QuantLens
JSONL for ML Pipelines
Streamable, line‑delimited JSON with stable schemas, integrity hashes, and per‑pack artifacts.
Why JSONL
Streamable
Consume one record at a time without loading entire files. Ideal for large‑scale training and ETL.
Integrity & Versioning
Per‑pack JSON Schema, rule sets, SHA‑256 manifests and optional Merkle roots.
Training‑Ready
Stable schema versions and contracts ensure reproducible pipelines and minimal parsing overhead.
Samples
Read JSONL
Python
import json
with open('sec_earnings_sample.jsonl', 'r', encoding='utf-8') as f:
for line in f:
rec = json.loads(line)
# use rec["transcript"], rec["metrics"], ...
Node.js
import fs from 'node:fs'
import readline from 'node:readline'
const rl = readline.createInterface({ input: fs.createReadStream('sec_earnings_sample.jsonl') })
for await (const line of rl) {
const rec = JSON.parse(line)
}
Schema & Integrity
Each pack publishes a JSON Schema and rule set; integrity via SHA‑256 per file and optional Merkle manifest.