def test_storage_remembers_type(self):
from model import TextModel, EquationModel, ImageModel, WrenModel
text_model = TextModel('text1')
text_model.save()
equation_model = EquationModel('equation1')
equation_model.save()
image_model = ImageModel('image1')
image_model.save()
self.assertIsInstance(WrenModel.load(text_model.key),
TextModel)
self.assertIsInstance(WrenModel.load(equation_model.key),
EquationModel)
self.assertIsInstance(WrenModel.load(image_model.key),
ImageModel)
def main(terminal_flag):
"""Main program."""
# Create PyQt5 application
app = QApplication(
sys.argv
) # Note: There must be exactly one instance of QApplication active at a time
app.setWindowIcon(QIcon('icons/ieviewer.png')) # Set program icon
# Create model, view and controller
model = ImageModel(terminal_flag)
view = View(model)
controller = Controller(model, view)
# Start main window
main_window = controller.get_main_window()
main_window.show()
sys.exit(app.exec_())
def image_list(self, unused_request):
qry = ImageModel.query()
my_items = []
for img in qry:
my_items.append(
ImageMessage(
name=img.name,
image=img.image,
uuid=img.uuid,
review_uuid=img.review_uuid,
revobj_uuid=img.revobj_uuid,
is_main=img.is_main,
deleted=img.deleted,
)
)
return ImageMessageCollection(items=my_items)
def image_sync_down(self, request):
query = ImageModel.query(ImageModel.last_modified > string_to_datetime(request.date))
my_items = []
for img in query:
my_items.append(
ImageMessage(
name=img.name,
image=img.image,
uuid=img.uuid,
review_uuid=img.review_uuid,
revobj_uuid=img.revobj_uuid,
is_main=img.is_main,
last_modified=img.last_modified,
deleted=img.deleted,
)
)
return ImageMessageCollection(items=my_items)
import base64
import json
from io import BytesIO
import cv2
import numpy as np
from flask import Flask, Response, render_template, request
from model import ImageModel
app = Flask(__name__)
model = ImageModel('sample_model')
def predict(data):
image_dec = cv2.imdecode(data, 1).astype(np.float32) / 255.0
x = np.expand_dims(image_dec, axis=0)
label = model.predict(x)
return label
@app.route('/predict_from_base64/', methods=['POST'])
def predict_from_base64():
data_json = json.loads(request.data)
image_bytes = base64.b64decode(data_json['image'])
image_np = np.frombuffer(image_bytes, dtype='uint8')
label = predict(image_np)
return Response(json.dumps({'label': label}))
@app.route('/predict_from_image/', methods=['POST'])
from tempfile import NamedTemporaryFile
import cv2
import numpy as np
import tensorflow as tf
from fastapi import FastAPI, File, UploadFile
from pydantic import BaseModel
from model import ImageModel
# create FastAPI instance
app = FastAPI()
# create model instance
# model = ImageModel('model folder path')
model = ImageModel('sample_model')
class ImageBase64(BaseModel):
base64_str: str
# get input shape for model from signature
input_shape = model.get_input_shape('sample_model/signature.json')
@app.post("/predict_from_image/")
async def predict(file: UploadFile = File(...)):
filepath = save_upload_file_tmp(file)
image = tf.keras.preprocessing.image.load_img(filepath,
color_mode='rgb',
