Files
ai/ingestor/src/symbol-metadata-generator.js

310 lines
8.8 KiB
JavaScript

// Symbol Metadata Generator - extracts market metadata from CCXT exchanges
import ccxt from 'ccxt';
export class SymbolMetadataGenerator {
constructor(config, kafkaProducer, logger) {
this.config = config;
this.kafkaProducer = kafkaProducer;
this.logger = logger;
this.exchanges = new Map();
this.lastGenerated = new Map(); // Track last generation time per exchange
this.publishedSymbols = new Set(); // Track published symbols to prevent duplicates
}
/**
* Get or create CCXT exchange instance
*/
getExchange(exchangeName) {
if (this.exchanges.has(exchangeName)) {
return this.exchanges.get(exchangeName);
}
const ExchangeClass = ccxt[exchangeName];
if (!ExchangeClass) {
throw new Error(`Unsupported exchange: ${exchangeName}`);
}
const exchange = new ExchangeClass({
enableRateLimit: true,
options: {
defaultType: 'spot'
}
});
this.exchanges.set(exchangeName, exchange);
this.logger.info({ exchange: exchangeName }, 'Created CCXT exchange instance for metadata');
return exchange;
}
/**
* Generate symbol metadata for all configured exchanges
*/
async generateAll() {
const supportedExchanges = this.config.supported_exchanges || ['binance'];
this.logger.info({ exchanges: supportedExchanges }, 'Generating symbol metadata for all exchanges');
const results = [];
for (const exchangeName of supportedExchanges) {
try {
const metadata = await this.generateForExchange(exchangeName);
results.push({ exchange: exchangeName, count: metadata.length, success: true });
} catch (error) {
this.logger.error(
{ error: error.message, exchange: exchangeName },
'Failed to generate metadata for exchange'
);
results.push({ exchange: exchangeName, error: error.message, success: false });
}
}
return results;
}
/**
* Generate symbol metadata for a single exchange
*/
async generateForExchange(exchangeName) {
const exchangeLower = exchangeName.toLowerCase();
const exchangeUpper = exchangeName.toUpperCase();
this.logger.info({ exchange: exchangeUpper }, 'Loading markets from exchange');
const exchange = this.getExchange(exchangeLower);
// Load all markets from the exchange
const markets = await exchange.loadMarkets();
this.logger.info(
{ exchange: exchangeUpper, marketCount: Object.keys(markets).length },
'Loaded markets from exchange'
);
const metadataList = [];
for (const [symbol, market] of Object.entries(markets)) {
try {
const metadata = this.convertMarketToMetadata(exchangeUpper, symbol, market);
// Debug log first few symbols
if (metadataList.length < 5) {
this.logger.info({
exchange: exchangeUpper,
symbol,
metadata
}, `Symbol metadata ${metadataList.length + 1} generated`);
}
metadataList.push(metadata);
} catch (error) {
this.logger.warn(
{ exchange: exchangeUpper, symbol, error: error.message },
'Failed to convert market to metadata'
);
}
}
// Write all metadata to Kafka
if (metadataList.length > 0) {
try {
await this.writeMetadataToKafka(metadataList);
} catch (error) {
this.logger.error(
{ exchange: exchangeUpper, error: error.message, stack: error.stack },
'Failed to write metadata to Kafka'
);
throw error;
}
}
// Update last generated timestamp
this.lastGenerated.set(exchangeUpper, Date.now());
this.logger.info(
{ exchange: exchangeUpper, count: metadataList.length },
'Generated and published symbol metadata'
);
return metadataList;
}
/**
* Convert CCXT market object to our Market metadata format
*/
convertMarketToMetadata(exchangeId, symbol, market) {
const base = market.base || '';
const quote = market.quote || '';
const marketType = market.type || 'spot';
// Extract precision information
const precision = market.precision || {};
const limits = market.limits || {};
// Derive Nautilus Instrument fields from CCXT market data
//
// CCXT precision.price can be:
// - Integer (decimal places): 2 means 0.01 tick size -> price_precision=2, tick_size=0.01
// - Float (tick size): 0.01 -> tick_size=0.01, price_precision=2
let price_precision;
let tick_size;
if (precision.price !== undefined) {
if (Number.isInteger(precision.price)) {
price_precision = precision.price;
tick_size = Math.pow(10, -precision.price);
} else {
tick_size = precision.price;
price_precision = Math.round(-Math.log10(precision.price));
}
} else if (limits.price?.min !== undefined) {
tick_size = limits.price.min;
price_precision = Math.round(-Math.log10(tick_size));
} else {
price_precision = 2;
tick_size = 0.01;
}
let size_precision;
let lot_size;
if (precision.amount !== undefined) {
if (Number.isInteger(precision.amount)) {
size_precision = precision.amount;
lot_size = Math.pow(10, -precision.amount);
} else {
lot_size = precision.amount;
size_precision = Math.round(-Math.log10(precision.amount));
}
} else if (limits.amount?.min !== undefined) {
lot_size = limits.amount.min;
size_precision = Math.round(-Math.log10(lot_size));
} else {
size_precision = 8;
lot_size = 0.00000001;
}
const min_notional = limits.cost?.min || 0;
const maker_fee = market.maker !== undefined ? market.maker : 0.001;
const taker_fee = market.taker !== undefined ? market.taker : 0.001;
// Standard supported periods (in seconds)
// Most exchanges support these timeframes
const supported_period_seconds = [
60, // 1m
300, // 5m
900, // 15m
1800, // 30m
3600, // 1h
14400, // 4h
86400, // 1d
];
// Build description
const description = `${base}/${quote} ${marketType} trading pair on ${exchangeId}`;
// NOTE: protobufjs expects camelCase field names, not snake_case!
return {
exchangeId: exchangeId,
marketId: symbol,
marketType: marketType,
description,
baseAsset: base,
quoteAsset: quote,
pricePrecision: price_precision,
sizePrecision: size_precision,
tickSize: tick_size,
lotSize: lot_size,
minNotional: min_notional,
makerFee: maker_fee,
takerFee: taker_fee,
contractMultiplier: 1.0,
supportedPeriodSeconds: supported_period_seconds,
// earliestTime can be added later if we track it
};
}
/**
* Write metadata list to Kafka topic
* Deduplicates symbols to prevent duplicate writes
*/
async writeMetadataToKafka(metadataList) {
const topic = 'symbol-metadata';
// Filter out duplicates
const uniqueMetadata = [];
let duplicateCount = 0;
for (const metadata of metadataList) {
const key = `${metadata.marketId}.${metadata.exchangeId}`;
// Debug first few to understand duplication
if (uniqueMetadata.length < 3 || (uniqueMetadata.length === 0 && duplicateCount < 3)) {
this.logger.info({
key,
exchangeId: metadata.exchangeId,
marketId: metadata.marketId,
isDuplicate: this.publishedSymbols.has(key),
setSize: this.publishedSymbols.size
}, 'Deduplication check');
}
if (!this.publishedSymbols.has(key)) {
uniqueMetadata.push(metadata);
this.publishedSymbols.add(key);
} else {
duplicateCount++;
}
}
if (uniqueMetadata.length === 0) {
this.logger.debug(
{ duplicateCount, topic },
'All symbols already published, skipping'
);
return;
}
// Convert each metadata to protobuf Market message
const messages = uniqueMetadata.map(metadata => {
const key = `${metadata.marketId}.${metadata.exchangeId}`;
return {
key,
value: metadata,
};
});
await this.kafkaProducer.writeMarketMetadata(topic, messages);
this.logger.info(
{ count: messages.length, duplicateCount, topic },
'Wrote symbol metadata to Kafka'
);
}
/**
* Get status of metadata generation
*/
getStatus() {
const status = {};
for (const [exchange, timestamp] of this.lastGenerated.entries()) {
status[exchange] = {
lastGenerated: new Date(timestamp).toISOString(),
ageMinutes: Math.floor((Date.now() - timestamp) / 60000),
};
}
return status;
}
/**
* Close all exchange connections
*/
async close() {
for (const [name, exchange] of this.exchanges) {
if (exchange.close) {
await exchange.close();
}
}
this.exchanges.clear();
}
}