Module `Shared_ast`Source

This module defines generic types for types, literals and expressions shared through several of the different ASTs.

Sourcemodule Runtime = Runtime_ocaml.Runtime

Sourcemodule ScopeName : sig ... end

Sourcemodule TopdefName : sig ... end

Sourcemodule StructName : sig ... end

Sourcemodule StructField : sig ... end

Sourcemodule EnumName : sig ... end

Sourcemodule EnumConstructor : sig ... end

Only used by surface

Sourcemodule RuleName : sig ... end

Sourcemodule LabelName : sig ... end

Used for unresolved structs/maps in desugared

Sourcemodule IdentName = Catala_utils.String

Only used by desugared/scopelang

Sourcemodule ScopeVar : sig ... end

Sourcemodule SubScopeName : sig ... end

Sourcemodule StateName : sig ... end

Abstract syntax tree

Define a common base type for the expressions in most passes of the compiler

Sourcetype desugared = [

| `Desugared

]

Phantom types used to select relevant cases on the generic AST

we instantiate them with a polymorphic variant to take advantage of sub-typing. The values aren't actually used.

Sourcetype scopelang = [

| `Scopelang

]

Sourcetype dcalc = [

| `Dcalc

]

Sourcetype lcalc = [

| `Lcalc

]

Sourcetype 'a any = [< desugared | scopelang | dcalc | lcalc ] as 'a

'a any is 'a, but adds the constraint that it should be restricted to valid AST kinds

Types

Sourcetype typ_lit =

| TBool
| TUnit
| TInt
| TRat
| TMoney
| TDate
| TDuration

Sourcetype typ = naked_typ Catala_utils.Marked.pos

Sourceand naked_typ =

| TLit of typ_lit
| TTuple of typ list
| TStruct of StructName.t
| TEnum of EnumName.t
| TOption of typ
| TArrow of typ list * typ
| TArray of typ
| TAny

Constants and operators

Sourcetype date = Runtime.date

Sourcetype duration = Runtime.duration

Sourcetype log_entry =

| VarDef of naked_typ
(*
During code generation, we need to know the type of the variable being logged for embedding
*)
| BeginCall
| EndCall
| PosRecordIfTrueBool

Sourcemodule Op : sig ... end

Classification of operators on how they should be typed

Sourcetype ('a, 'k) operator = ('a any, 'k) Op.t

Sourcetype except =

| ConflictError
| EmptyError
| NoValueProvided
| Crash

Generic expressions

Define a common base type for the expressions in most passes of the compiler

Sourcetype 'a glit =

| LBool : bool -> 'a glit
| LEmptyError : [< desugared | scopelang | dcalc ] glit
| LInt : Runtime.integer -> 'a glit
| LRat : Runtime.decimal -> 'a glit
| LMoney : Runtime.money -> 'a glit
| LUnit : 'a glit
| LDate : date -> 'a glit
| LDuration : duration -> 'a glit

Literals are the same throughout compilation except for the LEmptyError case which is eliminated midway through.

Sourcetype 'a glocation =

| DesugaredScopeVar : ScopeVar.t Catala_utils.Marked.pos * StateName.t option -> desugared glocation
| ScopelangScopeVar : ScopeVar.t Catala_utils.Marked.pos -> scopelang glocation
| SubScopeVar : ScopeName.t * SubScopeName.t Catala_utils.Marked.pos * ScopeVar.t Catala_utils.Marked.pos -> [< desugared | scopelang ] glocation
| ToplevelVar : TopdefName.t Catala_utils.Marked.pos -> [< desugared | scopelang ] glocation

Locations are handled differently in desugared and scopelang

Sourcetype ('a, 't) gexpr = (('a, 't) naked_gexpr, 't) Catala_utils.Marked.t

General expressions: groups all expression cases of the different ASTs, and uses a GADT to eliminate irrelevant cases for each one. The 't annotations are also totally unconstrained at this point. The dcalc exprs, for example, are then defined with type naked_expr = dcalc naked_gexpr plus the annotations.

A few tips on using this GADT:

To write a function that handles cases from different ASTs, explicit the type variables: fun (type a) (x: a naked_gexpr) -> ...
For recursive functions, you may need to additionally explicit the generalisation of the variable: let rec f: type a . a naked_gexpr -> ...
Always think of using the pre-defined map/fold functions in Expr rather than completely defining your recursion manually.

Sourceand ('a, 't) naked_gexpr =

| ELit : 'a glit -> ('a any, 't) naked_gexpr
| EApp : {
1. f : ('a, 't) gexpr;
2. args : ('a, 't) gexpr list;
} -> ('a any, 't) naked_gexpr
| EOp : {
1. op : ('a, _) operator;
2. tys : typ list;
} -> ('a any, 't) naked_gexpr
| EArray : ('a, 't) gexpr list -> ('a any, 't) naked_gexpr
| EVar : ('a, 't) naked_gexpr Bindlib.var -> ('a any, 't) naked_gexpr
| EAbs : {
1. binder : (('a, 't) naked_gexpr, ('a, 't) gexpr) Bindlib.mbinder;
2. tys : typ list;
} -> ('a any, 't) naked_gexpr
| EIfThenElse : {
1. cond : ('a, 't) gexpr;
2. etrue : ('a, 't) gexpr;
3. efalse : ('a, 't) gexpr;
} -> ('a any, 't) naked_gexpr
| EStruct : {
1. name : StructName.t;
2. fields : ('a, 't) gexpr StructField.Map.t;
} -> ('a any, 't) naked_gexpr
| EInj : {
1. name : EnumName.t;
2. e : ('a, 't) gexpr;
3. cons : EnumConstructor.t;
} -> ('a any, 't) naked_gexpr
| EMatch : {
1. name : EnumName.t;
2. e : ('a, 't) gexpr;
3. cases : ('a, 't) gexpr EnumConstructor.Map.t;
} -> ('a any, 't) naked_gexpr
| ETuple : ('a, 't) gexpr list -> ('a any, 't) naked_gexpr
| ETupleAccess : {
1. e : ('a, 't) gexpr;
2. index : int;
3. size : int;
} -> ('a any, 't) naked_gexpr
| ELocation : 'a glocation -> ([< desugared | scopelang ] as 'a, 't) naked_gexpr
| EScopeCall : {
1. scope : ScopeName.t;
2. args : ('a, 't) gexpr ScopeVar.Map.t;
} -> ([< desugared | scopelang ] as 'a, 't) naked_gexpr
| EDStructAccess : {
1. name_opt : StructName.t option;
2. e : ('a, 't) gexpr;
3. field : IdentName.t;
} -> (desugared as 'a, 't) naked_gexpr
(*
desugared has ambiguous struct fields
*)
| EStructAccess : {
1. name : StructName.t;
2. e : ('a, 't) gexpr;
3. field : StructField.t;
} -> ([< scopelang | dcalc | lcalc ] as 'a, 't) naked_gexpr
(*
Resolved struct/enums, after desugared
*)
| EAssert : ('a, 't) gexpr -> ([< dcalc | lcalc ] as 'a, 't) naked_gexpr
| EDefault : {
1. excepts : ('a, 't) gexpr list;
2. just : ('a, 't) gexpr;
3. cons : ('a, 't) gexpr;
} -> ([< desugared | scopelang | dcalc ] as 'a, 't) naked_gexpr
| EErrorOnEmpty : ('a, 't) gexpr -> ([< desugared | scopelang | dcalc ] as 'a, 't) naked_gexpr
| ERaise : except -> (lcalc as 'a, 't) naked_gexpr
| ECatch : {
1. body : ('a, 't) gexpr;
2. exn : except;
3. handler : ('a, 't) gexpr;
} -> (lcalc as 'a, 't) naked_gexpr

Source

type ('a, 't) boxed_gexpr =
  (('a, 't) naked_gexpr Bindlib.box, 't) Catala_utils.Marked.t

The annotation is lifted outside of the box for expressions

Sourcetype 'e boxed = ('a, 't) boxed_gexpr constraint 'e = ('a, 't) gexpr

('a, 't) gexpr boxed is ('a, 't) boxed_gexpr. The difference with ('a, 't) gexpr Bindlib.box is that the annotations is outside of the box, and can therefore be accessed without the need to resolve the box

Sourcetype ('e, 'b) binder = (('a, 't) naked_gexpr, 'b) Bindlib.binder constraint 'e = ('a, 't) gexpr

The expressions use the Bindlib library, based on higher-order abstract syntax

Sourcetype ('e, 'b) mbinder = (('a, 't) naked_gexpr, 'b) Bindlib.mbinder constraint 'e = ('a, 't) gexpr

Markings

Sourcetype untyped = {

pos : Catala_utils.Pos.t;

}

Sourcetype typed = {

pos : Catala_utils.Pos.t;
ty : typ;

}

Sourcetype _ mark =

| Untyped : untyped -> untyped mark
| Typed : typed -> typed mark

The generic type of AST markings. Using a GADT allows functions to be polymorphic in the marking, but still do transformations on types when appropriate. Expected to fill the 't parameter of gexpr and gexpr (a 't annotation different from this type is used in the middle of the typing processing, but all visible ASTs should otherwise use this.

Sourcetype any_expr =

| AnyExpr : (_, _ mark) gexpr -> any_expr

Useful for errors and printing, for example

Higher-level program structure

Constructs scopes and programs on top of expressions. The 'e type parameter throughout is expected to match instances of the gexpr type defined above. Markings are constrained to the mark GADT defined above. Note that this structure is at the moment only relevant for dcalc and lcalc, as scopelang has its own scope structure, as the name implies.

Sourcetype scope_let_kind =

| DestructuringInputStruct
(*
let x = input.field
*)
| ScopeVarDefinition
(*
let x = error_on_empty e
*)
| SubScopeVarDefinition
(*
let s.x = fun _ -> e or let s.x = error_on_empty e for input-only subscope variables.
*)
| CallingSubScope
(*
let result = s ({ x = s.x; y = s.x; ...})
*)
| DestructuringSubScopeResults
(*
let s.x = result.x *
*)
| Assertion
(*
let _ = assert e
*)

This kind annotation signals that the let-binding respects a structural invariant. These invariants concern the shape of the expression in the let-binding, and are documented below.

Sourcetype 'e scope_let = {

scope_let_kind : scope_let_kind;
scope_let_typ : typ;
scope_let_expr : 'e;
scope_let_next : ('e, 'e scope_body_expr) binder;
scope_let_pos : Catala_utils.Pos.t;

} constraint 'e = (_ any, _ mark) gexpr

This type is parametrized by the expression type so it can be reused in later intermediate representations.

Sourceand 'e scope_body_expr =

| Result of 'e
| ScopeLet of 'e scope_let

constraint 'e = (_ any, _ mark) gexpr

A scope let-binding has all the information necessary to make a proper let-binding expression, plus an annotation for the kind of the let-binding that comes from the compilation of a Scopelang.Ast statement.

Sourcetype 'e scope_body = {

scope_body_input_struct : StructName.t;
scope_body_output_struct : StructName.t;
scope_body_expr : ('e, 'e scope_body_expr) binder;

} constraint 'e = (_ any, _ mark) gexpr

Instead of being a single expression, we give a little more ad-hoc structure to the scope body by decomposing it in an ordered list of let-bindings, and a result expression that uses the let-binded variables. The first binder is the argument of type scope_body_input_struct.

Sourcetype 'e code_item =

| ScopeDef of ScopeName.t * 'e scope_body
| Topdef of TopdefName.t * typ * 'e

Sourcetype 'e code_item_list =

| Nil
| Cons of 'e code_item * ('e, 'e code_item_list) binder

Sourcetype struct_ctx = typ StructField.Map.t StructName.Map.t

Sourcetype enum_ctx = typ EnumConstructor.Map.t EnumName.Map.t

Sourcetype scope_out_struct = {

out_struct_name : StructName.t;
out_struct_fields : StructField.t ScopeVar.Map.t;

}

Sourcetype decl_ctx = {

ctx_enums : enum_ctx;
ctx_structs : struct_ctx;
ctx_struct_fields : StructField.t StructName.Map.t IdentName.Map.t;
(*
needed for disambiguation (desugared -> scope)
*)
ctx_scopes : scope_out_struct ScopeName.Map.t;

}

Sourcetype 'e program = {

decl_ctx : decl_ctx;
code_items : 'e code_item_list;

}

Sourcemodule Var : sig ... end

Sourcemodule Type : sig ... end

Sourcemodule Operator : sig ... end

Sourcemodule Expr : sig ... end

Functions handling the expressions of shared_ast

Sourcemodule Scope : sig ... end

Functions handling the code item structures of shared_ast, in particular the scopes

Sourcemodule Program : sig ... end

Sourcemodule Print : sig ... end

Printing functions for the default calculus AST

Sourcemodule Typing : sig ... end

Typing for the default calculus. Because of the error terms, we perform type inference using the classical W algorithm with union-find unification.

Install

Dune Dependency

Authors

Maintainers

Sources

doc/catala.shared_ast/Shared_ast/index.html

Module `Shared_ast`Source

Abstract syntax tree

Types

Constants and operators

Generic expressions

Markings

Higher-level program structure

package catala

Install

Dune Dependency

Authors

Maintainers

Sources

doc/catala.shared_ast/Shared_ast/index.html

Module Shared_astSource

Abstract syntax tree

Types

Constants and operators

Generic expressions

Markings

Higher-level program structure

Module `Shared_ast`Source