# Boring news about Federico - June 2016

Go forward in time to July 2016.

Tue 2016/Jun/28
• It is Hack Week at SUSE, and I am working on La Mapería (the map store), a little program to generate beautiful printed maps from OpenStreetMap data.

I've gotten to the point of having something working: the tool downloads rendered map tiles, assembles them with Cairo as a huge PDF surface, centers the map on a sheet of paper, and prints nice margins and a map scale. This was harder to me than it looks: I am pretty good at dealing with pixel coordinates and transformations, but a total newbie with geodetic calculations, geographical coodinate conversions, and thinking in terms of a physical map scale instead of just a DPI and a paper size.

The resulting chart has a map and a frame with arc-minute markings, and a map scale rule. I want to have a 1-kilometer UTM grid if I manage to wrap my head around map projections.

## Coordinates and printed maps

The initial versions of this tool evolved in an interesting way. Assembling a map from map tiles is basically this:

1. Figure out the tile numbers for the tiles in the upper-left and the lower-right corners of the map.
2. Composite each tile into a large image, like a mosaic.

The first step is pretty easy if you know the (latitude, longitude) of the corners: the relevant conversion from coordinates to tile numbers is in the OpenStreetMap wiki. The second step is just two nested for() loops that paste tile images onto a larger image.

When looking at a web map, it's reasonably easy to find the coordinates for each corner. However, I found that printed maps want one to think in different terms. The map scale corresponds to the center of the map (it changes slightly towards the corners, due to the map's projection). So, instead of thinking of "what fits inside the rectangle given by those corners", you have to think in terms of "how much of the map will fit given your paper size and the map scale... around a center point".

So, my initial tool looked like

```python3 make-map.py
--from-lat=19d30m --from-lon=-97d
--to-lat=19d22m --to-lon=-96d47m
--output=output.png```

and then I had to manually scale that image to print it at the necessary DPI for a given map scale (1:50,000). This was getting tedious. It took me a while to convert the tool to think in terms of these:

• Paper size and margins
• Coordinates for the center point of the map
• Printed map scale

Instead of providing all of these parameters in the command line, the program now takes a little JSON configuration file.

La Mapería generates a PDF or an SVG (for tweaking with Inkscape before sending it off to a printing bureau). It draws a nice frame around the map, and clips the map to the frame's dimensions.

La Mapería is available on github. It may or may not work out of the box right now; it includes my Mapbox access token — it's public — but I really would like to avoid people eating my Mapbox quota. I'll probably include the map style data with La Mapería's source code so that people can create their own Mapbox accounts.

Over the rest of the week I will be documenting how to set up a Mapbox account and a personal TileStache cache to avoid downloading tiles repeatedtly.

Go backward in time to May 2016.