feat(romhacks): add Discord rom-hack ingestion stack (DCE -> JDownloader -> RomM)

Inert by default: empty token keeps the orchestrator idle and DRY_RUN=true
suppresses downloads. Orchestrator exports per-generation forum threads via
DiscordChatExporter, extracts download links + art, stages metadata, and writes
JDownloader crawljobs. Armed here with an alt account token and all 11 romhack
forums, still in DRY_RUN pending first-cycle validation.
This commit is contained in:
ginnoir
2026-06-07 18:53:23 -05:00
parent 91d6724175
commit ba207a0fcd
8 changed files with 460 additions and 0 deletions
+86
View File
@@ -0,0 +1,86 @@
# romhacks stack
Automated ingestion of Pokémon rom-hacks from a Discord server's per-generation
threads into the RomM library.
```
romhacks (orchestrator) jdownloader (jlesage) RomM (roms stack)
DiscordChatExporter ── JSON ──▶ FolderWatch reads .crawljob, scans
parse links + art ── crawljob ▶ resolves ~all hosts, ──────▶ /storage1/Emulation
stage metadata ── art ─────▶ downloads to /output
```
## State: INERT
Ships idle and safe. It will not export or download until armed:
- `DISCORD_TOKEN` / `GUILD_ID` empty → orchestrator loop idles (see `run.sh`).
- `DRY_RUN=true` → even configured, it parses + stages metadata but writes **no**
`.crawljob`, so JDownloader fetches nothing.
## Why this design
- **Member-only Discord access** → no official bot possible. DiscordChatExporter
(DCE) does a **read-only** export with a **user token**. That violates Discord
ToS, so this is wired for an **alt account** (joined to the server purely for
this) to keep any ban risk off the main account.
- **Links span many file hosts** → JDownloader2 auto-resolves hundreds of hosts
and handles waits/captcha via its noVNC UI; far more robust than per-host glue.
- **Art + blurb** → DCE `--media` downloads inline thread images; the orchestrator
stages them with the thread name/description for a later RomM custom-cover
hand-off (pairs with the `deck-sgdb` art pipeline).
## Arming checklist
1. **Alt account**: create it, join the server, grab its user token → `DISCORD_TOKEN`
in `stack.env`. Set `GUILD_ID` (right-click server → Copy Server ID; Developer
Mode on).
2. **Channels**: fill `orchestrator/channels.json` with the per-generation channel
(or forum) IDs. Add any new file hosts you see to `hosts`.
3. **Register the stack** in Portainer once (new git stacks need a one-time
registration — see memory `portainer-new-stack-registration`). Build is on:
Portainer builds `homelab/romhacks-orchestrator` from `./orchestrator`.
4. **JDownloader FolderWatch**: open the noVNC UI at `http://valhalla:8998`
(set `VNC_PASSWORD` first), Settings → Extensions → **Folder Watch** → enable,
add watch folder `/watch`. This is what consumes the orchestrator's crawljobs.
5. **Dry run**: leave `DRY_RUN=true`, let one cycle run, inspect
`/storage1/labdata/romhacks/metadata/*` — confirm names, blurbs, art, and that
the extracted `links` look like real ROM links (not random URLs).
6. **Arm**: flip `DRY_RUN=false`. Downloads land in
`/storage1/labdata/romhacks/incoming/<game>/`.
7. **Caddy (optional, internal-only)** — expose the JD UI on the LAN:
```
jd.ginnoir.com {
import internal_only
reverse_proxy jdownloader:5800
}
```
then `./scripts/gen-bookmarks.ps1` and push (runner reloads Caddy).
8. **RomM hand-off** (next phase, not built yet): sort `incoming/<game>` into the
right `/storage1/Emulation/<platform>` folder and push the staged art to RomM's
custom-cover API.
## Editing config
`channels.json` and the scripts are **baked into the image** (Portainer's git-stack
checkout doesn't reliably bind sibling repo files — see the compose header). So a
change there needs a **push + rebuild**; if a content-only edit doesn't trigger a
rebuild, use Portainer → the stack → **Re-pull and redeploy**.
## Paths
| Host path | Purpose |
|---|---|
| `/config/romhacks/exports` | DCE JSON + downloaded media (SSD) |
| `/config/romhacks/state` | processed-message-id dedupe (SSD) |
| `/config/romhacks/crawljobs` | orchestrator → JDownloader FolderWatch (`/watch`) |
| `/config/romhacks/jd` | JDownloader config (SSD) |
| `/storage1/labdata/romhacks/incoming` | JDownloader downloads (ZFS) |
| `/storage1/labdata/romhacks/metadata` | staged name/blurb/art per game (ZFS) |
## Gotchas
- **ToS / ban risk** is real but low for read-only export. Use the alt account.
- **Pre-patched ROMs vs patches**: bundled base-game ROMs are piracy; IPS/BPS
patches are broadly tolerated. Prefer patch links where the thread offers both.
- DCE flags / dll path are pinned in `orchestrate.py`; verify on first run (the
code globs for the dll if the pinned path moves between image versions).
+83
View File
@@ -0,0 +1,83 @@
# romhacks stack — automated Pokémon rom-hack ingestion from a Discord server's
# per-generation threads into the RomM library.
#
# FLOW
# romhacks (orchestrator) jdownloader (jlesage) RomM (roms stack)
# ┌───────────────────────┐ ┌──────────────────┐ ┌──────────────┐
# │ DiscordChatExporter ──▶│ JSON │ FolderWatch picks│ files │ scans │
# │ parse links + art ───▶│ crawljob│ up .crawljob, │ ───────▶ │ /storage1/ │
# │ stage metadata ───▶│ ──────▶ │ resolves ~all │ │ Emulation │
# └───────────────────────┘ │ hosts, downloads │ └──────────────┘
# └──────────────────┘
#
# WHY THIS SHAPE
# - Member-only access => no official Discord bot is possible. DiscordChatExporter
# (DCE) does a READ-ONLY export with a USER token. Self-token use violates
# Discord ToS, so this is wired for an ALT account (see stack.env) to isolate
# ban risk from the main account.
# - The download links span many file hosts (Mega/Drive/MediaFire/Pixeldrain/...),
# so JDownloader2 (jlesage image) does the actual fetching — it auto-resolves
# hundreds of hosts and handles waits/captcha via its noVNC UI.
# - DCE `--media` downloads the inline thread images (box art / screenshots); the
# orchestrator stages those + the thread blurb as metadata for a later RomM
# custom-cover hand-off (dovetails with the deck-sgdb art pipeline).
#
# INERT BY DEFAULT. Does nothing until armed: DISCORD_TOKEN/GUILD_ID are empty and
# DRY_RUN=true, so even if deployed it idles and never enqueues a download. See
# README.md for the arming checklist.
#
# BUILD (not bind) for the orchestrator: Portainer git-stack checkouts do NOT
# reliably materialize sibling repo files at relative FILE bind paths (Docker then
# auto-creates a directory — see the `share` stack note). So the orchestrator's
# scripts + channels.json are baked into a built image from ./orchestrator instead
# of bind-mounted. Editing channels.json => push => Portainer rebuild (use
# "Re-pull and redeploy" if a content-only change doesn't trigger a rebuild).
#
# Tiered binds: working state (exports/state/crawljobs + JD config) -> /config
# (SSD); bulk output + art -> /storage1/labdata/romhacks (ZFS, root-auto-created).
#
# Only jdownloader joins `edge` (Caddy proxies its noVNC UI, internal-only).
services:
romhacks:
build:
context: ./orchestrator
image: homelab/romhacks-orchestrator:latest
container_name: romhacks
restart: unless-stopped
networks: [romhacks]
env_file:
- stack.env
volumes:
- /config/romhacks/exports:/exports
- /config/romhacks/state:/state
- /config/romhacks/crawljobs:/crawljobs
- /storage1/labdata/romhacks/metadata:/metadata
jdownloader:
image: jlesage/jdownloader-2:latest
container_name: jdownloader
restart: unless-stopped
networks: [romhacks, edge]
env_file:
- stack.env
environment:
# root so it can write to the ZFS-tier /output bind (dirs auto-created as root)
- USER_ID=0
- GROUP_ID=0
volumes:
- /config/romhacks/jd:/config
# shared with the orchestrator: it drops .crawljob files here; the JD
# FolderWatch extension must be enabled + pointed at /watch (one-time, README).
- /config/romhacks/crawljobs:/watch
- /storage1/labdata/romhacks/incoming:/output
ports:
- "8998:5800"
networks:
romhacks:
name: romhacks
driver: bridge
edge:
name: edge
external: true
+2
View File
@@ -0,0 +1,2 @@
__pycache__/
*.pyc
+15
View File
@@ -0,0 +1,15 @@
# Orchestrator = DiscordChatExporter (has .NET + the DCE CLI) + Python for parsing.
# Baking scripts + channels.json into the image avoids Portainer's relative
# FILE-bind quirk (see docker-compose.yml header).
FROM tyrrrz/discordchatexporter:stable
USER root
RUN apt-get update \
&& apt-get install -y --no-install-recommends python3 python3-requests \
&& rm -rf /var/lib/apt/lists/*
WORKDIR /app
COPY orchestrate.py run.sh channels.json ./
RUN chmod +x run.sh
ENTRYPOINT ["./run.sh"]
@@ -0,0 +1,26 @@
{
"_comment": "Forum channels to harvest (type 15 — each thread is one rom-hack). DCE's --include-threads all captures the threads beneath each forum. 'platform' is a hint for the later RomM hand-off and is not used yet. Editing this file requires a push + Portainer rebuild (it is baked into the orchestrator image). Cheat-code forums are intentionally omitted.",
"hosts": [
"mega.nz",
"mediafire.com",
"drive.google.com",
"pixeldrain.com",
"gofile.io",
"1fichier.com",
"workupload.com",
"krakenfiles.com"
],
"channels": [
{ "id": "1474174360926290105", "label": "gen-1-romhacks", "platform": "gb" },
{ "id": "1474174532196761802", "label": "gen-2-romhacks", "platform": "gbc" },
{ "id": "1474174608587489416", "label": "gen-3-romhacks", "platform": "gba" },
{ "id": "1474174672210755714", "label": "gen-4-romhacks", "platform": "nds" },
{ "id": "1474174736438132756", "label": "gen-5-romhacks", "platform": "nds" },
{ "id": "1474174789680632084", "label": "gen-6-romhacks", "platform": "3ds" },
{ "id": "1474174859671240898", "label": "gen-7-romhacks", "platform": "3ds" },
{ "id": "1474174950138183880", "label": "gen-8-romhacks", "platform": "switch" },
{ "id": "1474175004248903935", "label": "gen-9-romhacks", "platform": "switch" },
{ "id": "1474175244209098986", "label": "joiplay-rpgxp-games", "platform": "pc" },
{ "id": "1475605095805747351", "label": "unique-romhacks", "platform": "" }
]
}
+205
View File
@@ -0,0 +1,205 @@
#!/usr/bin/env python3
"""romhacks orchestrator.
One cycle:
1. For each configured channel, run DiscordChatExporter (DCE) to dump the
channel + all its threads to JSON, downloading inline media (box art).
2. Parse every exported JSON. For each NOT-yet-seen message, pull download
links (filtered to known file hosts) and any image attachments.
3. Stage thread name + blurb + art + links as metadata (for a later RomM
custom-cover hand-off), and — unless DRY_RUN — write a JDownloader
.crawljob so JD fetches the links.
Idempotent: the full thread JSON is re-parsed each cycle; a message-id state file
ensures each post is only enqueued once. INERT guards live in run.sh (token/guild)
and here (DRY_RUN).
"""
import glob
import hashlib
import json
import os
import pathlib
import re
import shutil
import subprocess
import sys
TOKEN = os.environ.get("DISCORD_TOKEN", "").strip()
DRY_RUN = os.environ.get("DRY_RUN", "true").lower() != "false"
JD_DOWNLOAD_ROOT = os.environ.get("JD_DOWNLOAD_ROOT", "/output")
# Paths default to the in-container mounts; overridable via env for local testing.
EXPORTS = pathlib.Path(os.environ.get("EXPORTS_DIR", "/exports"))
STATE = pathlib.Path(os.environ.get("STATE_FILE", "/state/processed.json"))
CRAWLJOBS = pathlib.Path(os.environ.get("CRAWLJOBS_DIR", "/crawljobs"))
METADATA = pathlib.Path(os.environ.get("METADATA_DIR", "/metadata"))
CHANNELS_FILE = pathlib.Path(os.environ.get("CHANNELS_FILE", "/app/channels.json"))
# Fallback host list; overridden by channels.json -> "hosts".
DEFAULT_HOSTS = [
"mega.nz", "mediafire.com", "drive.google.com", "pixeldrain.com",
"gofile.io", "1fichier.com", "workupload.com", "krakenfiles.com",
"bunkr", "anonfiles",
]
URL_RE = re.compile(r"https?://[^\s<>()\[\]]+", re.I)
IMG_EXTS = (".png", ".jpg", ".jpeg", ".webp", ".gif")
def load_json(p):
try:
return json.loads(pathlib.Path(p).read_text(encoding="utf-8"))
except Exception:
return None
def slug(s):
return re.sub(r"[^A-Za-z0-9._-]+", "_", s or "").strip("_")[:120] or "unknown"
def find_dce():
direct = "/opt/app/DiscordChatExporter.Cli.dll"
if os.path.exists(direct):
return direct
hits = glob.glob("/opt/**/DiscordChatExporter.Cli.dll", recursive=True)
return hits[0] if hits else None
def run_dce(channel_id):
dll = find_dce()
if not dll:
print("[romhacks] DiscordChatExporter.Cli.dll not found in image", file=sys.stderr)
return False
out = EXPORTS / channel_id
out.mkdir(parents=True, exist_ok=True)
cmd = [
"dotnet", dll, "export",
"-t", TOKEN,
"-c", channel_id,
"-f", "Json",
"--include-threads", "all",
"--media", "True",
"--reuse-media", "True",
"-o", str(out / "%c.json"),
]
print(f"[romhacks] exporting channel {channel_id}")
return subprocess.run(cmd).returncode == 0
def load_state():
return set((load_json(STATE) or {}).get("processed", []))
def save_state(processed):
STATE.parent.mkdir(parents=True, exist_ok=True)
STATE.write_text(json.dumps({"processed": sorted(processed)}), encoding="utf-8")
def host_match(url, hosts):
u = url.lower()
return any(h in u for h in hosts)
def write_crawljob(game, links):
CRAWLJOBS.mkdir(parents=True, exist_ok=True)
job = [{
"text": "\n".join(links),
"packageName": game,
"downloadFolder": f"{JD_DOWNLOAD_ROOT}/{slug(game)}",
"enabled": "TRUE",
"autoStart": "TRUE",
"autoConfirm": "TRUE",
}]
digest = hashlib.sha1("|".join(links).encode()).hexdigest()[:8]
f = CRAWLJOBS / f"{slug(game)}-{digest}.crawljob"
f.write_text(json.dumps(job, indent=2), encoding="utf-8")
print(f"[romhacks] enqueued {len(links)} link(s) for '{game}' -> {f.name}")
def stage_metadata(game, description, links, art_files):
d = METADATA / slug(game)
d.mkdir(parents=True, exist_ok=True)
meta = {
"name": game,
"description": description,
"links": links,
"art": [a.name for a in art_files],
}
(d / "metadata.json").write_text(json.dumps(meta, indent=2), encoding="utf-8")
for a in art_files:
try:
shutil.copy2(a, d / a.name)
except Exception as e:
print(f"[romhacks] art copy failed {a}: {e}", file=sys.stderr)
def process_file(path, hosts, processed):
data = load_json(path)
if not data:
return 0
channel = data.get("channel", {})
game = channel.get("name") or "unknown"
msgs = data.get("messages", [])
description = next((m.get("content", "") for m in msgs if (m.get("content") or "").strip()), "")
links, art, new = [], [], 0
for m in msgs:
mid = m.get("id")
if not mid or mid in processed:
continue
# Links usually live in the message body as markdown links, but some posts
# only surface them via an auto-embed — scan both.
blobs = [m.get("content") or ""]
for emb in m.get("embeds", []):
blobs.append(emb.get("url") or "")
blobs.append(emb.get("description") or "")
for fld in emb.get("fields", []):
blobs.append(fld.get("value") or "")
for blob in blobs:
for u in URL_RE.findall(blob):
if host_match(u, hosts):
links.append(u)
for att in m.get("attachments", []):
if att.get("fileName", "").lower().endswith(IMG_EXTS):
# DCE --media rewrites attachment "url" to a path relative to the export
local = pathlib.Path(path).parent / att.get("url", "")
if local.exists():
art.append(local)
processed.add(mid)
new += 1
links = list(dict.fromkeys(links))
if links:
if DRY_RUN:
print(f"[romhacks][DRY_RUN] would enqueue {len(links)} link(s) for '{game}'")
else:
write_crawljob(game, links)
if links or art:
stage_metadata(game, description, links, art)
return new
def main():
cfg = load_json(CHANNELS_FILE) or {}
channels = cfg.get("channels", [])
hosts = cfg.get("hosts", DEFAULT_HOSTS)
if not channels:
print("[romhacks] no channels configured in channels.json — nothing to do")
return
processed = load_state()
for ch in channels:
cid = str(ch.get("id", "")).strip()
if not cid:
continue
if not run_dce(cid):
print(f"[romhacks] export failed for {cid}, skipping", file=sys.stderr)
total_new = 0
for jf in EXPORTS.glob("**/*.json"):
total_new += process_file(jf, hosts, processed)
save_state(processed)
print(f"[romhacks] cycle done — {total_new} new message(s); DRY_RUN={DRY_RUN}")
if __name__ == "__main__":
main()
+15
View File
@@ -0,0 +1,15 @@
#!/bin/sh
# Harvest loop. Stays idle (no export, no enqueue) until the alt-account token and
# guild id are configured, so deploying the stack unconfigured is harmless.
set -u
POLL_INTERVAL="${POLL_INTERVAL:-21600}"
while true; do
if [ -z "${DISCORD_TOKEN:-}" ] || [ -z "${GUILD_ID:-}" ]; then
echo "[romhacks] DISCORD_TOKEN/GUILD_ID not set — idle. Sleeping ${POLL_INTERVAL}s."
else
python3 /app/orchestrate.py || echo "[romhacks] cycle failed (exit $?)"
fi
sleep "${POLL_INTERVAL}"
done
+28
View File
@@ -0,0 +1,28 @@
# romhacks stack — INERT until armed.
#
# Use an ALT Discord account's token. Self-token use violates Discord ToS; read-only
# export is the low-risk end of that gray zone, but an alt account that has joined
# the server purely for this isolates any ban risk from your main account.
TZ=America/Chicago
# ── Discord (ALT account) ────────────────────────────────────────────────────
# User token from the ALT account. Read-only export only. Rotating the alt's
# password / full logout invalidates this — re-grab via DevTools Network tab.
DISCORD_TOKEN=MTUxMzMyNDc1MTYxMzg1Mzc5Ng.GjT6qU.A-VRmDoca1iT2ys54IGQJhTbudqteb69Klhjbk
# Guild (server) ID.
GUILD_ID=1474050123347525652
# ── Behavior ─────────────────────────────────────────────────────────────────
# DRY_RUN=true => export, parse, and stage metadata, but DO NOT enqueue any
# downloads. Inspect /storage1/labdata/romhacks/metadata first.
# DRY_RUN=false => write .crawljob files so JDownloader starts fetching.
DRY_RUN=true
# Seconds between harvest cycles (21600 = 6h).
POLL_INTERVAL=21600
# JDownloader in-container download root (maps to /storage1/labdata/romhacks/incoming).
JD_DOWNLOAD_ROOT=/output
# ── JDownloader noVNC UI (jlesage) ───────────────────────────────────────────
# Password for the web UI (reachable at http://valhalla:8998, or via Caddy
# internal-only once you add a site block). Set this before exposing it.
VNC_PASSWORD=