Prefill stash scene studio codes from JAV video filenames

stash_jav_codes_from_filename.py

#!/usr/bin/env python3
# coding: utf-8

# utility script for https://github.com/stashapp/stash
# 2025, poontology

# Requires: working regexp_capture function,
#   wget -q https://github.com/nalgeon/sqlean/releases/download/0.28.0/sqlean-linux-x64.zip
#   unzip -q sqlean-linux-x64.zip -d /usr/lib/sqlite3/

import sqlite3
import sys
import os


def update_scene_codes_from_filenames(cur, prefix):
    rx = '(%s-\d+)' % prefix
    # ignores files <10MB in size
    cur.execute("""UPDATE scenes
                    SET code = regexp_capture(f.basename, ?, 1)
                    FROM files f
                    JOIN scenes_files sf ON sf.file_id = f.id
                    WHERE sf.scene_id = scenes.id
                      AND f.basename REGEXP ? AND f.size > 1e7
                      AND scenes.code IS NULL;
                   """, (rx, rx))

def ask_continue():
    try:
        answer = input("Continue (y/n)? ")
        return answer.lower() == 'y'
    except KeyboardInterrupt:
        return False


if __name__ == '__main__':
    if len(sys.argv) != 3:
        print("Syntax:", sys.argv[0], "stash-go.sqlite JAV_ID_PREFIX")
        sys.exit(1)

    dbfile, jav = sys.argv[1:3]
    if not os.access(dbfile, os.R_OK):
        print("Error: file not readable: ", dbfile)
        sys.exit(2)
    if not isinstance(jav, str) or not (3 <= len(jav) <= 5):
        print("Warning: JAV prefix doesn't seem valid (3-5 char long alphabetic value):", jav)
        if not ask_continue():
            print("cancelled")
            sys.exit(3)

    con = sqlite3.connect(dbfile)
    con.enable_load_extension(True)
    con.execute("SELECT load_extension('/usr/lib/sqlite3/regexp.so')")

    try:
        cur = con.cursor()

        update_scene_codes_from_filenames(cur, jav)
        print("copied", cur.rowcount, "JAV codes for prefix:", jav)
        input("Press enter to continue (ctrl-c to cancel)")

        con.commit()
        print("done")
    except KeyboardInterrupt:
        print("cancelled")
    finally:
        con.close()