@xtrata/reconstruction

Deterministic reconstruction helpers for rebuilding Xtrata inscription bytes from public chain data.

Core helpers:

• Chunk assembly
• Incremental hash verification
• Strict verification assertions
• Batch-first chunk reads with per-chunk fallback
• Ordered fallback source reconstruction for migrated content
• Provenance and read diagnostics
• Dependency graph resolution
• End-to-end reconstruction primitive

Current packaging mode:

• Source of truth: src/
• Build output: dist/
• Package entrypoint resolves from dist/index.js

Strict reconstruction

import { createXtrataReadClient, createXtrataReconstructionSources } from '@xtrata/sdk/simple';
import { reconstructXtrataInscription } from '@xtrata/reconstruction';

const core = createXtrataReadClient({
  contractId: 'SP3JNSEXAZP4BDSHV0DN3M8R3P0MY0EEBQQZX743X.xtrata-v2-1-1',
  senderAddress: 'SP3JNSEXAZP4BDSHV0DN3M8R3P0MY0EEBQQZX743X'
});
const legacy = createXtrataReadClient({
  contractId: 'SP3JNSEXAZP4BDSHV0DN3M8R3P0MY0EEBQQZX743X.xtrata-v2-1-0',
  senderAddress: 'SP3JNSEXAZP4BDSHV0DN3M8R3P0MY0EEBQQZX743X'
});

const result = await reconstructXtrataInscription({
  tokenId: 287n,
  sources: createXtrataReconstructionSources(core, [legacy]),
  strict: true,
  batchSize: 8,
  concurrency: 4
});

console.log(result.mimeType, result.bytes.length, result.diagnostics.chunkSourceId);

strict: true throws ReconstructionVerificationError if the rebuilt payload does not match the on-chain final-hash. It also rejects unsealed metadata when that flag is exposed.

batchSize and concurrency are optional. batchSize is clamped to the production ceiling of 30 chunks per read batch. These settings let production callers keep read-only reconstruction fast while preserving deterministic chunk order and strict hash verification. Platform adapters can also provide isTerminalReadError to stop immediately when retry or per-chunk fallback would only amplify an exhausted upstream quota.

Full rules: docs/reconstruction-spec.md in the main repository.