Data-MV Dataset

Collection of Japanese AV movies, organized by movie code.

Structure

• {CODE}.mp4 — Video file named by its unique movie code
• metadata.json — Full metadata (title, actress, genre, release date, etc.)
• movies.csv — Searchable table view (code, title, actress, release date)

How to Search by Actress or Title

Option 1: Browse movies.csv (easiest)

Open movies.csv — Hugging Face renders CSV files as an interactive table. You can search by actress name, title keyword, or release date directly in the browser.

Option 2: Download metadata.json

import requests, json
url = "https://huggingface.co/datasets/invokeaibase/data-mv/raw/main/metadata.json"
data = requests.get(url).json()
# Search by actress
mako_movies = [m for m in data if "Oda Mako" in m.get("actress", [])]
# Search by genre
bondage = [m for m in data if "Bondage" in m.get("genre", [])]

Option 3: Hugging Face Dataset Viewer

The movies.csv file is auto-indexed by HF's dataset viewer — just click the file in the repo browser.

Stats

• Total movies: 101
• With metadata: 70
• Actresses: Maria Himesaki, Oda Mako, and more compilation appearances

Indexing metadata for remaining movies (in progress)

The pipeline currently captures metadata for Oda Mako movies and will update metadata.json as more are processed.