Cypher-lite front-end: tokenize then parse·14 min

Parse a path pattern

0 / 1

# grade this stage (from your course directory)
npx theinference test --stage parse-path-pattern --watch

Parse a full path: a start node, then a chain of (relationship, node) segments, decoding direction from the arrow tokens and the optional *min..max. Reuse parseNodePattern for every node; loop while the next token starts a relationship (- or <-).

Your task

Implement parsePathPattern

Each segment is -[:TYPE]-> / <-[:TYPE]- / -[:TYPE]-. Decode the arrows: a right arrow (->) and no left arrow is OUT; a left arrow (<-) and no right is IN; neither is BOTH. Parse optional * var-length (*, *2, *1..3, *..3).

src/parser.ts

// PathPattern = { start: NodePattern; segments: Array<{ rel: RelPattern; node: NodePattern }> }
// RelPattern  = { type: string|null; dir: 'OUT'|'IN'|'BOTH'; varLength: {min,max}|null }
parsePathPattern(): PathPattern {
  const start = this.parseNodePattern();
  const segments = [];
  while (this.isPunct('-') || this.isPunct('<-')) {
    const rel = this.parseRelPattern();   // reads <- / - , [:TYPE *m..n] , -> / -
    const node = this.parseNodePattern();
    segments.push({ rel, node });
  }
  return { start, segments };
}

Process command

your_program.sh parse-path "(a)-[:ACTED_IN]->(m)<-[:ACTED_IN]-(b)"
# two segments: first rel dir OUT, second rel dir IN

How it's graded

Process mode, stdout_json_match against the exact two-segment AST. The co-star shape (a)-[:ACTED_IN]->(m)<-[:ACTED_IN]-(b) is precisely what the capstone executor will run.

Parse a node pattern Parse the whole query