Spaces:
Paused
Paused
Update engine.py
Browse files
engine.py
CHANGED
|
@@ -1,98 +1,141 @@
|
|
| 1 |
-
import os, threading, shutil, json, time, glob, subprocess, urllib.parse
|
| 2 |
import warnings, requests, struct, socket
|
| 3 |
import libtorrent as lt
|
| 4 |
-
from flask import Flask, request, Response
|
| 5 |
from huggingface_hub import HfApi
|
| 6 |
|
| 7 |
-
# ---
|
| 8 |
-
|
| 9 |
-
|
|
|
|
| 10 |
|
| 11 |
app = Flask(__name__)
|
| 12 |
api = HfApi()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
active_vaults = set()
|
| 14 |
-
active_handles = {} # Registry for
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
-
#
|
|
|
|
|
|
|
| 17 |
settings = {
|
| 18 |
'listen_interfaces': '0.0.0.0:6881',
|
| 19 |
'enable_dht': True,
|
|
|
|
| 20 |
'connections_limit': 350,
|
|
|
|
|
|
|
| 21 |
'aio_threads': 16,
|
| 22 |
-
'
|
|
|
|
| 23 |
'peer_connect_timeout': 8,
|
| 24 |
'announce_to_all_trackers': True,
|
| 25 |
-
'enable_outgoing_utp': False, # Force TCP
|
| 26 |
'enable_incoming_utp': False,
|
| 27 |
}
|
| 28 |
ses = lt.session(settings)
|
| 29 |
|
|
|
|
|
|
|
|
|
|
| 30 |
def trigger_github_scout(u_id, info_hash):
|
| 31 |
"""Automatically deploys the GitHub Scout via API"""
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
# ⚠️ UPDATE THESE THREE VARIABLES ⚠️
|
| 35 |
-
gh_owner = "amrit123101"
|
| 36 |
-
gh_repo = "Scout"
|
| 37 |
-
workflow_filename = "peer_radar.yml"
|
| 38 |
-
|
| 39 |
-
if not gh_token:
|
| 40 |
print("⚠️ [API] GitHub Token missing. Scout deployment skipped.", flush=True)
|
| 41 |
return
|
| 42 |
|
| 43 |
-
url = f"https://api.github.com/repos/
|
| 44 |
headers = {
|
| 45 |
"Accept": "application/vnd.github+json",
|
| 46 |
-
"Authorization": f"Bearer {
|
| 47 |
"X-GitHub-Api-Version": "2022-11-28"
|
| 48 |
}
|
| 49 |
payload = {
|
| 50 |
-
"ref": "main",
|
| 51 |
"inputs": {
|
| 52 |
"info_hash": info_hash,
|
| 53 |
"u_id": u_id,
|
| 54 |
"hf_url": f"https://{os.environ.get('SPACE_HOST', 'amri12310-sys-b-engine.hf.space')}"
|
| 55 |
}
|
| 56 |
}
|
| 57 |
-
|
| 58 |
try:
|
| 59 |
r = requests.post(url, headers=headers, json=payload, timeout=5)
|
| 60 |
if r.status_code == 204:
|
| 61 |
print(f"🚀 [COMMAND] Automated GitHub Scout deployed for {u_id}", flush=True)
|
| 62 |
-
|
| 63 |
-
print(f"❌ [COMMAND] Scout deployment failed: {r.status_code} - {r.text}", flush=True)
|
| 64 |
-
except Exception as e:
|
| 65 |
-
print(f"❌ [COMMAND] Network error: {e}", flush=True)
|
| 66 |
|
| 67 |
@app.route('/inject_peers', methods=['POST'])
|
| 68 |
def inject_peers():
|
| 69 |
"""Endpoint for the GitHub Scout to push discovered UDP peers"""
|
| 70 |
data = request.get_json()
|
| 71 |
-
u_id = data.get('id')
|
| 72 |
-
peers = data.get('peers', [])
|
| 73 |
-
|
| 74 |
handle = active_handles.get(u_id)
|
| 75 |
-
if not handle:
|
| 76 |
-
return f"No active handle for {u_id}", 404
|
| 77 |
-
|
| 78 |
count = 0
|
| 79 |
for p in peers:
|
| 80 |
try:
|
| 81 |
handle.connect_peer((p['ip'], int(p['port'])), 0)
|
| 82 |
count += 1
|
| 83 |
except: continue
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 89 |
def do_archive(u_id, info_hash):
|
| 90 |
handle = None
|
| 91 |
try:
|
| 92 |
work_dir, upload_dir = f"/app/temp/{u_id}", f"/app/upload/{u_id}"
|
| 93 |
os.makedirs(work_dir, exist_ok=True); os.makedirs(upload_dir, exist_ok=True)
|
| 94 |
|
| 95 |
-
# Build
|
| 96 |
r_trackers = requests.get("https://raw.githubusercontent.com/ngosang/trackerslist/master/trackers_all.txt", timeout=5)
|
| 97 |
all_trackers = [t for t in r_trackers.text.split('\n') if t.startswith('http')]
|
| 98 |
magnet = f"magnet:?xt=urn:btih:{info_hash}" + "".join([f"&tr={urllib.parse.quote(t)}" for t in all_trackers])
|
|
@@ -102,7 +145,7 @@ def do_archive(u_id, info_hash):
|
|
| 102 |
handle = ses.add_torrent(atp)
|
| 103 |
active_handles[u_id] = handle
|
| 104 |
|
| 105 |
-
# Launch Scout
|
| 106 |
threading.Thread(target=trigger_github_scout, args=(u_id, info_hash)).start()
|
| 107 |
|
| 108 |
print(f"[ENGINE] Waiting for metadata for {u_id}...", flush=True)
|
|
@@ -111,26 +154,18 @@ def do_archive(u_id, info_hash):
|
|
| 111 |
while True:
|
| 112 |
s = handle.status()
|
| 113 |
if s.state in [lt.torrent_status.seeding, lt.torrent_status.finished]: break
|
| 114 |
-
|
| 115 |
done_gb = s.total_done / 1073741824
|
| 116 |
total_gb = s.total_wanted / 1073741824
|
| 117 |
speed = s.download_rate / 1000000
|
| 118 |
print(f"ID: {u_id[:8]} | {s.progress*100:.1f}% ({done_gb:.2f}/{total_gb:.2f} GB) | Seeds: {s.num_seeds} | Peers: {s.num_peers} | Speed: {speed:.2f} MB/s", flush=True)
|
| 119 |
time.sleep(10)
|
| 120 |
|
| 121 |
-
#
|
| 122 |
-
files = []
|
| 123 |
-
for root, _, filenames in os.walk(work_dir):
|
| 124 |
-
for f in filenames:
|
| 125 |
-
if f.endswith(('.mkv', '.mp4', '.avi', '.iso', '.zim', '.gguf')):
|
| 126 |
-
files.append(os.path.join(root, f))
|
| 127 |
-
|
| 128 |
-
if not files: raise Exception("No media found.")
|
| 129 |
video = max(files, key=os.path.getsize)
|
| 130 |
temp_path = os.path.join(upload_dir, "proc.tmp")
|
| 131 |
shutil.move(video, temp_path)
|
| 132 |
|
| 133 |
-
print(f"[ENGINE] Splitting file into 500MB chunks...", flush=True)
|
| 134 |
subprocess.run(['split', '-b', '500M', '-d', '--suffix-length=2', temp_path, f"{upload_dir}/chunk_"])
|
| 135 |
os.remove(temp_path)
|
| 136 |
|
|
@@ -140,11 +175,10 @@ def do_archive(u_id, info_hash):
|
|
| 140 |
with open(f"{upload_dir}/meta.json", "w") as f:
|
| 141 |
json.dump({"size": os.path.getsize(video), "chunks": len(raw_chunks), "chunk_size_mb": 500}, f)
|
| 142 |
|
| 143 |
-
# Sync to Vault
|
| 144 |
api.upload_folder(folder_path=upload_dir, path_in_repo=u_id, repo_id=REPO_ID, repo_type="dataset", token=TOKEN)
|
| 145 |
-
print(f"[ENGINE] ✅
|
| 146 |
|
| 147 |
-
except Exception as e: print(f"[ERROR] {e}", flush=True)
|
| 148 |
finally:
|
| 149 |
if handle: ses.remove_torrent(handle)
|
| 150 |
active_handles.pop(u_id, None)
|
|
@@ -154,7 +188,6 @@ def do_archive(u_id, info_hash):
|
|
| 154 |
@app.route('/archive')
|
| 155 |
def archive():
|
| 156 |
u_id, info_hash = request.args.get('id'), request.args.get('hash')
|
| 157 |
-
if not u_id or not info_hash: return "Missing params", 400
|
| 158 |
if u_id in active_vaults: return "Busy", 200
|
| 159 |
active_vaults.add(u_id)
|
| 160 |
threading.Thread(target=do_archive, args=(u_id, info_hash)).start()
|
|
|
|
| 1 |
+
import os, threading, shutil, json, time, glob, subprocess, urllib.parse, logging
|
| 2 |
import warnings, requests, struct, socket
|
| 3 |
import libtorrent as lt
|
| 4 |
+
from flask import Flask, request, Response, stream_with_context
|
| 5 |
from huggingface_hub import HfApi
|
| 6 |
|
| 7 |
+
# --- SILENCE WARNINGS ---
|
| 8 |
+
warnings.filterwarnings("ignore", category=DeprecationWarning)
|
| 9 |
+
log = logging.getLogger('werkzeug')
|
| 10 |
+
log.setLevel(logging.ERROR)
|
| 11 |
|
| 12 |
app = Flask(__name__)
|
| 13 |
api = HfApi()
|
| 14 |
+
|
| 15 |
+
REPO_ID = "amri12310/hf-vault-99281-xk9-p02-mlq8-r4x-v9928-zb-v2-alpha-sync"
|
| 16 |
+
TOKEN = os.environ.get("HF_TOKEN")
|
| 17 |
+
GH_TOKEN = os.environ.get("GH_PAT")
|
| 18 |
+
|
| 19 |
active_vaults = set()
|
| 20 |
+
active_handles = {} # Registry for peer injection
|
| 21 |
+
|
| 22 |
+
@app.after_request
|
| 23 |
+
def after_request(response):
|
| 24 |
+
response.headers.add('Access-Control-Allow-Origin', '*')
|
| 25 |
+
response.headers.add('Access-Control-Allow-Headers', 'Content-Type,Authorization,Range')
|
| 26 |
+
response.headers.add('Access-Control-Allow-Methods', 'GET,PUT,POST,DELETE,OPTIONS')
|
| 27 |
+
response.headers.add('Accept-Ranges', 'bytes')
|
| 28 |
+
return response
|
| 29 |
|
| 30 |
+
# ==========================================
|
| 31 |
+
# THE GOLD STANDARD: I/O UNLEASHED ENGINE
|
| 32 |
+
# ==========================================
|
| 33 |
settings = {
|
| 34 |
'listen_interfaces': '0.0.0.0:6881',
|
| 35 |
'enable_dht': True,
|
| 36 |
+
'dht_bootstrap_nodes': 'router.bittorrent.com:6881,router.utorrent.com:6881,dht.transmissionbt.com:6881',
|
| 37 |
'connections_limit': 350,
|
| 38 |
+
'unchoke_slots_limit': 150,
|
| 39 |
+
'active_downloads': 1,
|
| 40 |
'aio_threads': 16,
|
| 41 |
+
'cache_size': 65536,
|
| 42 |
+
'max_queued_disk_bytes': 512 * 1024 * 1024,
|
| 43 |
'peer_connect_timeout': 8,
|
| 44 |
'announce_to_all_trackers': True,
|
| 45 |
+
'enable_outgoing_utp': False, # Force TCP for Muscle
|
| 46 |
'enable_incoming_utp': False,
|
| 47 |
}
|
| 48 |
ses = lt.session(settings)
|
| 49 |
|
| 50 |
+
# ==========================================
|
| 51 |
+
# GITHUB SCOUT AUTOMATION
|
| 52 |
+
# ==========================================
|
| 53 |
def trigger_github_scout(u_id, info_hash):
|
| 54 |
"""Automatically deploys the GitHub Scout via API"""
|
| 55 |
+
if not GH_TOKEN:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
print("⚠️ [API] GitHub Token missing. Scout deployment skipped.", flush=True)
|
| 57 |
return
|
| 58 |
|
| 59 |
+
url = f"https://api.github.com/repos/amrit123101/Scout/actions/workflows/peer_radar.yml/dispatches"
|
| 60 |
headers = {
|
| 61 |
"Accept": "application/vnd.github+json",
|
| 62 |
+
"Authorization": f"Bearer {GH_TOKEN}",
|
| 63 |
"X-GitHub-Api-Version": "2022-11-28"
|
| 64 |
}
|
| 65 |
payload = {
|
| 66 |
+
"ref": "main",
|
| 67 |
"inputs": {
|
| 68 |
"info_hash": info_hash,
|
| 69 |
"u_id": u_id,
|
| 70 |
"hf_url": f"https://{os.environ.get('SPACE_HOST', 'amri12310-sys-b-engine.hf.space')}"
|
| 71 |
}
|
| 72 |
}
|
|
|
|
| 73 |
try:
|
| 74 |
r = requests.post(url, headers=headers, json=payload, timeout=5)
|
| 75 |
if r.status_code == 204:
|
| 76 |
print(f"🚀 [COMMAND] Automated GitHub Scout deployed for {u_id}", flush=True)
|
| 77 |
+
except: pass
|
|
|
|
|
|
|
|
|
|
| 78 |
|
| 79 |
@app.route('/inject_peers', methods=['POST'])
|
| 80 |
def inject_peers():
|
| 81 |
"""Endpoint for the GitHub Scout to push discovered UDP peers"""
|
| 82 |
data = request.get_json()
|
| 83 |
+
u_id, peers = data.get('id'), data.get('peers', [])
|
|
|
|
|
|
|
| 84 |
handle = active_handles.get(u_id)
|
| 85 |
+
if not handle: return "No handle", 404
|
|
|
|
|
|
|
| 86 |
count = 0
|
| 87 |
for p in peers:
|
| 88 |
try:
|
| 89 |
handle.connect_peer((p['ip'], int(p['port'])), 0)
|
| 90 |
count += 1
|
| 91 |
except: continue
|
| 92 |
+
if count > 0: print(f"🎯 [SCOUT] Injected {count} peers into {u_id}", flush=True)
|
| 93 |
+
return "OK", 200
|
| 94 |
+
|
| 95 |
+
# ==========================================
|
| 96 |
+
# STREAMING FRONTEND (STREMIO COMPATIBLE)
|
| 97 |
+
# ==========================================
|
| 98 |
+
@app.route('/stream/<u_id>')
|
| 99 |
+
def stream_vault(u_id):
|
| 100 |
+
auth = {"Authorization": f"Bearer {TOKEN}"}
|
| 101 |
+
meta_url = f"https://huggingface.co/datasets/{REPO_ID}/resolve/main/{u_id}/meta.json"
|
| 102 |
+
try:
|
| 103 |
+
meta = requests.get(meta_url, headers=auth, timeout=10).json()
|
| 104 |
+
total_size = meta.get('size', 0)
|
| 105 |
+
num_chunks, c_size_mb = meta.get('chunks', 0), meta.get('chunk_size_mb', 500)
|
| 106 |
+
except: return "Metadata not found", 404
|
| 107 |
+
|
| 108 |
+
range_header = request.headers.get('Range', None)
|
| 109 |
+
byte_start = 0
|
| 110 |
+
if range_header:
|
| 111 |
+
try: byte_start = int(range_header.replace('bytes=', '').split('-')[0])
|
| 112 |
+
except: pass
|
| 113 |
|
| 114 |
+
def generate():
|
| 115 |
+
chunk_size_bytes = c_size_mb * 1024 * 1024
|
| 116 |
+
start_chunk_idx = byte_start // chunk_size_bytes
|
| 117 |
+
for i in range(start_chunk_idx, num_chunks):
|
| 118 |
+
chunk_url = f"https://huggingface.co/datasets/{REPO_ID}/resolve/main/{u_id}/{i}.bin"
|
| 119 |
+
headers = auth.copy()
|
| 120 |
+
if i == start_chunk_idx and byte_start % chunk_size_bytes > 0:
|
| 121 |
+
headers["Range"] = f"bytes={byte_start % chunk_size_bytes}-"
|
| 122 |
+
with requests.get(chunk_url, headers=headers, stream=True) as r:
|
| 123 |
+
for data in r.iter_content(chunk_size=1024*512): yield data
|
| 124 |
+
|
| 125 |
+
headers = {"Content-Length": str(total_size - byte_start), "Content-Type": "video/x-matroska", "Accept-Ranges": "bytes"}
|
| 126 |
+
if range_header: headers["Content-Range"] = f"bytes {byte_start}-{total_size-1}/{total_size}"
|
| 127 |
+
return Response(stream_with_context(generate()), status=206 if range_header else 200, headers=headers)
|
| 128 |
+
|
| 129 |
+
# ==========================================
|
| 130 |
+
# CORE ENGINE LOGIC
|
| 131 |
+
# ==========================================
|
| 132 |
def do_archive(u_id, info_hash):
|
| 133 |
handle = None
|
| 134 |
try:
|
| 135 |
work_dir, upload_dir = f"/app/temp/{u_id}", f"/app/upload/{u_id}"
|
| 136 |
os.makedirs(work_dir, exist_ok=True); os.makedirs(upload_dir, exist_ok=True)
|
| 137 |
|
| 138 |
+
# Build Tracker Bomb
|
| 139 |
r_trackers = requests.get("https://raw.githubusercontent.com/ngosang/trackerslist/master/trackers_all.txt", timeout=5)
|
| 140 |
all_trackers = [t for t in r_trackers.text.split('\n') if t.startswith('http')]
|
| 141 |
magnet = f"magnet:?xt=urn:btih:{info_hash}" + "".join([f"&tr={urllib.parse.quote(t)}" for t in all_trackers])
|
|
|
|
| 145 |
handle = ses.add_torrent(atp)
|
| 146 |
active_handles[u_id] = handle
|
| 147 |
|
| 148 |
+
# Launch Scout
|
| 149 |
threading.Thread(target=trigger_github_scout, args=(u_id, info_hash)).start()
|
| 150 |
|
| 151 |
print(f"[ENGINE] Waiting for metadata for {u_id}...", flush=True)
|
|
|
|
| 154 |
while True:
|
| 155 |
s = handle.status()
|
| 156 |
if s.state in [lt.torrent_status.seeding, lt.torrent_status.finished]: break
|
|
|
|
| 157 |
done_gb = s.total_done / 1073741824
|
| 158 |
total_gb = s.total_wanted / 1073741824
|
| 159 |
speed = s.download_rate / 1000000
|
| 160 |
print(f"ID: {u_id[:8]} | {s.progress*100:.1f}% ({done_gb:.2f}/{total_gb:.2f} GB) | Seeds: {s.num_seeds} | Peers: {s.num_peers} | Speed: {speed:.2f} MB/s", flush=True)
|
| 161 |
time.sleep(10)
|
| 162 |
|
| 163 |
+
# Splitting & Syncing
|
| 164 |
+
files = [os.path.join(r, f) for r, d, fs in os.walk(work_dir) for f in fs if f.endswith(('.mkv', '.mp4', '.avi', '.iso', '.zim', '.gguf'))]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 165 |
video = max(files, key=os.path.getsize)
|
| 166 |
temp_path = os.path.join(upload_dir, "proc.tmp")
|
| 167 |
shutil.move(video, temp_path)
|
| 168 |
|
|
|
|
| 169 |
subprocess.run(['split', '-b', '500M', '-d', '--suffix-length=2', temp_path, f"{upload_dir}/chunk_"])
|
| 170 |
os.remove(temp_path)
|
| 171 |
|
|
|
|
| 175 |
with open(f"{upload_dir}/meta.json", "w") as f:
|
| 176 |
json.dump({"size": os.path.getsize(video), "chunks": len(raw_chunks), "chunk_size_mb": 500}, f)
|
| 177 |
|
|
|
|
| 178 |
api.upload_folder(folder_path=upload_dir, path_in_repo=u_id, repo_id=REPO_ID, repo_type="dataset", token=TOKEN)
|
| 179 |
+
print(f"[ENGINE] ✅ Vaulting complete for {u_id}", flush=True)
|
| 180 |
|
| 181 |
+
except Exception as e: print(f"[CRITICAL ERROR] {e}", flush=True)
|
| 182 |
finally:
|
| 183 |
if handle: ses.remove_torrent(handle)
|
| 184 |
active_handles.pop(u_id, None)
|
|
|
|
| 188 |
@app.route('/archive')
|
| 189 |
def archive():
|
| 190 |
u_id, info_hash = request.args.get('id'), request.args.get('hash')
|
|
|
|
| 191 |
if u_id in active_vaults: return "Busy", 200
|
| 192 |
active_vaults.add(u_id)
|
| 193 |
threading.Thread(target=do_archive, args=(u_id, info_hash)).start()
|