Nomos Block Explorer

This is a Proof of Concept (PoC) for a block explorer for the Nomos blockchain.

Features

• Frontend (React-like SPA)
  • Client-side routing with Home, Block, and Transaction pages.
    • Home: Live stream of the latest Blocks and Transactions.
    • Block: Details of a Block, including a list of its transactions.
    • Transaction: Details of a Transaction.
• Backend (FastAPI)
  • API
    • REST API to query Blocks and Transactions.
    • SSE API to stream live Blocks (and its transactions).
  • Node Management
    • Pluggable API (e.g. fake, http) to query nodes.
    • Pluggable Manager (e.g. noop, docker) to manage local nodes.
  • Simple backfilling mechanism to populate historical blocks.

Architecture

The Nomos Block Explorer follows a three-tier architecture with a clear separation of concerns:

High-Level Overview

graph LR;
A[Nomos<br/>Node] -->|REST/SSE| B["Backend<br/>(FastAPI)"]
B -->|REST/SSE| C["Frontend<br/>(Preact)"]
B <--> D["Database<br/>(SQLite)"]

Components

1. Frontend (/static)

• Framework: Preact (lightweight React alternative)
• Routing: Client-side SPA routing
• Architecture: Component-based with functional components
• Communication: REST API calls and Server-Sent Events (SSE) for real-time updates

2. Backend (/src)

• Framework: FastAPI (Python async web framework)
• API Layer (/src/api): Serializers, REST and streaming endpoints
• Core (/src/core): Application setup, configuration, base types and mixins
• Database Layer (/src/db): Repository pattern for data access
• Models (/src/models): Domain models (Block, Transaction, Header, etc.)
• Node Integration (/src/node):
  • API: Pluggable adapters to communicate with Nomos nodes (fake, http) and serializers
  • Manager: Pluggable node lifecycle management (noop, docker)

3. Data Flow

1. Node Updates: On startup, the backend starts listening for new blocks from the node and stores them in the database
2. Backfilling: After at least one block is in the database, the backend fetches historical blocks from the node and stores them
3. Client Updates: Frontend subscribes to SSE endpoints for real-time block and transaction updates
4. Data Access: All queries route through repository classes for consistent data access

4. Key Design Patterns

• Repository Pattern: Abstraction layer for database operations (BlockRepository, TransactionRepository)
• Strategy Pattern: Pluggable Node API implementations (fake for testing, HTTP for production)
• Adapter Pattern: Serializers convert between Node API formats and internal domain models
• Observer Pattern: SSE streams for pushing real-time updates to clients

Requirements

• Python 3.14
• UV Package Manager

Optional

• Docker: To run a local node.

How to run

1. Install the dependencies:

   uv sync
   

2. Run the block explorer:

   PYTHONPATH=src uv run python -m main
   

By default, this will try to connect to a local Node running on port 18080.

• You can optionally run it via Docker with:

  docker build -t nomos-block-explorer . && docker run -p 8000:8000 nomos-block-explorer
  

• If you want to run the Explorer without a Node, make sure to set the NBE_NODE_API environment variable to fake:

  1. NBE_NODE_API=fake python src/main.py
     

  2. docker run -e NBE_NODE_API=fake -p 8000:8000 nomos-block-explorer
     

Configuration

The block explorer is configured through environment variables. The following variables are available:

NBE_LOG_LEVEL=DEBUG  # DEBUG, INFO, WARNING, ERROR, CRITICAL

NBE_DEBUG=true  # Randomizes transactions in BlockSerializer

NBE_NODE_MANAGER=noop  # noop, docker
NBE_NODE_COMPOSE_FILEPATH=/path/to/docker-compose.yml  # Only used if NODE_MANAGER=docker

NBE_NODE_API=http  # fake, http
NBE_NODE_API_HOST=localhost  # Only used if NODE_API=http
NBE_NODE_API_PORT=18080  # Only used if NODE_API=http
NBE_NODE_API_TIMEOUT=60  # Only used if NODE_API=http
NBE_NODE_API_PROTOCOL=http  # Only used if NODE_API=http

NBE_HOST=0.0.0.0  # Block Explorer's listening host
NBE_PORT=8000  # Block Explorer's listening port

If running the Block Explorer with Docker, these can be overridden.

Considerations

This PoC makes simplifications to focus on the core features:

• Each slot has exactly one block.
• When backfilling, the block explorer will only backfill from the earliest block's slot to genesis.

Ideas and improvements

• Fix aforementioned assumptions
• Backfilling
  • Make requests concurrently
  • Backfill all slots
  • Upsert received blocks and transactions
• Database
  • Update to Postgres
  • Add migrations management
  • Add relevant indexes to columns
• Add interfaces to database repositories: BlockRepository and TransactionRepository
• Add tests
• Colour logs by level
• Reconnections
  • Failures to connect to Node
  • Timeouts
  • Stream closed
• Frontend
  • Add a block / transaction search barImprove
  • Make pages work with block/transaction hash, rather than the id
• Error handling
  • Exceptions raised within async code pop up as ugly stack traces
  • Better error messages
• Remove DB IDs from API responses and use hashes instead