list, add a list item in that new list.
If you did not add a new unordered list, add the list item after the last list item in the last existing unordered list.
- Add a link to the new list item, be sure sure to include the link url and whosonfirst-sole-link and whosonfirst-dynamically-generated-post as classes.
make blogfrom the command line. The blog page should now be updated with the new blog post.
Updating the subnav
In our last example we created a new page "foo" in the "docs" section. Let's add a link to it in the subnav files.
There are two places we need to add it:
components/subnav/docs/subnav-top.html- for desktop/tablet/etc.
components/subnav/docs/subnav-bottom.html- for mobile
subnav-top.html and add new
Do the same for
Note that because we are using a really dumb
sed replacement, the last
class value should be the same as all the other links.
You may notice that there are some other component files in the
subnav directory. Those are for sub-sections like properties, whose subsubnav is defined by
components/subnav/docs/properties-top.html (for desktop/tablet/etc.) and
components/subnav/docs/properties-bottom.html (for mobile).
These get included by defining a
SUBSUBNAV variable (e.g.,
SUBSUBNAV=properties) when invoking each page build.
When defined, the subsubnav HTML component gets inserted into the subnav based on an HTML comment
For example, here's how the
properties subsubnav gets placed within
Building the site
make- download and build all the pages
make docs- download and build pages in the docs section
make download- download all the content
make build-pages- build all the pages
Or, if you want to deploy to just dev or prod S3 buckets:
make www-devmake www-prod
Cloning to S3
We are using a home-grown tool called
wof-clone-website which is part of the go-whosonfirst-www repo, mostly so that file permissions are set correctly. There is a binary copy for OS X included in this repo so if you're trying to deploy things from not-a-Mac then you will need to build an OS-specific version of the tool, which is left as an exercise to the reader.
Thanks to 2017 summer intern Scott Dombkowski for designing and building the new site.
To restore the repository download the bundle
git clone whosonfirst-whosonfirst-www_-_2017-10-26_05-15-17.bundle
Upload date: 2017-10-26
- 2017-10-26 05:15:17
- Internet Archive Python library 1.7.3
- iagitup - v1.2