Horizontal42/summarum
1
0
Fork 0
mirror of https://github.com/Horizontal42/summarum.git synced 2026-11-10 23:51:59 +03:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge 7c9c9e1b28 into 37cfd2144b

Browse source
This commit is contained in:
Andrew Pimenov 2026-09-04 10:12:32 +03:00 • committed by GitHub
commit d7e2db4d57
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 50 additions and 46 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

96
src/engine/tokenizer.ts
View file

@ -1,7 +1,7 @@
// Lexes a line and resolves phrases through the registry into semantic tokens. // Lexes a line and resolves phrases through the registry into semantic tokens.
import { DateOrder, Decimal, NumeralRepr, Unit } from "./types"; import { DateOrder, Decimal, NumeralRepr, Unit } from "./types";
import { lexLine } from "./lexer"; import { Lex, lexLine } from "./lexer";
import { Registry, PctOp, DateWord, BitOp } from "./registry"; import { Registry, PctOp, DateWord, BitOp, Payload } from "./registry";
export type Token = export type Token =
| { t: "num"; v: Decimal; repr: NumeralRepr; start: number; end: number } | { t: "num"; v: Decimal; repr: NumeralRepr; start: number; end: number }
@ -60,55 +60,15 @@ export function tokenize(line: string, reg: Registry, dateOrder: DateOrder = "dm
const m = reg.match(lexes, lowers, i); const m = reg.match(lexes, lowers, i);
if (m) { if (m) {
const end = lexes[i + m.length - 1].end; const end = lexes[i + m.length - 1].end;
const s = { start: lx.start, end }; tokens.push(handleRegistryMatch(m.payload, { start: lx.start, end }));
const p = m.payload;
switch (p.t) {
case "unit": tokens.push({ t: "unit", unit: p.unit, ...s }); break;
case "currency": tokens.push({ t: "currency", code: p.code, ...s }); break;
case "op": tokens.push({ t: "op", op: p.op, ...s }); break;
case "bitop": tokens.push({ t: "bitop", op: p.op, ...s }); break;
case "special": tokens.push({ t: "special", name: p.name, ...s }); break;
case "conv": tokens.push({ t: "conv", ...s }); break;
case "assign": tokens.push({ t: "assign", ...s }); break;
case "pctop": tokens.push({ t: "pctop", op: p.op, ...s }); break;
case "percent": tokens.push({ t: "percent", ...s }); break;
case "func": tokens.push({ t: "func", name: p.name, ...s }); break;
case "agg": tokens.push({ t: "agg", name: p.name, ...s }); break;
case "scale": tokens.push({ t: "scale", mult: p.mult, ...s }); break;
case "repr": tokens.push({ t: "repr", repr: p.repr, ...s }); break;
case "date": tokens.push({ t: "date", word: p.word, ...s }); break;
case "const": tokens.push({ t: "const", name: p.name, ...s }); break;
}
i += m.length; i += m.length;
continue; continue;
} }
if (lx.type === "sym") { if (lx.type === "sym") {
// << and >> arrive as two single-char lexemes const [tk, skip] = handleSymbol(lx, lexes, i, span);
const nextLx = lexes[i + 1]; tokens.push(tk);
if ((lx.raw === "<" || lx.raw === ">") && nextLx?.type === "sym" && nextLx.raw === lx.raw && nextLx.start === lx.end) { i += skip;
tokens.push({ t: "bitop", op: lx.raw === "<" ? "shl" : "shr", start: lx.start, end: nextLx.end });
i += 2;
continue;
}
switch (lx.raw) {
case "&": tokens.push({ t: "bitop", op: "band", ...span }); break;
case "|": tokens.push({ t: "bitop", op: "bor", ...span }); break;
case "+": tokens.push({ t: "op", op: "plus", ...span }); break;
case "-": case "−": case "–": tokens.push({ t: "op", op: "minus", ...span }); break;
case "*": case "×": case "·": tokens.push({ t: "op", op: "mul", ...span }); break;
case "/": case "÷": tokens.push({ t: "op", op: "div", ...span }); break;
case "^": tokens.push({ t: "op", op: "pow", ...span }); break;
case "%": tokens.push({ t: "percent", ...span }); break;
case "(": tokens.push({ t: "lparen", ...span }); break;
case ")": tokens.push({ t: "rparen", ...span }); break;
case "=": tokens.push({ t: "assign", ...span }); break;
case ";": tokens.push({ t: "semicolon", ...span }); break;
case "!": tokens.push({ t: "bang", ...span }); break;
case "?": tokens.push({ t: "unknown", ...span }); break;
default: tokens.push({ t: "junk", raw: lx.raw, ...span });
}
i++;
continue; continue;
} }
@ -118,6 +78,50 @@ export function tokenize(line: string, reg: Registry, dateOrder: DateOrder = "dm
return disambiguateMinAgg(disambiguateIn(tokens, reg, line), reg, line); return disambiguateMinAgg(disambiguateIn(tokens, reg, line), reg, line);
} }
function handleRegistryMatch(p: Payload, s: { start: number; end: number }): Token {
switch (p.t) {
case "unit": return { t: "unit", unit: p.unit, ...s };
case "currency": return { t: "currency", code: p.code, ...s };
case "op": return { t: "op", op: p.op, ...s };
case "bitop": return { t: "bitop", op: p.op, ...s };
case "special": return { t: "special", name: p.name, ...s };
case "conv": return { t: "conv", ...s };
case "assign": return { t: "assign", ...s };
case "pctop": return { t: "pctop", op: p.op, ...s };
case "percent": return { t: "percent", ...s };
case "func": return { t: "func", name: p.name, ...s };
case "agg": return { t: "agg", name: p.name, ...s };
case "scale": return { t: "scale", mult: p.mult, ...s };
case "repr": return { t: "repr", repr: p.repr, ...s };
case "date": return { t: "date", word: p.word, ...s };
case "const": return { t: "const", name: p.name, ...s };
}
}
function handleSymbol(lx: Lex, lexes: Lex[], i: number, span: { start: number; end: number }): [Token, number] {
const nextLx = lexes[i + 1];
if ((lx.raw === "<" || lx.raw === ">") && nextLx?.type === "sym" && nextLx.raw === lx.raw && nextLx.start === lx.end) {
return [{ t: "bitop", op: lx.raw === "<" ? "shl" : "shr", start: lx.start, end: nextLx.end }, 2];
}
switch (lx.raw) {
case "&": return [{ t: "bitop", op: "band", ...span }, 1];
case "|": return [{ t: "bitop", op: "bor", ...span }, 1];
case "+": return [{ t: "op", op: "plus", ...span }, 1];
case "-": case "−": case "–": return [{ t: "op", op: "minus", ...span }, 1];
case "*": case "×": case "·": return [{ t: "op", op: "mul", ...span }, 1];
case "/": case "÷": return [{ t: "op", op: "div", ...span }, 1];
case "^": return [{ t: "op", op: "pow", ...span }, 1];
case "%": return [{ t: "percent", ...span }, 1];
case "(": return [{ t: "lparen", ...span }, 1];
case ")": return [{ t: "rparen", ...span }, 1];
case "=": return [{ t: "assign", ...span }, 1];
case ";": return [{ t: "semicolon", ...span }, 1];
case "!": return [{ t: "bang", ...span }, 1];
case "?": return [{ t: "unknown", ...span }, 1];
default: return [{ t: "junk", raw: lx.raw, ...span }, 1];
}
}
/** /**
* "5 ft 4 in in cm" / "6 ft 3 in": the word "in" right after a number, * "5 ft 4 in in cm" / "6 ft 3 in": the word "in" right after a number,
* followed by another conversion word, an operator or the end of the line, * followed by another conversion word, an operator or the end of the line,