jamesmcm / wwf-solver Goto Github PK

Letter distribution: A=9 D=5 G=3 J=1 M=2 P=2 S=5 V=2 Y=2 B=2 E=13 H=4 K=1 N=5 Q=1 T=7 W=2 Z=1 C=2 F=2 I=8 L=4 O=8 R=6 U=4 X=1 *=2

Letter values: A=1 D=2 G=3 J=10 M=4 P=4 S=1 V=5 Y=3 B=4 E=1 H=3 K=5 N=2 Q=10 T=1 W=4 Z=10 C=4 F=4 I=1 L=2 O=1 R=1 U=2 X=8 *=0

DONE:

• Horizontal scan , without hooking - i.e. where you play one letter horizontally to form a word whilst also playing letters vertically to form another word) - supporting hooking requires a single vertical scan to be conducted in the case that only a single letter is played.

• The Horizontal scan relies upon the splitRow method which splits a row into substrings separated by white space and returns the relevant information. Construction of a vertical scan method will require an analogous 'splitColumn' method.

• Have commented on splitRow, and begun looking at boardScan but cannot proceed without internet access as I cannot look up the libraries.

• Planned out a method for splitColumn that will give the same results as splitRow but hopefully be more elegant via using the re library - cannot code without internet.

TODO:

• As mentioned above first an analogous column method to splitRow must be constructed.

• This will then allow for the construction of the vertical scan method, which can be used to support hooking in the horizontal scan and also completes the scanning part of the code, in order to support hooking a single horizontal scan would have to be called in the case of a single letter played vertically

• Because of the requirements of hooking, and to make the code more understandable, it seems to be strongly advisable to break the horizontal scan and vertical scan into 2 methods, which are then called by the boardScan method.

• Finally the splitRow method should probably be improved upon to make it clearer via additional comments and refactoring to explain what is going on, this would also make the construction of the analogous column method far easier.

• That appears to be all that is necessary to have a basic working solver - I would strongly recommend that once such a solver is complete the python debugger/profiler is used to check for bottlenecks and any obvious optimisation is conducted

• The scoring function also needs to be written (this looks like it might already be done... is it finished??) although that shouldn't be too difficult? Just remember the special cases like the 35 point bonus for using all the letters, and the special tiles etc., we also need to handle blank tiles which is a difficulty...

• There is a blank function called checkWordplay in the code, but I don't see why it is necessary as surely splitRow/column would return the number of spaces available and that can be used to determine whether placement is possible? But perhaps placing such a check in it's own method is better I don't know.

• Also need to handle the special case that it is the first turn (i.e. the board is blank, have to place a tile on the centre?), but this doesn't seem like it should be too hard...