@chr33s/pdf-dfa 5.0.0

package/dist/utils.d.ts

@@ -0,0 +1,12 @@
+/**
+ * Returns a new set representing the union of a and b.
+ */
+export declare function union<T>(a: Set<T>, b: Iterable<T>): Set<T>;
+/**
+ * Adds all items from the set b to a.
+ */
+export declare function addAll<T>(target: Set<T>, source: Iterable<T>): void;
+/**
+ * Returns whether two sets are equal
+ */
+export declare function equal<T>(a: Set<T>, b: Set<T>): boolean;

package/dist/utils.js

@@ -0,0 +1,34 @@
+/**
+ * Returns a new set representing the union of a and b.
+ */
+export function union(a, b) {
+    const result = new Set(a);
+    addAll(result, b);
+    return result;
+}
+/**
+ * Adds all items from the set b to a.
+ */
+export function addAll(target, source) {
+    for (const item of source) {
+        target.add(item);
+    }
+}
+/**
+ * Returns whether two sets are equal
+ */
+export function equal(a, b) {
+    if (a === b) {
+        return true;
+    }
+    if (a.size !== b.size) {
+        return false;
+    }
+    for (const item of a) {
+        if (!b.has(item)) {
+            return false;
+        }
+    }
+    return true;
+}
+//# sourceMappingURL=utils.js.map

package/dist/utils.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"utils.js","sourceRoot":"","sources":["../src/utils.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,MAAM,UAAU,KAAK,CAAI,CAAS,EAAE,CAAc;IAChD,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC;IAC1B,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;IAClB,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,MAAM,CAAI,MAAc,EAAE,MAAmB;IAC3D,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE,CAAC;QAC1B,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACnB,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,KAAK,CAAI,CAAS,EAAE,CAAS;IAC3C,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACZ,OAAO,IAAI,CAAC;IACd,CAAC;IAED,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;QACtB,OAAO,KAAK,CAAC;IACf,CAAC;IAED,KAAK,MAAM,IAAI,IAAI,CAAC,EAAE,CAAC;QACrB,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;YACjB,OAAO,KAAK,CAAC;QACf,CAAC;IACH,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC"}

package/package.json

@@ -0,0 +1,41 @@
+{
+  "name": "@chr33s/pdf-dfa",
+  "version": "5.0.0",
+  "description": "Deterministic finite automata compiler",
+  "type": "module",
+  "sideEffects": false,
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "default": "./dist/index.js"
+    }
+  },
+  "scripts": {
+    "build": "npm run clean && npm run generate && tsc",
+    "clean": "rm -rf dist src/grammar.js",
+    "generate": "node scripts/build-grammar.ts",
+    "test": "vitest run",
+    "typecheck": "tsc --project tsconfig.typecheck.json"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/chr33s/pdf.git"
+  },
+  "keywords": [
+    "state",
+    "machine",
+    "compiler"
+  ],
+  "author": "Devon Govett <devongovett@gmail.com>",
+  "bugs": {
+    "url": "https://github.com/chr33s/pdf/issues"
+  },
+  "license": "MIT",
+  "homepage": "https://github.com/chr33s/pdf",
+  "devDependencies": {
+    "@types/node": "24.10.1",
+    "peggy": "5.0.6",
+    "typescript": "5.9.3",
+    "vitest": "4.0.15"
+  }
+}

package/scripts/build-grammar.ts

@@ -0,0 +1,33 @@
+import { readFile, writeFile } from "node:fs/promises";
+import path from "node:path";
+import { fileURLToPath } from "node:url";
+import peggy from "peggy";
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = path.dirname(__filename);
+const projectRoot = path.resolve(__dirname, "..");
+const srcDir = path.join(projectRoot, "src");
+const grammarPath = path.join(srcDir, "grammar.peg");
+const outputPath = path.join(srcDir, "grammar.js");
+
+async function buildGrammar(): Promise<void> {
+  const grammar = await readFile(grammarPath, "utf8");
+
+  const parserSource = peggy.generate(grammar, {
+    cache: true,
+    output: "source",
+    format: "es",
+  });
+
+  const banner = [
+    "// @ts-nocheck",
+    "// This file is generated by scripts/build-grammar.ts. Do not edit by hand.\n",
+  ].join("\n");
+  await writeFile(outputPath, `${banner}${parserSource}`);
+  console.log(`Generated ${path.relative(projectRoot, outputPath)}`);
+}
+
+buildGrammar().catch((error) => {
+  console.error("Failed to build grammar:", error);
+  process.exitCode = 1;
+});

package/src/compile.ts

@@ -0,0 +1,31 @@
+import buildDFA, { DFAState } from "./dfa.js";
+import { parse as parseGrammar } from "./grammar.js";
+import * as nodes from "./nodes.js";
+import StateMachine, { StateMachineConfig } from "./state-machine.js";
+import SymbolTable from "./symbol-table.js";
+
+export type ExternalSymbols = Record<string, number>;
+
+export function parse(source: string, externalSymbols: ExternalSymbols = {}): SymbolTable {
+  const ast = parseGrammar(source, { nodes });
+  return new SymbolTable(ast, externalSymbols);
+}
+
+export function build(symbolTable: SymbolTable): StateMachine {
+  const states: DFAState[] = buildDFA(symbolTable.main, symbolTable.size);
+
+  const config: StateMachineConfig = {
+    stateTable: states.map((state) => Array.from(state.transitions)),
+    accepting: states.map((state) => state.accepting),
+    tags: states.map((state) => Array.from(state.tags)),
+  };
+
+  return new StateMachine(config);
+}
+
+export default function compile(
+  source: string,
+  externalSymbols: ExternalSymbols = {},
+): StateMachine {
+  return build(parse(source, externalSymbols));
+}

package/src/dfa.ts

@@ -0,0 +1,104 @@
+import { Concatenation, EndMarker, ExpressionNode, Literal, PositionNode, Tag } from "./nodes.js";
+import { addAll, equal } from "./utils.js";
+
+const END_MARKER = new EndMarker();
+
+export interface DFAState {
+  positions: Set<PositionNode>;
+  transitions: Uint16Array;
+  accepting: boolean;
+  marked: boolean;
+  tags: Set<string>;
+}
+
+/**
+ * This is an implementation of the direct regular expression to DFA algorithm described
+ * in section 3.9.5 of "Compilers: Principles, Techniques, and Tools" by Aho,
+ * Lam, Sethi, and Ullman. http://dragonbook.stanford.edu
+ * There is a PDF of the book here:
+ * http://www.informatik.uni-bremen.de/agbkb/lehre/ccfl/Material/ALSUdragonbook.pdf
+ */
+export default function buildDFA(root: ExpressionNode, numSymbols: number): DFAState[] {
+  const augmentedRoot = new Concatenation(root, END_MARKER);
+  augmentedRoot.calcFollowpos();
+
+  const failState = new State(new Set<PositionNode>(), numSymbols);
+  const initialState = new State(augmentedRoot.firstpos, numSymbols);
+  const dstates: State[] = [failState, initialState];
+
+  // while there is an unmarked state S in dstates
+  while (true) {
+    let s: State | null = null;
+
+    for (let j = 1; j < dstates.length; j++) {
+      if (!dstates[j].marked) {
+        s = dstates[j];
+        break;
+      }
+    }
+
+    if (s == null) {
+      break;
+    }
+
+    // mark S
+    s.marked = true;
+
+    // for each input symbol a
+    for (let a = 0; a < numSymbols; a++) {
+      // let U be the union of followpos(p) for all
+      //  p in S that correspond to a
+      const u = new Set<PositionNode>();
+      for (const p of s.positions) {
+        if (p instanceof Literal && p.value === a) {
+          addAll(u, p.followpos);
+        }
+      }
+
+      if (u.size === 0) {
+        continue;
+      }
+
+      // if U is not in dstates
+      let ux = -1;
+      for (let i = 0; i < dstates.length; i++) {
+        if (equal(u, dstates[i].positions)) {
+          ux = i;
+          break;
+        }
+      }
+
+      if (ux === -1) {
+        // Add U as an unmarked state to dstates
+        dstates.push(new State(u, numSymbols));
+        ux = dstates.length - 1;
+      }
+
+      s.transitions[a] = ux;
+    }
+  }
+
+  return dstates;
+}
+
+class State implements DFAState {
+  readonly positions: Set<PositionNode>;
+  readonly transitions: Uint16Array;
+  readonly accepting: boolean;
+  marked: boolean;
+  readonly tags: Set<string>;
+
+  constructor(positions: Set<PositionNode>, len: number) {
+    this.positions = positions;
+    this.transitions = new Uint16Array(len);
+    this.accepting = positions.has(END_MARKER);
+    this.marked = false;
+    this.tags = new Set<string>();
+
+    for (const pos of positions) {
+      if (pos instanceof Tag) {
+        this.tags.add(pos.name);
+      }
+    }
+  }
+}