naughtyamericavr scene identifcation by filename

crontab.sh

#!/bin/bash
# install in `crontab -e` as: @daily  $HOME/cvr_id_gist/crontab.sh

cd "${0%/*}"
./update_navr_csv.py && git add . && git commit -m "updates $(date +%Y-%m-%d)" && git push

navr.csv

navr_url.py

"""
navr_url - scraper for stashapp
2022, PoKeH

This script is meant to identify a scene where only filename is known
To collect metadata once the identification is done use this scraper:
  https://github.com/stashapp/CommunityScrapers/blob/master/scrapers/NaughtyAmericaVR.yml

Install:
- put this .py in config/scrapers directory
- also create new file navr_url.yml with following contents:
```
name: "navr_url"
sceneByFragment:
  action: script
  script:
    - python3
    - navr_url.py
```
- reload scrapers
Test:
- open NA scene for editing (eg. clsrname1name2_180x180_3dh)
- "Scrape with.. / navr_url", if scene is known it offers a new URL
- click button next to URL field to fetch metadata
Use with lots of files:
- in Scenes list mass-select NA scenes
- under "..." menu select Identify
- add navr_url as Source, click Identify
- with URLs in place run metadata scraper
"""
from urllib.request import Request, urlopen
import time
import json
import sys
import os


def format_rfc1123(timestamp):
    return time.strftime('%a, %d %b %Y %H:%M:%S GMT', time.gmtime(timestamp))


def download_if_expired(url, path, maxage=60*30):
    headers = {}
    if os.path.exists(path):
        mtime = os.path.getmtime(path)
        headers["If-Modified-Since"] = format_rfc1123(mtime)

    if not os.path.exists(path) or time.time() - mtime > maxage:
        r = urlopen(Request(url, None, {}))
        if r.status == 200:
            with open(path, 'wb') as fh:
                fh.write(r.read())


FRAGMENT = json.loads(sys.stdin.read())
# print(FRAGMENT, file=sys.stderr)
TITLE = FRAGMENT.get("title")
slug = TITLE.split("_")[0]

CACHE_DIR = os.getenv("STASH_CACHE", ".")
CSV_URL = "https://gist.githubusercontent.com/poontology/a093624982eb43befc8f1bb93d56d6c2/raw/navr.csv"
CSVFILE = CACHE_DIR + "/navr.csv"
download_if_expired(CSV_URL, CSVFILE)
#res = urlopen(Request(navr_url, None, {}))
#lst = [l.split(",") for l in res.read().decode().split("\n")]
lst = [l.split(",") for l in open(CSVFILE).read().split("\n")]

match = list(filter(lambda s: s[0].replace("/", "") == slug, lst))

if not match and slug.startswith("nam"):
    # they renamed some old releases from nam-prefix to new "studios"
    match = list(filter(lambda s: s[0].endswith(slug[3:]), lst))
    print("== nam-prefix didn't match known scenes, trying to fuzz", file=sys.stderr)

print("== %s matches %s" % (slug, match), file=sys.stderr)

if not match:
    print(json.dumps({}))
    sys.exit(0)

NA_BASE = "https://www.naughtyamericavr.com/scene/-"
res = {"url": NA_BASE + match[0][1]}
print(json.dumps(res))

update_navr_csv.py

#!/usr/bin/env python3
"""
update_navr_csv - crawls NAVR scene index and collects URL slug to id mappings
2022, PoKeH

NAVR video filenames typically use the slug as prefix but to get to
the metainfo on the webpage also the id number needs to be known
"""

from urllib.request import Request, urlopen
from urllib.parse import urlparse
from operator import itemgetter
from lxml import html
import time
import re


DEFAULT_UA = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/100.0.4896.127 Safari/537.36"

text = lambda e: e.xpath('normalize-space(text())') if type(e)!=list else list(map(text, e))


def retrieve(url, headers={}):
    if 'User-Agent' not in headers:
        headers['User-Agent'] = DEFAULT_UA
    page = urlopen(Request(url, None, headers))
    return html.parse(page).getroot()


def getnascenes(page):
    res = retrieve("https://www.naughtyamericavr.com/vr-videos?page=%s" % page)
    ro = []
    for i,t in enumerate(res.cssselect('.pornfinder-scene')):
        d = {
            'path': urlparse(t.cssselect('a')[0].attrib['href']).path,
            'title': text(t.cssselect('.pornfinder-title'))[0],
            'picture': t.cssselect('img')[0].attrib['src'],
        }
        d['na_id'] = d['path'].split('-')[-1]
        picm = re.match(r'.*scenes/(.*)/scene', d['picture'])
        if picm:
            d['slug'] = picm.group(1)
        ro.append(d)
    ro = sorted(ro, key=itemgetter('na_id'))
    return ro


def getnalastpage():
    res = retrieve("https://www.naughtyamericavr.com/vr-videos")
    el = res.cssselect('#last-page')
    return int(el[0].attrib['href'].split('=')[-1])


if __name__ == "__main__":
    filename = "navr.csv"
    csv_data = open(filename).read().strip()
    scene_lst = [l.split(",") for l in csv_data.split('\n') if "," in l]
    last_page = getnalastpage()
    items_per_page = 4 * 3
    max_pg = max(1, int(last_page - len(scene_lst) / items_per_page))
    # last_id = int(scene_lst[-1][1])
    scene_dic = {d[1]:d[0] for d in scene_lst}

    for pg in range(max_pg, 0, -1):
        new_items = []

        for d in getnascenes(pg):
            if d["na_id"] in scene_dic: continue
            new_items.append("%(slug)s,%(na_id)s" % d)

        with open(filename, "a") as fh:
            fh.write("\n".join(new_items) + "\n")
            print("pg %i had %i new items" % (pg, len(new_items)))
            time.sleep(0.5)

Looks like over time they might rename a scene, release it with new id and drop the old one from the site. I'm only going to run incremental updates to the CSV in this gist so theoretically it's possible that some id in the list no longer exists on the site.
At the moment I think bcb/marierachelrileykyle,25727 -> no/marierachelrileykyle,30765 is the only one where that's happened, I guess we could just map the old name to the new id since the metadata is essentially same. Comment if you care.