Skip to content

Instantly share code, notes, and snippets.

View tripolskypetr's full-sized avatar
💻
Frontend dev fueled by a passion for UI/UX and art, design

Petr Tripolsky tripolskypetr

💻
Frontend dev fueled by a passion for UI/UX and art, design
View GitHub Profile
@tripolskypetr
tripolskypetr / report.json
Created April 13, 2026 20:02
report.json
{
"report": {
"date": "01 февраля 2026 09:00 UTC",
"instruments": ["BTC"],
"prepared_for": "Краткосрочный трейдинг (1 час)",
"summary": {
"bias": "NEUTRAL-LOW / BEARISH",
"overall_sentiment": "Рынок в состоянии сильного страха, продолжается отток ликвидности, BTC упал ниже $80K",
"key_levels_btc": {
"current_price": 78978.73,
//@version=5
indicator("SuperTrend", overlay=true)
plotcandle(open, high, low, close,
color = close > open ? color.new(color.green, 70) : color.new(color.red, 70),
wickcolor = color.new(color.gray, 70),
bordercolor = color.new(color.gray, 70))
atrPeriod = input(10, "ATR Period")
factor = input(3.0, "Factor")
# GSI Mobile SDK
## Техническая документация
**Версия документа:** 1.0
**Дата:** Декабрь 2024
---
## Глоссарий
import { Subject, singleshot, randomString } from "functools-kit";
declare global {
interface Window {
onFetchApiResponse: (requestId: string, text: string) => void;
AndroidBinding: {
beginFetchApi: (requestId: string, url: string, totalChunks: number) => void;
sendFetchApiChunk: (requestId: string, chunkIndex: number, chunk: string) => void;
endFetchApi: (requestId: string, contentLength: number) => void;
};
@tripolskypetr
tripolskypetr / terminal.txt
Created December 23, 2025 08:37 — forked from joulgs/terminal.txt
How install libssl1.1 on ubuntu 22.04
wget http://archive.ubuntu.com/ubuntu/pool/main/o/openssl/libssl1.1_1.1.0g-2ubuntu4_amd64.deb
sudo dpkg -i libssl1.1_1.1.0g-2ubuntu4_amd64.deb
@tripolskypetr
tripolskypetr / README.md
Last active December 22, 2025 08:55
FX Cache Demo - Демонстрация критических проблем в финансовых системах

FX Cache Demo - Демонстрация критических проблем в финансовых системах

Этот проект воспроизводит три критические уязвимости в системе обработки курсов валют и банковского клиринга, которые приводят к финансовым потерям и проблемам производительности.

Проблема 1: Устаревшие курсы валют без кеширования

Описание проблемы

const ROI_BORDER_RATIO = 0.01;
interface IFaceDetect {
detectId: string;
bbox: { left: number; top: number; right: number; bottom: number };
detectionScore: number;
lowQuality: boolean;
}
interface IBbox {
@tripolskypetr
tripolskypetr / ETHUSDT_BUY.json
Created August 11, 2025 16:48
SIGNAL_HISTORY
{"history":[{"role":"system","content":"Вы - специализированная AI-система для анализа криптовалютных рынков для LONG позиций (покупка актива)\nПредоставьте торговую рекомендацию на основе технического анализа с четким обоснованием для LONG стратегии"},{"role":"system","content":"КРИТИЧНО: При action=\"buy\" покупка выполняется НЕМЕДЛЕННО! НЕ рекомендуй покупку, если нужны дополнительные подтверждения или ожидания. Рекомендация = немедленное исполнение. Только ГОТОВЫЕ к покупке сигналы с полным подтверждением СЕЙЧАС."},{"role":"system","content":"ПРИОРИТЕТ АНАЛИЗА - АДАПТИВНЫЙ ПОДХОД НА ОСНОВЕ ОБЪЕМОВ:\n- Индикаторы ЗАПАЗДЫВАЮТ. При стабилизации цены после падения индикаторы могут быть негативными - это НОРМАЛЬНО\n- ПРИ СТАБИЛЬНЫХ ОБЪЕМАХ (низкие колебания объемов свечей): 1) Минутные свечи/движение, 2) Объёмы, 3) Индикаторы\n- ПРИ ШУМНЫХ ОБЪЕМАХ (высокие колебания, резкие спайки, нестабильность): 1) Технические индикаторы, 2) Объёмы, 3) Минутные свечи (меньше веса)\n- Приоритет для ФИЛЬТРАЦИИ: часовой тренд

System: You are Grok 3 built by xAI.

The issue with the trend being consistently classified as "sideways" stems from the overly strict thresholds in the trend detection logic within the predict method of your OptimizedPricePredictor class. The conditions for identifying a bullish or bearish trend are too restrictive, especially for 1-minute candles, which often exhibit smaller price movements. As a result, the model defaults to "sideways" too frequently. Below, I’ll outline the key changes needed to make the trend detection more sensitive to short-term movements and incorporate additional indicators for better accuracy.

Proposed Fixes

  1. Adjust Thresholds: Lower the priceThreshold, slopeThreshold, and momentumThreshold to be more sensitive to the smaller price movements typical of 1-minute candles.
  2. Incorporate Volatility: Use volatility as an additional factor to gauge trend strength, ensuring it complements the trend detection logic.
  3. Enhanced Trend Logic: Include more indica