Chimpanzee is a library written in TypeScript with which you can write a "Schema" for traversing trees structures and capturing values of certain nodes. This is especially useful for analyzing ASTs, which can be fairly large and complex. Schemas can be composed, allowing common structures to be abstracted out.
npm install chimpanzee
import { match, types, capture } from "chimpanzee";
const input = {
hello: "world"
}
const schema = {
hello: capture()
}
const result = match(schema, input);
//result is Match { value: { hello: "world" } }
The result of a match is one of four types.
- Match
- value: object
- Skip
- message: string
- Fault
- message: string
- Empty (is a Match, but has no value)
import { Match } from "chimpanzee";
const input = {
hello: "world"
}
const schema = {
hello: item => context => new Match(`${item}!!!`)
}
const result = match(schema, input);
//result is Match { value: { hello: "world!!!" } }
Allows you to capture the value of a node.
const input = {
hello: "world"
}
const schema = {
hello: capture()
}
const result = match(schema, input);
//result is Match { value: { hello: "world" } }
Capture the value of a node and assign a name to it.
const input = {
hello: "world"
}
const schema = {
hello: capture("prop1")
}
const result = match(schema, input);
//result is Match { value: { prop1: "world" } }
Capture the value of a node and assign a name to it.
const input = {
hello: "world"
}
const schema = {
hello: modify(
x => x === "world",
x => `${x}!!!`
)
}
const result = match(schema, input);
//result is Match { value: { prop1: "world!!!" } }
const input = {
hello: "world"
}
const schema = {
something: "else",
hello: capture()
}
const result = match(schema, input);
//result is Skip { message: "Expected something to be defined." }
const input = {
hello: "world"
}
const schema = {
hello: captureIf(x => x === "world")
}
const result = match(schema, input);
//result is Match { value: { hello: "world" } }
const input = {
name: "JPK",
isHuman: true,
age: 36,
meta: { x: 100 },
getCoordinates: () => "BLR"
}
const schema = {
name: string(),
isHuman: bool(),
age: number(),
meta: object(),
getCoordinates: func()
}
const input = {
name: "JPK",
}
const schema = {
name: literal("JPK"),
}
const input = {
myArray: [1, 2, 3]
}
const schema = { myArray: [number(), number(), number()] };
const input = {
level1: [
"one",
"two",
"three"
]
}
const schema = {
level1: [
repeating(string())
]
};
const input = {
level1: [
"one",
"two",
true
]
}
const schema = {
level1: [
unordered(string()),
unordered(bool())
]
};
const input = {
level1: [
20,
"HELLO",
true,
100
]
}
const schema = {
level1: [
optional(number()),
string(),
bool()
]
};
Matches if found. Does not emit a Skip() if not found.
const input = {
level1: {
prop1: "hello",
}
}
const schema = {
level1: {
prop1: capture(),
prop2: optional(capture())
}
};
const input = {
inner1: {
hello: "world"
}
}
const schema = {
prop1: "HELLO"
inner1: {
hello: capture()
}
}
const result = match(schema, input);
//result is Match { value: { inner1: { hello: "world" } } }
const input = {
inner1: {
hello: "world"
}
}
const schema = types.obj(
{
inner1: {
hello: capture()
}
},
{ replace: true }
)
const result = match(schema, input);
//result is Match { value: { hello: "world" } }
const input = {
level1: {
level2: "hello world"
}
}
const schema = {
level1: captureAndParse(
{
level2: capture("prop2")
},
"prop1"
)
}
const result = match(schema, input);
//result is Match { prop1: { level2: 'hello world', prop2: 'hello world' } }
const input = {
level1: {
level2: "world"
}
}
schema1 = {
level4: capture("hello")
}
schema2 = {
level1: {
level2: capture("hello")
}
};
const schema = any([schema1, schema2]);
const input = {
level1: {
prop1: "hello",
level2: {
level3: {
prop2: "something"
}
},
level2a: {
level3a: {
level4a: {
level5a: {
prop3: "world"
}
},
level4b: "yoyo"
}
}
}
}
const schema = {
level1: {
level2a: deep(
{
level5a: {
prop3: capture()
}
},
"prop1"
)
}
}
const result = match(schema, input);
//result is Match { prop1: { prop3: 'world' } }
const input = {
prop1: "hello",
prop2: undefined
}
const schema = {
prop1: capture(),
prop2: empty()
}
const result = match(schema, input);
//result is Match({ prop1: "hello" })
export const input = {
hello: "world",
prop1: "val1"
}
export const schema = {
hello: exists(),
prop1: capture()
}
//This makes sure that hello exists. Or it returns a Skip.
const input = {
hello: "world"
}
const schema = {
hello: regex(/^world$/)
}
Let's you apply multiple traversal strategies on a single tree. Traversal options are specified via a selector. { selector: }
In the following example the property "prop" is traversed without any modifiers, since the selector is set to "alt".
If ownParams are set, the consolidated result of traversals can be modified again.
const input = {
node: {
something: "else",
jeff: "buckley",
hello: "world",
},
prop: "something"
}
const schema = composite(
{
something: "else",
hello: capture({ key: "first" }),
prop: capture({ key: "second", selector: "alt" })
},
[
{ name: "default", modifiers: { object: x => x.node } },
{ name: "alt" },
]
)
Property Modifier. Use this if your input tree isn't a simple object.
const input = {
getItem(item) {
return item === "hello" ? "world" : "nothing";
}
}
const schema = types.obj(
{
hello: capture()
},
{ modifiers: { property: (obj, key) => obj.getItem(key) } }
)
The build option lets you modify the result of a parse.
import { builtins as $, capture, Match } from "../../../index.js";
export const input = {
hello: "world"
};
export const schema = $.obj(
{
hello: capture()
},
{
build: obj => context => result =>
result instanceof Match ? new Match({ hello: `${result.value.hello}!!!` }) : result
}
);
//The result is Match({ hello: "world!!!" })