UnityAILab/Deviant-Art
4
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
Schizo Art
4 commits 3 branches 0 tags 108 KiB
  • Python 100%
Find a file
Exact
GFourteen 88e227985b merge(develop): catalog-descriptions lands in main
2026-06-23 06:43:54 -06:00
.gitignore chore: initialize Deviant-Art catalog repo (README + gitignore) 2026-06-23 06:43:37 -06:00
deviantart_descriptions.json feat(catalog): harvest + descriptions list for all 355 DeviantArt pieces 2026-06-23 06:43:47 -06:00
deviantart_descriptions.md feat(catalog): harvest + descriptions list for all 355 DeviantArt pieces 2026-06-23 06:43:47 -06:00
harvest_descriptions.py feat(catalog): harvest + descriptions list for all 355 DeviantArt pieces 2026-06-23 06:43:47 -06:00
README.md chore: initialize Deviant-Art catalog repo (README + gitignore) 2026-06-23 06:43:37 -06:00

README.md

Deviant-Art — GFourteen Portfolio Catalog

A catalog of GFourteen's DeviantArt portfolio: every piece's name paired with its Description write-up, harvested straight from deviantart.com.

The artwork itself (932MB of images) is not tracked here — it lives on disk under downloads/ (gitignored) and on DeviantArt. What this repo tracks is the catalog and the tool that builds it.

Contents

File What it is
harvest_descriptions.py The harvester. Parses each image filename for its deviation ID + title, fetches the deviation page, extracts the artist Description, and writes the list. Resume-aware, rate-limit-aware.
deviantart_descriptions.md Human-readable list — ## Name then the Description body. 355 pieces.
deviantart_descriptions.json Structured form — id, name, filename_title, description, file per piece.

Stats

  • 355 pieces catalogued
  • 336 with descriptions (pulled verbatim)
  • 19 with no description (genuinely blank on DeviantArt)

Regenerating

Drop the DeviantArt export under downloads/deviantart/GFourteen/ (files named deviantart_<id>_<title>.<ext>), then:

python harvest_descriptions.py

It resumes from deviantart_descriptions.json if present (skips already-harvested pieces), paces requests at 3s, and backs off on HTTP 403/429 rate limits.