Spaces:
Sleeping
Sleeping
Download lib/cloudinary_service.py from thesis-sd3a/digital-art-api: direct link, hf CLI and curl.
- Browser
- Download file 3.83 kB
-
https://huggingface.co/spaces/thesis-sd3a/digital-art-api/resolve/main/lib/cloudinary_service.py
- Command line
-
hf download hf://spaces/thesis-sd3a/digital-art-api/lib/cloudinary_service.py
-
curl -L -o cloudinary_service.py https://huggingface.co/spaces/thesis-sd3a/digital-art-api/resolve/main/lib/cloudinary_service.py
3.83 kB
| import cloudinary | |
| import cloudinary.uploader | |
| import logging | |
| import os | |
| import time | |
| import requests | |
| from dotenv import load_dotenv | |
| from fastapi import HTTPException, UploadFile, File | |
| load_dotenv() | |
| logger = logging.getLogger(__name__) | |
| cloudinary.config( | |
| cloud_name=os.getenv("CLOUDINARY_NAME"), | |
| api_key=os.getenv("CLOUDINARY_KEY"), | |
| api_secret=os.getenv("CLOUDINARY_SECRET"), | |
| ) | |
| def cloudinary_uploader(image_file: UploadFile = File(...)): | |
| try: | |
| image_file.file.seek(0) | |
| response = cloudinary.uploader.upload( | |
| image_file.file, | |
| resource_type="image", | |
| folder="temp_plag_scans", | |
| # Serper's crawler must fetch this URL without credentials, so the | |
| # temp upload must be explicitly public with a unique name. | |
| access_mode="public", | |
| use_filename=True, | |
| unique_filename=True, | |
| overwrite=False, | |
| ) | |
| return response | |
| except Exception as e: | |
| raise HTTPException(status_code=400, detail=str(e)) | |
| def wait_for_url_ready(url, timeout=15.0, initial_delay=1.0, max_delay=4.0): | |
| """Poll a Cloudinary delivery URL until it is publicly fetchable. | |
| Serper fetches ``secure_url`` immediately after upload; if the CDN has | |
| not propagated yet it sees a 404/403 and returns HTTP 200 with an empty | |
| ``organic`` list. This gate distinguishes that case from a genuine | |
| "no online match" so callers can degrade explicitly instead of | |
| reporting a silent clean negative. | |
| Returns (ready, info) where info carries attempts/last_status/elapsed_ms | |
| for ``web_diagnostics``. Never raises. | |
| """ | |
| start = time.monotonic() | |
| attempts = 0 | |
| last_status = None | |
| delay = initial_delay | |
| deadline = start + max(timeout, 0.1) | |
| try: | |
| while True: | |
| attempts += 1 | |
| try: | |
| # HEAD is cheap; some CDNs block it, so fall back to a | |
| # 1-byte ranged GET on 403/405. | |
| resp = requests.head(url, timeout=8, allow_redirects=True) | |
| last_status = resp.status_code | |
| content_type = (resp.headers.get("Content-Type") or "").lower() | |
| if resp.status_code == 200 and ( | |
| not content_type or "image" in content_type | |
| ): | |
| break | |
| if resp.status_code in (403, 405): | |
| g = requests.get( | |
| url, | |
| timeout=8, | |
| headers={"Range": "bytes=0-0"}, | |
| stream=True, | |
| ) | |
| last_status = g.status_code | |
| g.close() | |
| if g.status_code in (200, 206): | |
| break | |
| except Exception as e: | |
| logger.debug("readiness poll attempt %d failed: %s", attempts, e) | |
| now = time.monotonic() | |
| if now >= deadline: | |
| break | |
| sleep_for = min(delay, max(0.0, deadline - now)) | |
| if sleep_for > 0: | |
| time.sleep(sleep_for) | |
| delay = min(delay * 1.5, max_delay) | |
| except Exception as e: # never let readiness break the pipeline | |
| logger.debug("readiness poll error: %s", e) | |
| elapsed_ms = int((time.monotonic() - start) * 1000) | |
| ready = last_status in (200, 206) | |
| return ready, { | |
| "attempts": attempts, | |
| "last_status": last_status, | |
| "elapsed_ms": elapsed_ms, | |
| } | |
| def delete_image(public_id): | |
| try: | |
| result = cloudinary.uploader.destroy( | |
| public_id, | |
| invalidate=True, | |
| ) | |
| if result.get("result") == "ok": | |
| return True | |
| else: | |
| return False | |
| except Exception as e: | |
| raise HTTPException(status_code=400, detail=str(e)) |