keywordbarrage/myntra-scraper

Python Code:

import sqlite3
from datasets import load_dataset
import pandas as pd

# Method 1: Load via Hugging Face datasets library for exploratory research
dataset = load_dataset("daas-intelligence/enterprise-sku-depletion-telemetry")
print("Hub Sample Record:", dataset["train"][0])

# Method 2: Load zero-latency local pre-indexed SQLite deliverable for production pipelines
conn = sqlite3.connect("enterprise_feed.sqlite")
df = pd.read_sql_query("""
    SELECT sku_id, price_usd, stock_level, depletion_velocity 
    FROM sku_stream 
    ORDER BY timestamp DESC LIMIT 10
""", conn)
print(df)

---

Dataset Card: Enterprise High-Frequency SKU Depletion & WAF Telemetry

Dataset Summary

This repository contains an open 1,000-row verified benchmark sample of the Myntra SKU depletion feed. It is structured to demonstrate how quantitative hedge funds and data engineering teams can bypass brittle scraping infrastructure and Akamai Bot Manager/Edge Shield defenses. The feed provides 99.9% verified pipeline availability, bypassing traditional edge challenge loops via protocol-level parity.

Executive Summary

Brittle scrapers and broken proxy pools destroy data pipelines for quantitative analysts. This dataset provides a production-grade solution, delivering schema-ready telemetry that survives upstream WAF updates. Our ingestion engine maintains session decoupling, TLS/JA4 fingerprinting consistency, and request velocity management to deliver clean, noise-free market intelligence.

Enterprise Deliverables Matrix

We provide four distinct formats optimized for different stages of the data engineering lifecycle:

• Pre-indexed SQLite (.sqlite): Instant local querying with B-Tree indices on sku_id and timestamp.
• Snappy Parquet: High-compression partitions optimized for direct ingestion into Snowflake, BigQuery, and AWS Athena.
• Offline DataGrid (.html): A self-contained, interactive UI for quantitative stakeholders to perform ad-hoc visual verification.
• CDC JSONL Stream: Change Data Capture format designed for low-latency RAG pipeline updates and vector DB synchronization.

Dataset Schema

Column	Type	Nullable	Description
sku_id	STRING	No	Unique Myntra Style ID / Product Code
timestamp	DATETIME	No	ISO-8601 ingestion time
price_usd	FLOAT	No	Real-time selling price (converted from INR)
stock_level	INTEGER	No	Current inventory count per size variant
depletion_velocity	FLOAT	Yes	Calculated units/hour depletion rate
target_domain	STRING	No	Originating host (myntra.com)
detected_waf	STRING	No	Active WAF signature (Akamai Bot Manager)
tls_ja4_signature	STRING	No	JA4 fingerprint for protocol parity
delivery_status	STRING	No	Verification flag (200 OK / WAF-Bypass)

JSON Sample Record

{
  "sku_id": "MYN-9928374",
  "timestamp": "2026-09-24T14:30:00Z",
  "price_usd": 42.50,
  "stock_level": 14,
  "depletion_velocity": 0.85,
  "detected_waf": "Akamai-Bot-Manager-v2",
  "tls_ja4_signature": "t13d1516h2_88a6d88c88a6_a6d8",
  "delivery_status": "verified-bypass"
}

Service SLA & Reproducibility

• Availability: 99.9% uptime for daily feed ingestion.
• Parity: Deterministic schema parity with upstream Myntra API microservices.
• Latency: Sub-120ms protocol execution benchmarks for per-request session validation.
• Commercial Terms: This dataset is a sample of our managed extraction service. For production-scale daily feeds, we provide a turnkey extraction SLA. Our model eliminates the need for internal proxy pool maintenance, CAPTCHA solving, or headless browser infrastructure.