HDR / app.py
THED2's picture
Update app.py
38251d0 verified
Raw History Blame Contribute Delete
24.3 kB
import asyncio
import json
import os
import re
import time
import uuid
import shutil
import gc
import zipfile
import threading
import urllib3
from urllib3.util import Timeout
from fastapi import FastAPI, BackgroundTasks, Request, File, UploadFile
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse, StreamingResponse
from pydantic import BaseModel
import aiohttp
app = FastAPI()
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
tasks = {}
DATA_DIR = "/data" # HF Space me /data mounted volume hai
os.makedirs(DATA_DIR, exist_ok=True)
STATE_FILE = os.path.join(DATA_DIR, ".tasks_state.json")
FLUSH_EVERY = 256 * 1024 * 1024 # 256MB ke baad cache drop
# HF free Space: 16GB RAM + 50GB ephemeral disk. 50GB upar container EVICT.
MAX_TOTAL_STORAGE = int(os.environ.get("MAX_TOTAL_STORAGE", 45 * 1024 ** 3))
DISK_MARGIN = 2 * 1024 ** 3
CHUNK = 4 * 1024 * 1024
CONNECT_TIMEOUT = 30
READ_TIMEOUT = 180
MAX_ATTEMPTS = 6
SEGMENTS = 8
MIN_SIZE_FOR_PARALLEL = 32 * 1024 * 1024
_http = urllib3.PoolManager(maxsize=32, block=True)
class DownloadRequest(BaseModel):
url: str
def clean_filename(url: str) -> str:
path = url.split("?")[0]
filename = path.split("/")[-1]
if not filename or "." not in filename:
return "downloaded_file.bin"
return filename
def _filename_from_cd(cd: str):
if not cd or "filename=" not in cd:
return None
fname = re.findall('filename="([^"]+)"', cd)
if not fname:
fname = re.findall("filename=([^;]+)", cd)
return fname[0] if fname else None
def _drop_cache(fd):
"""Page cache drop -> container ka memory limit safe, OOM-kill se bachav."""
try:
os.fsync(fd)
os.posix_fadvise(fd, 0, 0, os.POSIX_FADV_DONTNEED)
except Exception:
pass
def save_state():
"""Task state disk par -> container restart hone par bhi task list bachi rahe."""
try:
clean = {}
for k, v in tasks.items():
clean[k] = {kk: vv for kk, vv in v.items() if not kk.startswith("_")}
tmp = STATE_FILE + ".tmp"
with open(tmp, "w") as f:
json.dump(clean, f)
os.replace(tmp, STATE_FILE)
except Exception:
pass
def load_state():
try:
with open(STATE_FILE) as f:
return json.load(f)
except Exception:
return {}
def probe_url(url: str):
"""Size, range-support aur filename nikalo (download se pehle)."""
info = {"total": 0, "ranges": False, "filename": None, "status": None}
sizes = [] # sab observed sizes me se SABSE BADA sachcha mana jayega,
# taaki koi chhota error/rate-limit page size na ban jaye
try:
r = _http.request("HEAD", url, preload_content=False,
timeout=Timeout(connect=CONNECT_TIMEOUT, read=60), retries=2)
info["status"] = r.status
cl = r.headers.get("Content-Length")
if cl:
sizes.append(int(cl))
ar = r.headers.get("Accept-Ranges")
if ar and "bytes" in ar.lower():
info["ranges"] = True
info["filename"] = _filename_from_cd(r.headers.get("Content-Disposition"))
r.release_conn()
except Exception:
pass
try:
r = _http.request("GET", url, headers={"Range": "bytes=0-0"}, preload_content=False,
timeout=Timeout(connect=CONNECT_TIMEOUT, read=60), retries=2)
info["status"] = r.status
cr = r.headers.get("Content-Range")
if r.status == 206 and cr and "/" in cr:
info["ranges"] = True
try:
sizes.append(int(cr.split("/")[-1]))
except Exception:
pass
else:
info["ranges"] = False # server Range ignore karta hai
cl = r.headers.get("Content-Length")
if cl:
sizes.append(int(cl))
if not info["filename"]:
info["filename"] = _filename_from_cd(r.headers.get("Content-Disposition"))
r.release_conn()
except Exception as e:
info["probe_error"] = str(e)
info["total"] = max(sizes) if sizes else 0
return info
def _download_range(url, path, start, end, state, idx, cancel):
"""start..end (inclusive) bytes download, apne aap resume + retry."""
pos = start
attempt = 0
fd = os.open(path, os.O_WRONLY)
try:
while pos <= end and attempt < MAX_ATTEMPTS and not cancel.is_set():
try:
r = _http.request(
"GET", url,
headers={"Range": f"bytes={pos}-{end}"},
preload_content=False,
retries=False,
timeout=Timeout(connect=CONNECT_TIMEOUT, read=READ_TIMEOUT),
)
if r.status != 206: # server range ignore karta hai
r.release_conn()
return "no_resume"
last_flush = pos
for chunk in r.stream(CHUNK):
if not chunk or cancel.is_set():
break
os.pwrite(fd, chunk, pos)
pos += len(chunk)
state[idx] = pos - start # is segment ne abhi tak kitne bytes liye
if pos - last_flush >= FLUSH_EVERY:
_drop_cache(fd)
last_flush = pos
_drop_cache(fd)
r.release_conn()
if pos > end:
return "done"
attempt += 1
time.sleep(min(2 * attempt, 20))
except Exception:
attempt += 1
time.sleep(min(2 * attempt, 20))
return "done" if pos > end else "failed"
finally:
os.close(fd)
def _download_parallel(url, path, total, state, cancel):
seg = total // SEGMENTS
bounds = []
for i in range(SEGMENTS):
s = i * seg
e = (total - 1) if i == SEGMENTS - 1 else (s + seg - 1)
bounds.append((s, e))
state[:] = [b[0] for b in bounds]
threads = []
results = [None] * SEGMENTS
for i, (s, e) in enumerate(bounds):
t = threading.Thread(
target=lambda i=i, s=s, e=e: results.__setitem__(
i, _download_range(url, path, s, e, state, i, cancel)))
t.daemon = True
t.start()
threads.append(t)
for t in threads:
t.join()
return results
def _download_single(url, path, state, cancel):
"""Non-resumable servers ke liye (jaise video-downloads.googleusercontent.com)."""
attempt = 0
while attempt < MAX_ATTEMPTS and not cancel.is_set():
pos = 0
try:
r = _http.request("GET", url, preload_content=False, retries=False,
timeout=Timeout(connect=CONNECT_TIMEOUT, read=READ_TIMEOUT))
cl = r.headers.get("Content-Length")
expected = tasks[state["task_id"]].get("total_size", 0) or 0
if cl and expected and int(cl) != expected:
r.release_conn()
raise RuntimeError(
f"SOURCE SIZE MISMATCH: server ne ab {int(cl)} bytes bheje, "
f"pehle {expected} tha (rate-limit / error page / expired link?)")
if cl and not expected:
tasks[state["task_id"]]["total_size"] = int(cl)
with open(path, "r+b") as f:
f.seek(0)
f.truncate(0) # pichhle attempt ka kachra saaf
last_flush = 0
for chunk in r.stream(CHUNK):
if not chunk or cancel.is_set():
break
f.write(chunk)
pos += len(chunk)
state["pos"] = pos
if pos - last_flush >= FLUSH_EVERY:
_drop_cache(f.fileno())
last_flush = pos
_drop_cache(f.fileno())
r.release_conn()
total = tasks[state["task_id"]].get("total_size", 0) or 0
if total == 0 or pos >= total:
return "done"
attempt += 1
except Exception:
attempt += 1
time.sleep(min(2 * attempt, 20))
return "failed"
if total:
used_here = 0
for fn in os.listdir(DATA_DIR):
fp = os.path.join(DATA_DIR, fn)
try:
if os.path.isfile(fp):
used_here += os.path.getsize(fp)
except Exception:
pass
if used_here + total > MAX_TOTAL_STORAGE:
raise RuntimeError(
f"FILE TOO BIG FOR THIS SERVER: file {total/1e9:.2f}GB + stored "
f"{used_here/1e9:.2f}GB = {(used_here+total)/1e9:.2f}GB > limit "
f"{MAX_TOTAL_STORAGE/1e9:.1f}GB. HF free Space me 50GB ephemeral "
f"disk hai, uske upar container EVICT ho jata hai. Fix: Space "
f"Settings me persistent storage on karo ya chhoti file use karo.")
# ---- DISK CHECK ----
if total:
free = shutil.disk_usage(DATA_DIR).free
if free < total + DISK_MARGIN:
raise RuntimeError(f"Not enough disk: need {total/1e9:.1f}GB, free {free/1e9:.1f}GB")
# ---- progress ticker (har second speed + downloaded update) ----
state = {"task_id": task_id, "pos": 0}
seg_state = [0]
tasks[task_id]["_state"] = state
stop = threading.Event()
def ticker():
last = 0
last_t = time.time()
while not stop.is_set():
time.sleep(1.0)
try:
done = sum(seg_state) if tasks[task_id].get("resumable") else state.get("pos", 0)
now = time.time()
dt = now - last_t
if dt > 0:
sp = max(0.0, (done - last) / dt)
tasks[task_id]["speed"] = sp
tsize = tasks[task_id].get("total_size", 0) or 0
if sp > 0 and tsize:
tasks[task_id]["eta_seconds"] = int((tsize - done) / sp)
tasks[task_id]["downloaded"] = done
last, last_t = done, now
save_state()
except Exception:
pass
threading.Thread(target=ticker, daemon=True).start()
# ---- DOWNLOAD ----
result = None
with open(file_path, "wb"):
pass # file pehle banao, segments isi me pwrite karenge
if info.get("ranges") and total >= MIN_SIZE_FOR_PARALLEL:
seg_state = [0] * SEGMENTS
tasks[task_id]["_state"] = seg_state
res = _download_parallel(url, file_path, total, seg_state, cancel)
if "no_resume" in res:
tasks[task_id]["resumable"] = False
with open(file_path, "r+b") as f:
f.truncate(0)
result = _download_single(url, file_path, state, cancel)
elif "failed" in res:
result = "failed"
else:
result = "done"
else:
result = _download_single(url, file_path, state, cancel)
stop.set()
# ---- VERIFY (SABSE IMPORTANT FIX) ----
actual = os.path.getsize(file_path)
tasks[task_id]["downloaded"] = actual
if total and actual != total:
tasks[task_id]["status"] = "error"
tasks[task_id]["error"] = (f"INCOMPLETE: got {actual} of {total} bytes "
f"(missing {total-actual}). Server ne beech me "
f"connection band kar diya.")
return
if result != "done":
tasks[task_id]["status"] = "error"
tasks[task_id]["error"] = "Download failed after maximum retries"
return
tasks[task_id]["status"] = "completed"
tasks[task_id]["speed"] = 0.0
tasks[task_id]["verified"] = True
except Exception as e:
tasks[task_id]["status"] = "error"
tasks[task_id]["error"] = str(e)
finally:
save_state()
gc.collect()
async def download_file(task_id: str, url: str):
loop = asyncio.get_event_loop()
await loop.run_in_executor(None, sync_download_worker, task_id, url)
@app.post("/start_download")
async def start_download(req: DownloadRequest, background_tasks: BackgroundTasks):
task_id = str(uuid.uuid4())
filename = clean_filename(req.url)
tasks[task_id] = {
"task_id": task_id,
"url": req.url,
"status": "downloading",
"total_size": 0,
"downloaded": 0,
"speed": 0.0,
"file_path": os.path.join(DATA_DIR, f"{task_id}_{filename}"),
"original_filename": filename,
"resumable": None,
"timestamp": time.time(),
}
background_tasks.add_task(download_file, task_id, req.url)
return {"task_id": task_id}
@app.get("/status/{task_id}")
async def get_status(task_id: str):
if task_id not in tasks:
return {"error": "Task not found"}
t = dict(tasks[task_id])
t.pop("_state", None)
t.pop("_cancel", None)
return t
@app.get("/download/{task_id}")
async def download(task_id: str):
if task_id not in tasks or tasks[task_id]["status"] != "completed":
return {"error": "File not ready"}
return FileResponse(tasks[task_id]["file_path"],
filename=tasks[task_id].get("original_filename"))
@app.get("/stream/{task_id}")
async def stream(task_id: str, request: Request):
if task_id not in tasks or tasks[task_id]["status"] != "completed":
return {"error": "File not ready"}
file_path = tasks[task_id]["file_path"]
file_size = os.path.getsize(file_path)
range_header = request.headers.get("Range")
def file_iterator(start, length):
with open(file_path, "rb") as f:
f.seek(start)
while length > 0:
data = f.read(min(1024 * 1024, length)) # length se zyada NAHI bhejna,
if not data: # warna Content-Length mismatch
break # -> client ko "bytes missing"
yield data
length -= len(data)
if range_header:
byte1, byte2 = 0, None
parts = range_header.replace("bytes=", "").split("-")
if parts[0]:
byte1 = int(parts[0])
if len(parts) > 1 and parts[1]:
byte2 = int(parts[1])
length = file_size - byte1 if byte2 is None else byte2 + 1 - byte1
headers = {
"Content-Range": f"bytes {byte1}-{byte1+length-1}/{file_size}",
"Accept-Ranges": "bytes",
"Content-Length": str(length),
}
return StreamingResponse(file_iterator(byte1, length), status_code=206, headers=headers)
headers = {"Accept-Ranges": "bytes", "Content-Length": str(file_size)}
return StreamingResponse(file_iterator(0, file_size), headers=headers,
media_type="application/octet-stream")
@app.get("/storage")
async def get_storage():
total, used, free = shutil.disk_usage(DATA_DIR)
stored = 0
try:
for fn in os.listdir(DATA_DIR):
fp = os.path.join(DATA_DIR, fn)
if os.path.isfile(fp):
stored += os.path.getsize(fp)
except Exception:
pass
return {"total": total, "used": used, "free": free,
"our_files_bytes": stored,
"our_budget_bytes": MAX_TOTAL_STORAGE,
"budget_left_bytes": max(0, MAX_TOTAL_STORAGE - stored)}
@app.get("/server_status")
async def get_server_status():
status = {}
try:
with open("/proc/meminfo") as f:
mem = {}
for line in f:
if ":" in line:
k, v = line.split(":", 1)
parts = v.strip().split()
if parts and parts[0].isdigit():
mem[k.strip()] = int(parts[0]) * 1024
total_ram = mem.get("MemTotal", 0)
available_ram = mem.get("MemAvailable", 0)
used_ram = total_ram - available_ram
status["ram_total"] = total_ram
status["ram_used"] = used_ram
status["ram_percent"] = round((used_ram / total_ram * 100) if total_ram else 0, 1)
except Exception:
status.update({"ram_total": 0, "ram_used": 0, "ram_percent": 0})
try:
total_d, used_d, free_d = shutil.disk_usage(DATA_DIR)
status["storage_total"] = total_d
status["storage_used"] = used_d
status["storage_free"] = free_d
status["storage_percent"] = round((used_d / total_d * 100) if total_d else 0, 1)
except Exception:
pass
status["temperature"] = -1
status["ping_ms"] = -1
status["server_online"] = True
status["active_tasks"] = len(tasks)
status["uptime"] = round(time.time(), 0)
return status
@app.post("/delete_all")
async def delete_all():
for filename in os.listdir(DATA_DIR):
filepath = os.path.join(DATA_DIR, filename)
try:
if os.path.isfile(filepath):
os.remove(filepath)
elif os.path.isdir(filepath):
shutil.rmtree(filepath)
except Exception as e:
print(f"Failed to delete {filepath}: {e}")
tasks.clear()
gc.collect()
return {"status": "success"}
@app.get("/history")
async def get_history():
out = []
for t in tasks.values():
out.append({k: v for k, v in t.items() if not k.startswith("_")})
return {"history": out}
@app.post("/upload")
async def upload_file(file: UploadFile = File(...)):
task_id = str(uuid.uuid4())
original_filename = file.filename if file.filename else "uploaded_file.bin"
file_path = os.path.join(DATA_DIR, f"{task_id}_{original_filename}")
with open(file_path, "wb") as buffer:
while True:
chunk = await file.read(CHUNK)
if not chunk:
break
buffer.write(chunk)
await file.close()
file_size = os.path.getsize(file_path)
tasks[task_id] = {
"task_id": task_id,
"url": "local_upload",
"status": "completed",
"total_size": file_size,
"downloaded": file_size,
"speed": 0.0,
"file_path": file_path,
"original_filename": original_filename,
"timestamp": time.time(),
}
save_state()
return {"task_id": task_id}
async def file_sender(file_path, chunk_size=CHUNK):
with open(file_path, "rb") as f:
while True:
chunk = f.read(chunk_size)
if not chunk:
break
yield chunk
@app.post("/start_gofile_transfer")
async def start_gofile_transfer(req: DownloadRequest, background_tasks: BackgroundTasks):
task_id = str(uuid.uuid4())
filename = clean_filename(req.url)
tasks[task_id] = {
"task_id": task_id,
"url": req.url,
"status": "downloading",
"total_size": 0,
"downloaded": 0,
"speed": 0.0,
"file_path": os.path.join(DATA_DIR, f"{task_id}_{filename}"),
"original_filename": filename,
"gofile_url": None,
"timestamp": time.time(),
}
background_tasks.add_task(process_gofile_transfer, task_id, req.url)
return {"task_id": task_id}
async def process_gofile_transfer(task_id: str, url: str):
try:
loop = asyncio.get_event_loop()
await loop.run_in_executor(None, sync_download_worker, task_id, url)
if tasks[task_id].get("status") != "completed":
raise Exception(tasks[task_id].get("error", "download failed"))
file_path = tasks[task_id]["file_path"]
tasks[task_id]["status"] = "uploading_to_gofile"
async with aiohttp.ClientSession() as session:
async with session.get("https://api.gofile.io/servers") as resp:
servers_data = await resp.json()
server = servers_data["data"]["servers"][0]["name"]
upload_url = f"https://{server}.gofile.io/uploadFile"
data = aiohttp.FormData()
data.add_field("file", file_sender(file_path),
filename=tasks[task_id]["original_filename"])
async with session.post(upload_url, data=data) as upload_resp:
upload_result = await upload_resp.json()
if upload_result["status"] == "ok":
tasks[task_id]["gofile_url"] = upload_result["data"]["downloadPage"]
tasks[task_id]["status"] = "completed"
else:
raise Exception("GoFile upload failed")
try:
os.remove(file_path)
except Exception:
pass
except Exception as e:
tasks[task_id]["status"] = "error"
tasks[task_id]["error"] = str(e)
finally:
save_state()
gc.collect()
@app.get("/list_zip/{task_id}")
async def list_zip_contents(task_id: str):
if task_id not in tasks:
return {"error": "Task not found"}
file_path = tasks[task_id]["file_path"]
if not os.path.exists(file_path):
return {"error": "File does not exist on server"}
if not zipfile.is_zipfile(file_path):
return {"error": "This file is not a valid ZIP archive"}
try:
with zipfile.ZipFile(file_path, "r") as z:
file_list = [{"filename": i.filename,
"file_size_mb": round(i.file_size / (1024 * 1024), 2),
"is_dir": i.is_dir()} for i in z.infolist()]
return {"task_id": task_id, "total_files": len(file_list), "files": file_list}
except Exception as e:
return {"error": f"Failed to read ZIP: {str(e)}"}
@app.post("/extract_zip/{task_id}")
async def extract_zip_file(task_id: str, background_tasks: BackgroundTasks):
if task_id not in tasks:
return {"error": "Task not found"}
file_path = tasks[task_id]["file_path"]
if not os.path.exists(file_path):
return {"error": "File does not exist on server"}
if not zipfile.is_zipfile(file_path):
return {"error": "This file is not a valid ZIP archive"}
extract_folder = os.path.join(DATA_DIR, f"extracted_{task_id}")
os.makedirs(extract_folder, exist_ok=True)
tasks[task_id]["status"] = "extracting"
tasks[task_id]["extract_path"] = extract_folder
background_tasks.add_task(process_zip_extraction, task_id, file_path, extract_folder)
return {"status": "extraction_started", "task_id": task_id, "target_folder": extract_folder}
def process_zip_extraction(task_id: str, zip_path: str, target_dir: str):
try:
with zipfile.ZipFile(zip_path, "r") as z:
z.extractall(path=target_dir)
tasks[task_id]["status"] = "extracted"
except Exception as e:
tasks[task_id]["status"] = "extraction_error"
tasks[task_id]["error"] = str(e)
finally:
save_state()
@app.on_event("startup")
async def startup_event():
async def cleanup_old_files():
while True:
try:
now = time.time()
for filename in os.listdir(DATA_DIR):
filepath = os.path.join(DATA_DIR, filename)
if os.path.isfile(filepath) and now - os.path.getmtime(filepath) > 86400:
os.remove(filepath)
for tid, tinfo in list(tasks.items()):
if tinfo.get("file_path") == filepath:
del tasks[tid]
except Exception as e:
print(f"Cleanup error: {e}")
await asyncio.sleep(3600)
asyncio.create_task(cleanup_old_files())
if __name__ == "__main__":
import uvicorn
uvicorn.run(app, host="0.0.0.0", port=7860)