Sort out the tokenizer and tree-construction tests

[gsnedders]

This:

• Gives us a single definition of all our tree types for the testsuite in support
• Moves tree walker tests over to being developed from tree_construction giving them more meaningful ids
• Moves tokenizer tests to being developed from a new tokenizer.py

[landscape-bot]

Code quality remained the same when pulling 7357bea on gsnedders:tree_types into 3b7e017 on html5lib:master.

[landscape-bot]

Code quality remained the same when pulling fbb5ae0 on gsnedders:tree_types into 3b7e017 on html5lib:master.

[landscape-bot]

Code quality remained the same when pulling fbb5ae0 on gsnedders:tree_types into 3b7e017 on html5lib:master.

[codecov-io]

Current coverage is 89.20%

Merging #240 into master will increase coverage by +0.28%

@@             master       #240   diff @@
==========================================
  Files            50         50          
  Lines          6675       6713    +38   
  Methods           0          0          
  Messages          0          0          
  Branches       1310       1314     +4   
==========================================
+ Hits           5935       5988    +53   
+ Misses          569        555    -14   
+ Partials        171        170     -1   

1. 3 files (not in diff) in html5lib were modified. more
   • Misses -9
   • Partials -3
   • Hits +11

Powered by Codecov. Last updated by b5b91cc...bfe955c

[landscape-bot]

Code quality remained the same when pulling 8e1bef2 on gsnedders:tree_types into 3b7e017 on html5lib:master.

[landscape-bot]

Code quality remained the same when pulling bfe955c on gsnedders:tree_types into 3b7e017 on html5lib:master.

[landscape-bot]

Code quality remained the same when pulling 68d6f34 on gsnedders:tree_types into b5b91cc on html5lib:master.