Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,9 +1,7 @@
|
|
| 1 |
# app.py
|
| 2 |
-
import io
|
| 3 |
import uvicorn
|
| 4 |
import numpy as np
|
| 5 |
import cv2
|
| 6 |
-
import re
|
| 7 |
from fastapi import FastAPI, UploadFile, File, HTTPException
|
| 8 |
from bson import ObjectId
|
| 9 |
from motor.motor_asyncio import AsyncIOMotorClient, AsyncIOMotorGridFSBucket
|
|
@@ -19,16 +17,15 @@ MONGO_URI = (
|
|
| 19 |
)
|
| 20 |
|
| 21 |
DB_NAME = "ocr_fastapi"
|
| 22 |
-
BUCKET_NAME = "
|
| 23 |
|
| 24 |
app = FastAPI()
|
| 25 |
|
| 26 |
-
# MongoDB
|
| 27 |
_client = AsyncIOMotorClient(MONGO_URI)
|
| 28 |
_db = _client[DB_NAME]
|
| 29 |
|
| 30 |
def gridfs():
|
| 31 |
-
# Always use default "fs" bucket (fs.files + fs.chunks)
|
| 32 |
return AsyncIOMotorGridFSBucket(_db, bucket_name=BUCKET_NAME)
|
| 33 |
|
| 34 |
ocr_engine = RapidOCR()
|
|
@@ -42,14 +39,11 @@ async def health():
|
|
| 42 |
# --------------------------------------------------
|
| 43 |
@app.post("/upload")
|
| 44 |
async def upload_image(file: UploadFile = File(...)):
|
| 45 |
-
"""
|
| 46 |
-
Uploads image to MongoDB GridFS
|
| 47 |
-
"""
|
| 48 |
try:
|
| 49 |
data = await file.read()
|
| 50 |
fs = gridfs()
|
| 51 |
|
| 52 |
-
# Upload to GridFS
|
| 53 |
oid = await fs.upload_from_stream(
|
| 54 |
file.filename,
|
| 55 |
data,
|
|
@@ -58,9 +52,9 @@ async def upload_image(file: UploadFile = File(...)):
|
|
| 58 |
|
| 59 |
print("π Uploaded ID:", oid)
|
| 60 |
|
| 61 |
-
# Debug: Confirm stored in
|
| 62 |
stored = await _db[f"{BUCKET_NAME}.files"].count_documents({"_id": oid})
|
| 63 |
-
print("π¦ Stored in
|
| 64 |
|
| 65 |
return {"image_id": str(oid)}
|
| 66 |
|
|
@@ -69,23 +63,20 @@ async def upload_image(file: UploadFile = File(...)):
|
|
| 69 |
|
| 70 |
# --------------------------------------------------
|
| 71 |
async def load_image_from_gridfs(image_id: str):
|
| 72 |
-
"""
|
| 73 |
-
Loads the image bytes from GridFS bucket
|
| 74 |
-
"""
|
| 75 |
try:
|
| 76 |
print("π Fetching from GridFS:", image_id)
|
| 77 |
|
| 78 |
oid = ObjectId(image_id)
|
| 79 |
fs = gridfs()
|
| 80 |
|
| 81 |
-
#
|
| 82 |
file_exists = await _db[f"{BUCKET_NAME}.files"].count_documents({"_id": oid})
|
| 83 |
-
print("π Exists in
|
| 84 |
|
| 85 |
if file_exists == 0:
|
| 86 |
-
raise HTTPException(status_code=404, detail="Image not found
|
| 87 |
|
| 88 |
-
# Read file
|
| 89 |
stream = await fs.open_download_stream(oid)
|
| 90 |
data = await stream.read()
|
| 91 |
await stream.close()
|
|
@@ -101,29 +92,27 @@ async def load_image_from_gridfs(image_id: str):
|
|
| 101 |
# --------------------------------------------------
|
| 102 |
@app.post("/generate/{image_id}")
|
| 103 |
async def generate(image_id: str):
|
| 104 |
-
|
| 105 |
-
Reads image β performs OCR β returns extracted text
|
| 106 |
-
"""
|
| 107 |
raw_bytes = await load_image_from_gridfs(image_id)
|
| 108 |
|
| 109 |
-
# Decode
|
| 110 |
img_array = np.frombuffer(raw_bytes, np.uint8)
|
| 111 |
img = cv2.imdecode(img_array, cv2.IMREAD_COLOR)
|
| 112 |
|
| 113 |
if img is None:
|
| 114 |
raise HTTPException(status_code=400, detail="Unable to decode image")
|
| 115 |
|
| 116 |
-
#
|
| 117 |
result, times = ocr_engine(img)
|
| 118 |
|
| 119 |
if not result:
|
| 120 |
raise HTTPException(status_code=500, detail="OCR returned empty result")
|
| 121 |
|
| 122 |
-
# Extract
|
| 123 |
extracted = [text for box, text, score in result]
|
| 124 |
full_text = "\n".join(extracted)
|
| 125 |
|
| 126 |
-
# Cache
|
| 127 |
OCR_RESULTS[image_id] = {
|
| 128 |
"text": full_text,
|
| 129 |
"details": result,
|
|
@@ -136,20 +125,11 @@ async def generate(image_id: str):
|
|
| 136 |
@app.get("/view/{image_id}")
|
| 137 |
async def view_details(image_id: str):
|
| 138 |
"""
|
| 139 |
-
Returns
|
| 140 |
"""
|
| 141 |
if image_id not in OCR_RESULTS:
|
| 142 |
-
raise HTTPException(status_code=404, detail="No OCR result found
|
| 143 |
return OCR_RESULTS[image_id]
|
| 144 |
-
@app.get("/debug/db")
|
| 145 |
-
async def debug_db():
|
| 146 |
-
return {"db_name": _db.name, "collections": await _db.list_collection_names()}
|
| 147 |
-
|
| 148 |
-
@app.get("/debug/test-write")
|
| 149 |
-
async def test_write():
|
| 150 |
-
res = await _db["debug_test"].insert_one({"ping": "ok"})
|
| 151 |
-
return {"inserted_id": str(res.inserted_id)}
|
| 152 |
-
|
| 153 |
|
| 154 |
# --------------------------------------------------
|
| 155 |
if __name__ == "__main__":
|
|
|
|
| 1 |
# app.py
|
|
|
|
| 2 |
import uvicorn
|
| 3 |
import numpy as np
|
| 4 |
import cv2
|
|
|
|
| 5 |
from fastapi import FastAPI, UploadFile, File, HTTPException
|
| 6 |
from bson import ObjectId
|
| 7 |
from motor.motor_asyncio import AsyncIOMotorClient, AsyncIOMotorGridFSBucket
|
|
|
|
| 17 |
)
|
| 18 |
|
| 19 |
DB_NAME = "ocr_fastapi"
|
| 20 |
+
BUCKET_NAME = "ocr_images" # β USE THIS BUCKET
|
| 21 |
|
| 22 |
app = FastAPI()
|
| 23 |
|
| 24 |
+
# Init MongoDB & GridFS
|
| 25 |
_client = AsyncIOMotorClient(MONGO_URI)
|
| 26 |
_db = _client[DB_NAME]
|
| 27 |
|
| 28 |
def gridfs():
|
|
|
|
| 29 |
return AsyncIOMotorGridFSBucket(_db, bucket_name=BUCKET_NAME)
|
| 30 |
|
| 31 |
ocr_engine = RapidOCR()
|
|
|
|
| 39 |
# --------------------------------------------------
|
| 40 |
@app.post("/upload")
|
| 41 |
async def upload_image(file: UploadFile = File(...)):
|
|
|
|
|
|
|
|
|
|
| 42 |
try:
|
| 43 |
data = await file.read()
|
| 44 |
fs = gridfs()
|
| 45 |
|
| 46 |
+
# Upload image bytes to GridFS
|
| 47 |
oid = await fs.upload_from_stream(
|
| 48 |
file.filename,
|
| 49 |
data,
|
|
|
|
| 52 |
|
| 53 |
print("π Uploaded ID:", oid)
|
| 54 |
|
| 55 |
+
# Debug: Confirm stored in ocr_images.files
|
| 56 |
stored = await _db[f"{BUCKET_NAME}.files"].count_documents({"_id": oid})
|
| 57 |
+
print("π¦ Stored in ocr_images.files:", stored)
|
| 58 |
|
| 59 |
return {"image_id": str(oid)}
|
| 60 |
|
|
|
|
| 63 |
|
| 64 |
# --------------------------------------------------
|
| 65 |
async def load_image_from_gridfs(image_id: str):
|
|
|
|
|
|
|
|
|
|
| 66 |
try:
|
| 67 |
print("π Fetching from GridFS:", image_id)
|
| 68 |
|
| 69 |
oid = ObjectId(image_id)
|
| 70 |
fs = gridfs()
|
| 71 |
|
| 72 |
+
# Check file existence
|
| 73 |
file_exists = await _db[f"{BUCKET_NAME}.files"].count_documents({"_id": oid})
|
| 74 |
+
print("π Exists in ocr_images.files:", file_exists)
|
| 75 |
|
| 76 |
if file_exists == 0:
|
| 77 |
+
raise HTTPException(status_code=404, detail="Image not found")
|
| 78 |
|
| 79 |
+
# Read file from GridFS
|
| 80 |
stream = await fs.open_download_stream(oid)
|
| 81 |
data = await stream.read()
|
| 82 |
await stream.close()
|
|
|
|
| 92 |
# --------------------------------------------------
|
| 93 |
@app.post("/generate/{image_id}")
|
| 94 |
async def generate(image_id: str):
|
| 95 |
+
# Load the image data
|
|
|
|
|
|
|
| 96 |
raw_bytes = await load_image_from_gridfs(image_id)
|
| 97 |
|
| 98 |
+
# Decode into OpenCV format
|
| 99 |
img_array = np.frombuffer(raw_bytes, np.uint8)
|
| 100 |
img = cv2.imdecode(img_array, cv2.IMREAD_COLOR)
|
| 101 |
|
| 102 |
if img is None:
|
| 103 |
raise HTTPException(status_code=400, detail="Unable to decode image")
|
| 104 |
|
| 105 |
+
# Run OCR
|
| 106 |
result, times = ocr_engine(img)
|
| 107 |
|
| 108 |
if not result:
|
| 109 |
raise HTTPException(status_code=500, detail="OCR returned empty result")
|
| 110 |
|
| 111 |
+
# Extract recognized text only
|
| 112 |
extracted = [text for box, text, score in result]
|
| 113 |
full_text = "\n".join(extracted)
|
| 114 |
|
| 115 |
+
# Cache OCR results for viewing
|
| 116 |
OCR_RESULTS[image_id] = {
|
| 117 |
"text": full_text,
|
| 118 |
"details": result,
|
|
|
|
| 125 |
@app.get("/view/{image_id}")
|
| 126 |
async def view_details(image_id: str):
|
| 127 |
"""
|
| 128 |
+
Returns OCR results stored in memory.
|
| 129 |
"""
|
| 130 |
if image_id not in OCR_RESULTS:
|
| 131 |
+
raise HTTPException(status_code=404, detail="No OCR result found")
|
| 132 |
return OCR_RESULTS[image_id]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 133 |
|
| 134 |
# --------------------------------------------------
|
| 135 |
if __name__ == "__main__":
|