Orchestrate Brazil CVM Filings with Flyte

The problem

ML/data platforms that version research artifacts need durable CVM document UUIDs in workflow outputs — not scraped HTML blobs.

Solution flow

workflow → resolve → list documents → download → store UUID + PDF path

Task sketch (Python)

import os, requests

BASE = os.environ["APICVM_URL"]
H = {"Authorization": f"Bearer {os.environ['APICVM_KEY']}"}

def list_itr(ticker: str, year: int = 2024) -> list[dict]:
    r = requests.get(
        f"{BASE}/v1/documents",
        headers=H,
        params={"ticker": ticker, "type": "ITR", "year": year, "perPage": 20},
        timeout=60,
    )
    r.raise_for_status()
    return r.json()["data"]

Wrap list_itr in a Flyte task; keep document id in typed outputs for audit.

Current limitations

• Extraction is async via callback (Pro)
• No push feed for new filings; schedule re-lists
• Empty lists can be corpus gaps

Next steps

• Read the API docs
• Get an API key
• Dagster pipeline
• Temporal pipeline

Related guides

• Getting started
• Async PDF extraction