Sign inSign up

jnovack/mcd

By jnovack

•Updated 8 days ago

Headless One-Shot Moxfield Collection Downloader

Image
Integration & delivery
Developer tools
Data science
0

3.9K

jnovack/mcd repository overview

⁠moxfield-collection-downloader

A headless Go CLI that downloads a Moxfield collection as JSON for local use or CI. It uses Playwright for session-aware requests and applies guarded pagination with integrity checks.

⁠Quickstart

mcd --id "cpfxIAEPH0aGHI-3r9F_xg"

# or

docker run --rm jnovack/mcd:2 --id "cpfxIAEPH0aGHI-3r9F_xg"

⁠Configuration

PurposeFlag(s)Environment Variable(s)DefaultNotes
Collection ID--id <collectionId>MCD_COLLECTION_ID, MCD_IDnoneUse either --id or --url.
Collection URL--url <collectionUrl>MCD_COLLECTION_URL, MCD_URLnoneMust be https://moxfield.com/collection/<id>.
Output path--output <path>MCD_OUTPUT./collection.jsonDirectory values resolve to collection.json.
Timeout (seconds)--timeout <seconds>MCD_TIMEOUT10Used as the base timeout for retries/backoff.
Log level--log-level <none|trace|debug|info|warn|error>MCD_LOG_LEVELinfoUse none to suppress all logs.
Force overwrite--forceMCD_FORCEfalseBypasses freshness guard. (Don't do this...)
Version only--versionn/an/aPrints version/build metadata and exits.

⁠Behavior

All good netizens need to be nice to APIs so we've implemented some guardrails:

  1. If output already exists and is newer than 72 hours, by default, retrieval is blocked.
  2. Retrieval starts with single-shot pageSize=totalResults and clamps only if totalResults > 20000.
  3. On timeout/block/mismatch, page size backs off: 20000 -> 10000 -> 5000 -> 1000 -> 500 -> 100.
  4. Pages are reassembled, deduplicated, and validated with totalResults == len(data).

⁠Disclaimer

This project is not affiliated with, endorsed by, or sponsored by Moxfield. But I do hope they secretly like this project.

Tag summary

Content type

Image

Digest

sha256:b33590c20…

Size

1.2 GB

Last updated

5 months ago

docker pull jnovack/mcd