阅读中文版本

Build a Graph Traversal Service

Graph Traversal Service — a complete walkthrough of graph traversal, covering adjacency store, bfs engine, dijkstra engine, and testing.

graph_traversal_service/
├── src/index.ts
├── src/adjacency-store.ts
├── src/bfs-engine.ts
├── tests/graph_traversal_service.test.ts
└── package.json

// Adjacency store: graph traversal core
export function adjacencystore(input: AdjacencystoreInput): AdjacencystoreResult {
  // Build the adjacency store that anchors graph traversal and defines the core data shape.
  return processAdjacencystore(input);
}

// BFS engine — handles graph traversal Add the bfs engine so
export class Bfsengine {
  constructor(private opts: BfsengineOptions) {}
  async run(ctx: Context): Promise<BfsengineOutput> {
    return this.execute(ctx);
  }
}

// Dijkstra engine: graph traversal stage
export const dijkstraengine = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

// Path materializer for graph traversal
export async function pathmaterializer(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

// Graph snapshot: graph traversal finalization
export function graphsnapshot(state: State): Result {
  validate(state);
  return finalize(state);
}

Project overview

This project builds a graph traversal system whose reason for existing is answering shortest-path and reachability queries over a graph that changes between queries. The temptation in graph traversal is to bolt the interesting part onto a working prototype and discover the hard cases in production; this build inverts that order and treats a query that hits a disconnected component and walks the entire reachable set before returning empty as the first thing designed, not the last.

You will implement Adjacency store, BFS engine, Dijkstra engine, Path materializer, and Graph snapshot, each as a module with one responsibility and a test that proves it in isolation. The graph traversal pipeline is then the wiring that composes them, and the integration test is where the composition is shown to survive a query that hits a disconnected component and walks the entire reachable set before returning empty.

The walkthrough is written for a developer who has used a graph traversal system but has not had to make one honest. It favors explicit data shapes over clever abstractions, and it names the failure each design choice prevents rather than presenting the choice as a style preference.

Each phase ends with something observable — a adjacency store that returns the right shape, a bfs engine that rejects bad input, an integration test that goes green — so you can stop at any point and still have a graph traversal system that works for the cases covered so far.

What you will build

The finished graph traversal system exposes five capabilities, each with a focused test:

• Adjacency store
• BFS engine
• Dijkstra engine
• Path materializer
• Graph snapshot

Adjacency store is the foundation that every later stage assumes; BFS engine is where graph traversal correctness lives; Dijkstra engine handles a query that hits a disconnected component and walks the entire reachable set before returning empty; Path materializer makes the system observable; and Graph snapshot is the graceful-degradation layer that keeps overload from becoming corruption.

The build keeps the graph traversal logic in ordinary modules you can read in one sitting, and pulls in dependencies only for the parts that are genuinely tedious to hand-roll. That keeps the diff between the tutorial and a real codebase small, and keeps the failure modes visible in code rather than hidden behind a framework.

Architecture and data model

The system is organized around Adjacency store, BFS engine, and Dijkstra engine, with Path materializer tying them together. Data flows as an adjacency list keyed by node id with weighted edges, backed by a read-mostly copy swapped atomically on updates.

Each component is a module with a pure interface where practical: the same input produces the same output, and side effects are pushed to the edges. That discipline is what makes the phases independently testable, because a test for Dijkstra engine does not need to stand up Adjacency store and BFS engine to exercise the interesting branch.

The data model is deliberately explicit. An adjacency list keyed by node id with weighted edges, backed by a read-mostly copy swapped atomically on updates Keeping the shape written down — in a schema, a type, or a fixture — means a change to the shape is a visible diff rather than a silent runtime shift, and it means the integration test can assert against a concrete shape instead of guessing.

Configuration and code are kept apart: the configuration says what graph traversal should do, the modules say how. A reviewer can change a threshold or a policy without reading the implementation, which keeps the risky changes small and lets the diff stand as the review.

The failure path through Dijkstra engine is designed before the happy path is finished, because a query that hits a disconnected component and walks the entire reachable set before returning empty is the case that distinguishes a graph traversal system from a demo. Every component declares what it does when its inputs are wrong, and the integration test asserts each of those declarations.

Prerequisites and dependencies

Have these ready before you install anything, so the graph traversal setup is not blocked by a missing tool:

• Node.js 20 LTS or later
• Familiarity with graph traversal fundamentals
• A local environment to run the test suite

The runtime dependencies are chosen to keep the graph traversal datapath legible:

• fastify (^5.0.0) — Query API server
• redis (^4.7.0) — Edge cache for hot subgraphs

# setup for graph-traversal-service
pnpm init
pnpm add fastify redis
pnpm add -D typescript tsx vitest

Versions are pinned to a major range so the graph traversal install is reproducible; commit the lockfile. The adjacency store reads its operational parameters from the environment and validates them at startup, so a missing value fails the boot rather than the first graph traversal request.

Directory structure

Keep each graph traversal component in its own file so it can be tested without starting the others:

graph_traversal_service/
├── src/index.ts
├── src/adjacency-store.ts
├── src/bfs-engine.ts
├── tests/graph_traversal_service.test.ts
└── package.json

Phase 1: Adjacency store

Build the adjacency store that anchors graph traversal and defines the core data shape.

// Adjacency store: graph traversal core
export function adjacencystore(input: AdjacencystoreInput): AdjacencystoreResult {
  // Build the adjacency store that anchors graph traversal and defines the core data shape.
  return processAdjacencystore(input);
}

After this phase the adjacency store works in isolation and is covered by a unit test. The adjacency store defines the data shape that every later stage assumes, so getting it right first means the later phases compile against a stable contract.

Treat the adjacency store as the single source of truth for the graph traversal data shape. Validate inputs at the boundary and reject early, so a malformed value never reaches BFS engine or Dijkstra engine where its effect would be harder to localize.

Phase 2: BFS engine

Add the bfs engine so graph traversal can answering shortest-path and reachability correctly.

// BFS engine — handles graph traversal Add the bfs engine so
export class Bfsengine {
  constructor(private opts: BfsengineOptions) {}
  async run(ctx: Context): Promise<BfsengineOutput> {
    return this.execute(ctx);
  }
}

After this phase the bfs engine works in isolation and is covered by a unit test. This is where graph traversal correctness lives: the bfs engine is what makes answering shortest-path and reachability queries over a graph that changes between queries true, and its test is the proof.

Keep the bfs engine pure where you can. A function that takes an input and returns an output is trivial to test and impossible to mock wrong; a class with hidden state is the opposite. Push side effects to the caller and keep the bfs engine about the transformation.

Phase 3: Dijkstra engine

Wire the dijkstra engine to handle a query that hits without losing data.

// Dijkstra engine: graph traversal stage
export const dijkstraengine = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

After this phase the dijkstra engine works in isolation and is covered by a unit test. The dijkstra engine is the difference between a graph traversal system that works on the happy path and one that survives a query that hits a disconnected component and walks the entire reachable set before returning empty.

Design the failure before the success. Decide what dijkstra engine does when a query that hits a disconnected component and walks the entire reachable set before returning empty: reject, retry, shed, or degrade — and write that test first. A graph traversal system whose failure behavior is unspecified will pick one at random under pressure, and it is usually the wrong one.

Phase 4: Path materializer

Integrate the path materializer to make graph traversal observable and testable end to end.

// Path materializer for graph traversal
export async function pathmaterializer(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

After this phase the path materializer works in isolation and is covered by a unit test. Observability is not a finishing touch; it is how you know the graph traversal system is doing what it claims in production.

Emit one structured event per meaningful action, with enough context to localize a failure without a second round-trip. The path materializer is what turns “something is slow” into “the bfs engine is slow for inputs matching this shape,” which is the difference between an alert and an investigation.

Phase 5: Graph snapshot

Finalize the graph snapshot so the graph traversal system degrades gracefully under overload.

// Graph snapshot: graph traversal finalization
export function graphsnapshot(state: State): Result {
  validate(state);
  return finalize(state);
}

After this phase the graph snapshot works in isolation and is covered by a unit test. The graph snapshot is what keeps overload from becoming corruption: when the graph traversal system is asked for more than it can do, it degrades in a defined way instead of failing in an undefined one.

Pick a degradation strategy for graph snapshot — shed, queue, or backpressure — and make it explicit in code and in a test. Leaving the behavior to chance is how a graph traversal traffic spike becomes a data loss incident.

Testing

Each phase has a focused test: adjacency store, bfs engine, and dijkstra engine are exercised in isolation, then an integration test drives the full graph traversal pipeline and asserts the a query that hits case is handled.

Start with unit tests that pin each component’s contract: the adjacency store returns the documented shape, the bfs engine transforms correctly, and the dijkstra engine fails in the documented way. These tests are fast and they localize regressions to the component that broke.

Then add an integration test that composes the components and drives a realistic graph traversal scenario end to end. Assert the happy path first, then deliberately trigger a query that hits a disconnected component and walks the entire reachable set before returning empty and assert the system degrades rather than corrupts. Because the components are pure where practical, the integration test stays deterministic rather than flaky.

A graph traversal suite must also prove the negative: bad input is rejected at the boundary, overload is shed rather than queued indefinitely, and a partial failure does not leave the system in an inconsistent state. These are the tests that catch the regressions a happy-path suite never would.

The entire graph traversal suite runs with no external service; every dependency is either in-process or stubbed, so every case is reproducible in CI. That is what lets the dijkstra engine failure-mode assertions live as tests rather than as a manual checklist that drifts out of date.

# run the graph traversal test suite
pnpm vitest run

Deployment and operations

Run at least two graph traversal instances behind a load balancer and share state in Redis. Set resource limits so graph traversal overload surfaces as backpressure rather than failure.

Run at least two stateless instances behind a load balancer, and move any shared state into a shared store so the graph traversal behavior is global rather than per instance. Size limits so the system sheds load before it runs out of memory, because a graph traversal system that OOMs under load fails in a way that is hard to reason about.

Drain on deploy: stop accepting new work, let in-flight graph traversal operations finish, then exit. Because the system holds no state that cannot be reconstructed, a rollout is a matter of cycling instances one at a time while watching the error rate.

When a dependency is unavailable, decide explicitly whether the graph traversal system fails closed or degrades. Leaving that choice to chance is how a dependency outage becomes a silent data loss, so make the choice visible in code and in a test.

Security considerations

Validate all graph traversal inputs at the boundary, reject oversized payloads before buffering them, and keep secrets in the environment rather than the repository.

The graph traversal system is a choke point, so it is the right place to reject malformed input before it reaches bfs engine or dijkstra engine. Validate the shape and the size at the edge, and let the inner components assume what they receive is well-formed.

Never log secrets or full payloads; log enough to localize a failure without recording data that should not be on disk. Treat any external call as untrusted and validate its response before using it, because a graph traversal system that trusts its inputs blindly is one malformed response away from corruption.

Keep dependencies pinned and run an audit in CI. A transitive dependency with a known vulnerability is still a vulnerability in your graph traversal system, and the lockfile is what makes the audit meaningful.

Observability

Emit one structured event per graph traversal operation with the inputs’ shape, the outcome, and a correlation id. Track error rate and latency per component and alert on a step change rather than a single spike.

The correlation id is the thread that ties a user-visible symptom to the graph traversal decision and to the component log entry for the same operation. Without it, an investigation starts from a timestamp and a guess.

Watch the dijkstra engine failure rate separately from the overall error rate. A jump in dijkstra engine failures usually means a query that hits a disconnected component and walks the entire reachable set before returning empty, not a bug in bfs engine, and the two have different owners and different fixes.

Chart the share of graph traversal operations that shed, queue, or succeed. A change in the mix points at the cause faster than an aggregate latency can, because each class has a different root.

Troubleshooting

A query that hits a disconnected component and walks the entire reachable set before returning empty. Reproduce with a minimal fixture, then add a guard test so the graph traversal system rejects the condition loudly.

Performance regresses under graph traversal load. Profile with the sampling analyzer, locate the hot path in bfs engine, and add a targeted benchmark to lock the regression.

Keep the graph traversal dashboards separate from dependency dashboards. When an operation fails, the correlation id lets you decide in seconds whether the fault is adjacency store, bfs engine, or dijkstra engine, instead of guessing from a timeout.

Limitations and next steps

• Single-region; multi-region replication is out of scope.
• The adjacency store is in-process and not horizontally shared by default.
• No auth layer; assume deployment behind an edge that enforces identity.

The graph traversal system is single-region by design; extending to multi-region means making adjacency store shareable across regions, which is a real change to the consistency model rather than a config flag. The dijkstra engine is in-process, which is correct for one instance but must move to a shared store before scaling out.

There is no authentication or authorization layer; the system assumes it is deployed behind an edge that enforces identity. Adding one is straightforward but out of scope for this build, and doing it here would distract from the graph traversal correctness that is the point.

The natural next step, once the graph traversal pipeline is trusted in production, is to add a short-TTL cache for idempotent reads and a richer path materializer that exports histograms rather than counters. Both are additive and neither changes the contracts this build establishes.