Upgrade draw_map.py

[alexmitchell]

• [x] Use whoami to figure out player and draw only the player's map. Command line argument can optionally draw the full map.
  • Hardcode who is player 1 and 2
• [x] Tiles that a player controls have their background color
• [x] Show tile ids and resources for explored or controlled territories
• [x] Show tile ids only for nodes adjacent to explored or controlled tiles
• [x] Don't show (delete) nodes that aren't explored
• [x] Change node shape for tiles with a battle. e.g doubleoctagon or polygon
• [ ] Scale map so that nodes don't overlap (see previous commit a044286)

Wishlist:

• [ ] Can use graphvis label html to create tables for the labels to keep all the information organized.
• [x] Start up a application to view the image (use subprocess module)

Docs for shape of the node: https://graphviz.org/docs/attrs/shape/

Docs for labels: https://graphviz.org/docs/attrs/label/

[alexmitchell]

@kalekundert Is there a reason the "control" attribute of map nodes is a list rather than an integer? The attribute is created in the compose_map function in seacow.py (lines 354-382 in my version). There can only be one owner at a time, correct?

If we keep the list, would the last element be the current owner?

[kalekundert]

The list is needed to know which player controlled the tile on past turns. For example, if I were to explore a tile on turn 3, and you were to take control of it on turn 4, my map should continue to show it as uncontrolled, because that's how it was at the time I explored it. In order for this to work, the game needs to track (i) what turn each explore action was taken and (ii) what turn any visible changes to the map happened.

The list is a collection of Control objects, where each control object has turn and player attributes, so the current owner is map.nodes[tile]['control'][-1].player. (Of course, the list could also be empty, indicating that neither player controls the tile.) This logic is implemented by seacow.who_controls(), though, so you don't need to do it yourself.

Now that I think about it, there should also be a variant on that function, maybe called seacow.who_controlled(), that calculates who controlled a tile on a given turn. That would be useful for some of this map stuff.

[alexmitchell]

Great! Thanks for the fast clarification! I see why a list is needed.

I completely missed that there is a Control class in the file. haha