speckl/compiler/dist/parser.js
Scott Scoble 8facd5c6c7
Some checks failed
Deploy Landing Page to Storj / deploy (push) Failing after 3s
Wire up compiler generators + fix parser for SpeckDL edge cases
- Connected PROV-O, CycloneDX, and SPDX generators to CLI entry point
- Fixed parseAuthor regex to handle angle-bracket email format
- Fixed parseDerives to handle hyphenated names and quoted rationale
- Fixed parseSatisfies to handle hyphenated requirement IDs
- Fixed parseSource to handle ref keyword with quoted value
- Added graceful fallbacks (warn instead of throw) for parse errors
- All 5 example .speck files compile and validate successfully
2026-05-02 15:14:31 -07:00

300 lines
No EOL
9 KiB
JavaScript

import fs from 'fs';
/**
* Parse a .speck file and return the AST
*/
export function parseSpeckFile(filePath) {
const content = fs.readFileSync(filePath, 'utf-8');
return parseSpeckContent(content, filePath);
}
/**
* Parse SpeckDL content and return the AST
*/
export function parseSpeckContent(content, filePath) {
const specks = [];
const lines = content.split('\n');
let i = 0;
while (i < lines.length) {
const line = lines[i].trim();
// Skip empty lines and comments
if (!line || line.startsWith('//') || line.startsWith('/*')) {
i++;
continue;
}
// Handle import statements
if (line.startsWith('import ')) {
const importNode = parseImportStatement(line);
// Import is a special case - it's not part of a speck
// We'll handle it separately in a future update
i++;
continue;
}
// Handle speck definitions
if (line.startsWith('speck ')) {
const speck = parseSpeck(lines, i);
if (speck) {
specks.push(speck);
// Move index past the speck we just parsed
i++; // Increment by 1, the loop will continue from next line
}
else {
i++;
}
continue;
}
i++;
}
return { specks };
}
function parseSpeck(lines, startIndex) {
// Parse speck header: speck <Name> {
const headerMatch = lines[startIndex].match(/^speck\s+(\w+)\s*{/);
if (!headerMatch) {
return null;
}
const name = headerMatch[1];
const members = [];
let i = startIndex + 1;
let braceCount = 1;
while (i < lines.length && braceCount > 0) {
const line = lines[i].trim();
// Skip empty lines and comments
if (!line || line.startsWith('//') || line.startsWith('/*')) {
i++;
continue;
}
if (line.includes('{')) {
braceCount++;
}
if (line.includes('}')) {
braceCount--;
if (braceCount === 0) {
break;
}
}
// Parse member
const member = parseMember(line);
if (member) {
members.push(member);
}
i++;
}
return {
type: 'speck',
name,
members,
};
}
function parseMember(line) {
// input:
if (line.startsWith('input:')) {
const typeExpr = parseTypeExpr(line.substring(6).trim());
return { type: 'input', typeExpr };
}
// output:
if (line.startsWith('output:')) {
const typeExpr = parseTypeExpr(line.substring(7).trim());
return { type: 'output', typeExpr };
}
// constraint:
if (line.startsWith('constraint:')) {
return { type: 'constraint', expr: line.substring(11).trim() };
}
// verify:
if (line.startsWith('verify:')) {
return { type: 'verify', temporalExpr: line.substring(7).trim() };
}
// interface:
if (line.startsWith('interface ')) {
return parseInterface(line);
}
// event:
if (line.startsWith('event ')) {
return parseEvent(line);
}
// provenance {
// review:
if (line.startsWith('review:')) {
const kind = line.substring(7).trim();
return { type: 'review', kind };
}
// derives from:
if (line.startsWith('derives from ')) {
return parseDerives(line);
}
// satisfies:
if (line.startsWith('satisfies ')) {
return parseSatisfies(line);
}
// author:
if (line.startsWith('author:')) {
return parseAuthor(line);
}
// source:
if (line.startsWith('source:')) {
return parseSource(line);
}
// bom {
if (line.startsWith('bom ')) {
// Simplified - full BOM block parsing would need multi-line context
return {
type: 'bom',
};
}
return null;
}
function parseImportStatement(line) {
const match = line.match(/^import\s+(['"])(.+?)\1(?:\s+as\s+(\w+))?(?:\s+version\s+(['"])(.+?)\4)?(?:\s+hash\s+(['"])(.+?)\6)?/);
if (!match) {
throw new Error(`Failed to parse import statement: ${line}`);
}
const [, , path, , alias, , version, , hash] = match;
return {
type: 'import',
path,
alias,
version: version || undefined,
hash: hash || undefined,
};
}
function parseTypeExpr(expr) {
expr = expr.trim();
// Primitive types
const primitives = ['Nat', 'Int', 'Real', 'Bool', 'String', 'Bytes'];
if (primitives.includes(expr)) {
return { type: 'primitive', name: expr };
}
// List type: [T]
if (expr.startsWith('[') && expr.endsWith(']')) {
const elementType = parseTypeExpr(expr.substring(1, expr.length - 1));
return { type: 'list', elementType };
}
// Record type: { ... }
if (expr.startsWith('{') && expr.endsWith('}')) {
const content = expr.substring(1, expr.length - 1).trim();
const fields = [];
if (content) {
// Split by comma and parse each field
const fieldStrs = content.split(',').map(s => s.trim());
for (const fieldStr of fieldStrs) {
const fieldMatch = fieldStr.match(/^(\w+)\s*:\s*(.+)$/);
if (fieldMatch) {
const [, name, typeStr] = fieldMatch;
fields.push({ name, type: parseTypeExpr(typeStr.trim()) });
}
}
}
return { type: 'record', fields };
}
// Identifier (for interface references, etc.)
return { type: 'ident', name: expr };
}
function parseInterface(line) {
// interface Name {
const match = line.match(/^interface\s+(\w+)\s*{/);
if (!match) {
throw new Error(`Failed to parse interface: ${line}`);
}
const name = match[1];
// This is a simplified parser - in a real implementation, we'd need to handle multi-line interfaces
return {
type: 'interface',
name,
methods: [],
};
}
function parseEvent(line) {
// event Name {
const match = line.match(/^event\s+(\w+)\s*{/);
if (!match) {
throw new Error(`Failed to parse event: ${line}`);
}
const name = match[1];
return {
type: 'event',
name,
fields: [],
};
}
function parseProvenanceBlock(lines, line) {
// This is a simplified parser - in a real implementation, we'd parse the full block
return {
type: 'provenance',
clauses: [],
};
}
function parseDerives(line) {
// derives from Name via "rationale"
const match = line.match(/^derives\s+from\s+(\w+(?:-\w+)*)(?:\s+via\s+(['"])(.+?)\2)?/);
if (!match) {
console.warn(`Failed to parse derives statement: ${line}`);
return { type: 'derives', from: 'unknown' };
}
const [, from, , via] = match;
return {
type: 'derives',
from,
via: via || undefined,
};
}
function parseSatisfies(line) {
// satisfies Name [clause "X.Y.Z"]
// e.g., satisfies RESILIENCE-REQ-01 clause "3.1.2"
// e.g., satisfies RESILIENCE-REQ-02
const match = line.match(/^satisfies\s+([\w-]+(?:-[\w]+)*)(?:\s+clause\s+(['"])(.+?)\2)?/);
if (!match) {
console.warn(`Failed to parse satisfies statement: ${line}`);
return { type: 'satisfies', requirement: 'unknown' };
}
const [, requirement, , clause] = match;
return {
type: 'satisfies',
requirement,
clause: clause || undefined,
};
}
function parseAuthor(line) {
// author: "Name" <"email">
// e.g., author: "Scott Scoble" <"scott@scoble.me">
const match = line.match(/^author:\s+(['"])(.+?)\1(?:\s+<(['"])(.+?)\3>)?/);
if (!match) {
// Try without angle brackets or with different format
console.warn(`Failed to parse author statement: ${line}`);
return { type: 'author', name: 'unknown' };
}
const [, , name, , email] = match;
return {
type: 'author',
name,
email: email || undefined,
};
}
function parseSource(line) {
// source: kind ref "identifier"
// e.g., source: architecture_review ref "AR-2024-0215"
const match = line.match(/^source:\s+(\w+)(?:\s+ref\s+(['"])(.+?)\2)?/);
if (!match) {
console.warn(`Failed to parse source statement: ${line}`);
return { type: 'source', kind: 'document' };
}
const [, kind, , ref] = match;
const validKinds = [
'conversation', 'meeting', 'document', 'regulation',
'architecture_review', 'threat_model', 'compliance_audit'
];
if (!validKinds.includes(kind)) {
console.warn(`Invalid source kind: ${kind}, defaulting to document`);
return { type: 'source', kind: 'document', ref: ref || undefined };
}
return {
type: 'source',
kind: kind,
ref: ref || undefined,
};
}
function parseBOMBlock(lines, line) {
// This is a simplified parser - in a real implementation, we'd parse the full block
return {
type: 'bom',
};
}
//# sourceMappingURL=parser.js.map