Implement the part of the language in black. No need to implement code generation for the red parts. No test for this task will feature programs containing language constructs in red. Your code generator can throw CodegenException for the ASTs that correspond to language constructs in red, or you can return RISC-V code, it does not matter.

PROG → DEC | DEC; PROG 
DEC → def ID (VARDEC) = E
VARDEC →  ε | VARDECNE 
VARDECNE → ID | VARDECNE, ID 
ID → ... (identifiers)
INT → ... (Integers)
E →  INT 
  | ID 
  | if E COMP E then E else E endif
  | (E BINOP E)
  | (E)
  | ID(ARGS)
  | skip
  | (E; E)
  | while E COMP E do E endwhile
  | repeat E until E COMP E endrepeat
  | ID := E
  | break
  | continue
ARGS → ε | ARGSNE
ARGSNE → E | ARGSNE, E 
COMP → == | < | > | <= | >= 
BINOP → + | - | * | / 

Recall that the relevant definitions are here, here and here (and in the Zip file / Github given on the parent page). If you don't want to implement a feature, simply throw CodegenException when the code generator encounters this feature.

Note that the translation of procedure invocation (given by the production E → ID(ARGS) in the grammar above) is part of Task 1. Since every program will have at least one procedure invocation, I strongly suggest that you focus on getting this right.