Spaces:
Runtime error
Runtime error
Update server.py
Browse files
server.py
CHANGED
|
@@ -9,13 +9,23 @@ from fastapi.middleware.cors import CORSMiddleware
|
|
| 9 |
from fastapi.staticfiles import StaticFiles
|
| 10 |
from huggingface_hub import snapshot_download
|
| 11 |
|
| 12 |
-
#
|
| 13 |
logging.basicConfig(level=logging.INFO)
|
| 14 |
logger = logging.getLogger(__name__)
|
| 15 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
app = FastAPI()
|
| 17 |
|
| 18 |
-
# Enable CORS for Hugging Face environment
|
| 19 |
app.add_middleware(
|
| 20 |
CORSMiddleware,
|
| 21 |
allow_origins=["*"],
|
|
@@ -24,19 +34,16 @@ app.add_middleware(
|
|
| 24 |
allow_headers=["*"],
|
| 25 |
)
|
| 26 |
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
DB_TARGET = "orbiitt.db"
|
| 30 |
|
| 31 |
-
#
|
| 32 |
engine = None
|
| 33 |
-
loading_status = "System
|
| 34 |
|
| 35 |
def normalize_filename(name):
|
| 36 |
-
"""Handles smart quotes
|
| 37 |
-
# Convert curly quotes/dashes to standard ASCII
|
| 38 |
name = name.replace('β', "'").replace('β', "'").replace('β', '"').replace('β', '"').replace('β', '-')
|
| 39 |
-
# Decompose Unicode characters and strip non-ASCII
|
| 40 |
name = unicodedata.normalize('NFKD', name).encode('ascii', 'ignore').decode('ascii')
|
| 41 |
return name.strip()
|
| 42 |
|
|
@@ -45,11 +52,14 @@ def setup_environment():
|
|
| 45 |
global engine, loading_status
|
| 46 |
token = os.environ.get("HF_TOKEN")
|
| 47 |
|
| 48 |
-
|
| 49 |
-
os.
|
|
|
|
|
|
|
| 50 |
|
| 51 |
try:
|
| 52 |
loading_status = "Syncing Orbit Assets..."
|
|
|
|
| 53 |
snapshot_download(repo_id=DATASET_REPO, repo_type="dataset", token=token, local_dir=".")
|
| 54 |
|
| 55 |
if os.path.exists("orbiitt_db.zip"):
|
|
@@ -62,12 +72,13 @@ def setup_environment():
|
|
| 62 |
if f.lower() in ["orbiitt_db", "orbiitt.db"] or (f.endswith(".db") and not f.startswith(".")):
|
| 63 |
shutil.copy(src, f"./{DB_TARGET}")
|
| 64 |
elif f.lower().endswith(('.png', '.jpg', '.jpeg', '.webp')):
|
| 65 |
-
#
|
| 66 |
clean_name = normalize_filename(urllib.parse.unquote(f))
|
| 67 |
shutil.copy(src, os.path.join(IMAGE_DIR, clean_name))
|
| 68 |
shutil.rmtree("temp_extract")
|
| 69 |
|
| 70 |
-
loading_status = "
|
|
|
|
| 71 |
try:
|
| 72 |
from orbiitt_engine import OrbiittEngine
|
| 73 |
engine = OrbiittEngine()
|
|
@@ -79,8 +90,7 @@ def setup_environment():
|
|
| 79 |
|
| 80 |
except Exception as e:
|
| 81 |
loading_status = f"Setup Failed: {e}"
|
| 82 |
-
|
| 83 |
-
app.mount("/Productimages", StaticFiles(directory=IMAGE_DIR), name="Productimages")
|
| 84 |
|
| 85 |
@app.get("/health")
|
| 86 |
def health():
|
|
@@ -95,11 +105,11 @@ async def search(text: str = Form(None), weight: float = Form(0.5), file: Upload
|
|
| 95 |
try:
|
| 96 |
if file:
|
| 97 |
t_path = f"tmp_{file.filename}"
|
| 98 |
-
with open(t_path, "wb") as b:
|
|
|
|
| 99 |
|
| 100 |
results = engine.search(text_query=text, image_file=t_path, text_weight=weight)
|
| 101 |
|
| 102 |
-
# Fuzzy match results against disk to ensure valid URLs
|
| 103 |
all_files = os.listdir(IMAGE_DIR)
|
| 104 |
valid_results = []
|
| 105 |
|
|
@@ -107,12 +117,13 @@ async def search(text: str = Form(None), weight: float = Form(0.5), file: Upload
|
|
| 107 |
raw_path = r.get('url') or r.get('path') or ""
|
| 108 |
fname = normalize_filename(os.path.basename(raw_path))
|
| 109 |
|
|
|
|
| 110 |
match = None
|
| 111 |
if fname in all_files:
|
| 112 |
match = fname
|
| 113 |
else:
|
| 114 |
for disk_file in all_files:
|
| 115 |
-
if fname[:20] in disk_file:
|
| 116 |
match = disk_file
|
| 117 |
break
|
| 118 |
|
|
@@ -122,9 +133,11 @@ async def search(text: str = Form(None), weight: float = Form(0.5), file: Upload
|
|
| 122 |
|
| 123 |
return {"results": valid_results}
|
| 124 |
except Exception as e:
|
|
|
|
| 125 |
return {"results": [], "error": str(e)}
|
| 126 |
finally:
|
| 127 |
-
if t_path and os.path.exists(t_path):
|
|
|
|
| 128 |
|
| 129 |
if __name__ == "__main__":
|
| 130 |
import uvicorn
|
|
|
|
| 9 |
from fastapi.staticfiles import StaticFiles
|
| 10 |
from huggingface_hub import snapshot_download
|
| 11 |
|
| 12 |
+
# 1. SETUP LOGGING
|
| 13 |
logging.basicConfig(level=logging.INFO)
|
| 14 |
logger = logging.getLogger(__name__)
|
| 15 |
|
| 16 |
+
# 2. DEFINE PATHS
|
| 17 |
+
DATASET_REPO = "aniketkumar1106/orbit-data"
|
| 18 |
+
IMAGE_DIR = "Productimages"
|
| 19 |
+
DB_TARGET = "orbiitt.db"
|
| 20 |
+
|
| 21 |
+
# 3. CRITICAL FIX: Create directory BEFORE FastAPI initialization
|
| 22 |
+
if not os.path.exists(IMAGE_DIR):
|
| 23 |
+
os.makedirs(IMAGE_DIR, exist_ok=True)
|
| 24 |
+
logger.info(f"Created initial directory: {IMAGE_DIR}")
|
| 25 |
+
|
| 26 |
+
# 4. INITIALIZE APP
|
| 27 |
app = FastAPI()
|
| 28 |
|
|
|
|
| 29 |
app.add_middleware(
|
| 30 |
CORSMiddleware,
|
| 31 |
allow_origins=["*"],
|
|
|
|
| 34 |
allow_headers=["*"],
|
| 35 |
)
|
| 36 |
|
| 37 |
+
# 5. MOUNT STORAGE (Will no longer crash)
|
| 38 |
+
app.mount("/Productimages", StaticFiles(directory=IMAGE_DIR), name="Productimages")
|
|
|
|
| 39 |
|
| 40 |
+
# 6. GLOBAL ENGINE STATE
|
| 41 |
engine = None
|
| 42 |
+
loading_status = "System Booting..."
|
| 43 |
|
| 44 |
def normalize_filename(name):
|
| 45 |
+
"""Handles smart quotes and complex symbols to prevent 404s."""
|
|
|
|
| 46 |
name = name.replace('β', "'").replace('β', "'").replace('β', '"').replace('β', '"').replace('β', '-')
|
|
|
|
| 47 |
name = unicodedata.normalize('NFKD', name).encode('ascii', 'ignore').decode('ascii')
|
| 48 |
return name.strip()
|
| 49 |
|
|
|
|
| 52 |
global engine, loading_status
|
| 53 |
token = os.environ.get("HF_TOKEN")
|
| 54 |
|
| 55 |
+
# We clean and re-create inside startup to ensure a fresh sync
|
| 56 |
+
if os.path.exists(IMAGE_DIR):
|
| 57 |
+
shutil.rmtree(IMAGE_DIR)
|
| 58 |
+
os.makedirs(IMAGE_DIR, exist_ok=True)
|
| 59 |
|
| 60 |
try:
|
| 61 |
loading_status = "Syncing Orbit Assets..."
|
| 62 |
+
logger.info(loading_status)
|
| 63 |
snapshot_download(repo_id=DATASET_REPO, repo_type="dataset", token=token, local_dir=".")
|
| 64 |
|
| 65 |
if os.path.exists("orbiitt_db.zip"):
|
|
|
|
| 72 |
if f.lower() in ["orbiitt_db", "orbiitt.db"] or (f.endswith(".db") and not f.startswith(".")):
|
| 73 |
shutil.copy(src, f"./{DB_TARGET}")
|
| 74 |
elif f.lower().endswith(('.png', '.jpg', '.jpeg', '.webp')):
|
| 75 |
+
# Normalize filename on disk
|
| 76 |
clean_name = normalize_filename(urllib.parse.unquote(f))
|
| 77 |
shutil.copy(src, os.path.join(IMAGE_DIR, clean_name))
|
| 78 |
shutil.rmtree("temp_extract")
|
| 79 |
|
| 80 |
+
loading_status = "Waking up AI Engine..."
|
| 81 |
+
logger.info(loading_status)
|
| 82 |
try:
|
| 83 |
from orbiitt_engine import OrbiittEngine
|
| 84 |
engine = OrbiittEngine()
|
|
|
|
| 90 |
|
| 91 |
except Exception as e:
|
| 92 |
loading_status = f"Setup Failed: {e}"
|
| 93 |
+
logger.error(loading_status)
|
|
|
|
| 94 |
|
| 95 |
@app.get("/health")
|
| 96 |
def health():
|
|
|
|
| 105 |
try:
|
| 106 |
if file:
|
| 107 |
t_path = f"tmp_{file.filename}"
|
| 108 |
+
with open(t_path, "wb") as b:
|
| 109 |
+
b.write(await file.read())
|
| 110 |
|
| 111 |
results = engine.search(text_query=text, image_file=t_path, text_weight=weight)
|
| 112 |
|
|
|
|
| 113 |
all_files = os.listdir(IMAGE_DIR)
|
| 114 |
valid_results = []
|
| 115 |
|
|
|
|
| 117 |
raw_path = r.get('url') or r.get('path') or ""
|
| 118 |
fname = normalize_filename(os.path.basename(raw_path))
|
| 119 |
|
| 120 |
+
# Fuzzy match check
|
| 121 |
match = None
|
| 122 |
if fname in all_files:
|
| 123 |
match = fname
|
| 124 |
else:
|
| 125 |
for disk_file in all_files:
|
| 126 |
+
if fname[:20] in disk_file:
|
| 127 |
match = disk_file
|
| 128 |
break
|
| 129 |
|
|
|
|
| 133 |
|
| 134 |
return {"results": valid_results}
|
| 135 |
except Exception as e:
|
| 136 |
+
logger.error(f"Search error: {e}")
|
| 137 |
return {"results": [], "error": str(e)}
|
| 138 |
finally:
|
| 139 |
+
if t_path and os.path.exists(t_path):
|
| 140 |
+
os.remove(t_path)
|
| 141 |
|
| 142 |
if __name__ == "__main__":
|
| 143 |
import uvicorn
|