Algebraic data types: Difference between revisions

Line 1,105:

jq does not have built-in support for pattern matching in the sense of the present task description, but the following `bindings` function takes advantage of the way in which singleton-key JSON objects can be used as variables for pattern-matching. In effect, jq expressions such as `{a}`

can be used as variables in the pattern definitions, and after matching, the corresponding values can be referenced by jq expressions such as `.a`.

Notice also how various features of jq come together to simplify the implementation of the `balance` function.

'''bindings.jq'''

Line 1,141:

Line 1,143:

<lang jq>include "bindings" {search: "."};

⚫

def E: []; # the empty node

# Each nonempty node is an array: [Color, Left, Value, Right]

# where Left and Right are nodes.

Line 1,146:

Line 1,149:

def B: "⚫";

def R: "🔴";

⚫

def E: []; # the empty node

def b(x): bindings({} | x) // empty;

Line 1,155:

Line 1,156:

def node: [R, [B, .a, .x, .b], .y, [B, .c, .z, .d]];

(b([B, [R, [R, {a}, {x}, {x}], {y}, {c}], {z}, {d}]~~) | node~~)

( b([B, [R, [R, {a}, {x}, {x}], {y}, {c}], {z}, {d}])

// (b([B, [R, {a}, {x}, [R, {b}, {y}, {c}]], {z}, {d}]~~) | node~~)

// b([B, [R, {a}, {x}, [R, {b}, {y}, {c}]], {z}, {d}])

// (b([B, {a},{x}, [R, [R, {b}, {y}, {c}], {z}, {d}]]~~) | node~~)

// b([B, {a},{x}, [R, [R, {b}, {y}, {c}], {z}, {d}]])

// (b([B, {a},{x}, [R, {b}, {y}, [R, {c}, {z}, {d}]]]~~)| node~~)

// b([B, {a},{x}, [R, {b}, {y}, [R, {c}, {z}, {d}]]])

| node) // . ;

// (b([{col}, {a}, {x}, {b}])

| [.col, .a, .x, .b ]) ;

# Input: a node