@absinthelabs/adapters

Shared types and Zod schemas used across the Absinthe indexing pipeline.
This package defines the canonical shape of enriched actions, enriched positions, and adapter configuration — ensuring all Absinthe services interpret protocol data consistently.

It is currently a types-only package that exports:

• Enriched Action schema & type
• Enriched Position schema & type
• RegisterConfig schema & type
• Chain/Asset/Denomination/Measurement enums
• POSITION_REASONS constant
• CommonFields + AssetFields shared structures

This package forms the foundation of the new adapter architecture (manifest-driven, priced, multi-protocol).

📦 Installation

npm install @absinthelabs/adapters
# or
pnpm add @absinthelabs/adapters

🛠 Usage Examples

Validate an Incoming Action

import { ActionSchema } from '@absinthelabs/adapters';

const action = ActionSchema.parse(rawData);What this does:

• ✅ Validates all required fields are present (user, ts_ms, height, tx_ref, etc.)
• ✅ Ensures field types match the schema
• ✅ Automatically strips extra fields not in the schema
• ❌ Throws ZodError if validation fails (use .safeParse() for non-throwing validation)

Validate a Position

script import { PositionSchema } from '@absinthelabs/adapters';

const pos = PositionSchema.parse(rawPos);What this does:

• ✅ Validates position window fields (window_utc_start_ts_ms, window_utc_end_ts_ms, etc.)
• ✅ Validates asset fields (asset_key, decimals, asset_type) - required for positions
• ✅ Validates common fields (chain metadata, measurement type, etc.)
• ❌ Throws if any required field is missing

Use TypeScript Types

import type { EnrichedAction, EnrichedPosition } from '@absinthelabs/adapters';

function consume(p: EnrichedPosition) { console.log(p.user, p.raw_after); // TypeScript knows all available fields and their types }Benefits:

• 🎯 Full type safety and autocomplete
• 🔍 Type inference from Zod schemas (single source of truth)
• ⚡ No runtime overhead (types are stripped at compile time)

📚 Exports

✔ Zod Schemas

| Schema | Purpose | Use Case | | ---------------------------- | --------------------------------------------------------------- | ---------------------------------- | | EnrichedActionSchema | Validates enriched action events | API ingestion, database validation | | EnrichedPositionSchema | Validates enriched position windows | Position tracking, analytics | | RegisterConfigSchema | Validates adapter registration config | Adapter setup | | ChainArchSchema | Validates chain architecture (evm | solana) | Chain-specific logic | | MeasurementTypeSchema | Validates measurement type (token_based | count | none) | Trackable configuration | | DenominationSchema | Validates denomination (usd | scaled_token | none) | Pricing configuration | | AssetEnum | Validates asset type (erc20 | erc721 | spl | custom) | Asset identification |

✔ TypeScript Types

| Type | Inferred From | Description | | ---------------------- | ------------------------ | ---------------------------------- | | EnrichedAction | EnrichedActionSchema | Type-safe action event shape | | EnrichedPosition | EnrichedPositionSchema | Type-safe position window shape | | RegisterConfig | RegisterConfigSchema | Adapter registration configuration |

⚠️ Common Validation Errors

Missing Required Fields

// ❌ This will throw
ActionSchema.parse({ user: '0x123' });
// Error: Required field 'ts_ms' is missingcript
// ✅ Use safeParse for graceful handling
const result = ActionSchema.safeParse({ user: '0x123' });
if (!result.success) {
console.error(result.error.errors);
}---

Type Mismatches

// ❌ This will throw
ActionSchema.parse({
user: '0x123',
ts_ms: 'not-a-number' // Should be number
});
// Error: Expected number, received string---

Unknown Fields

// ✅ Extra fields are automatically stripped
const action = ActionSchema.parse({
user: '0x123',
ts_ms: Date.now(),
height: 12345,
tx_ref: '0xabc',
extraField: 'will be removed', // Automatically stripped
// ... all required fields
});
// Result: action.extraField is undefined

Note: The schema uses .strip() which automatically removes any fields not defined in the schema definition. This ensures data consistency across all consumers.