main.py

from fastapi import FastAPI, File, Request
from functions.ImageDetectorV1 import detectorV1
from functions.ImageDetectorV2 import detectorV2
from functions.newassignmentV1 import new_assignmentV1
from functions.newassignmentV2 import new_assignmentV2
from fastapistats import Stats
import json

import os
from datetime import datetime
from apscheduler.schedulers.background import BackgroundScheduler
from fastapi.responses import RedirectResponse
import uuid


app = FastAPI()
update = Stats.update_stats


# THese 2 functions help count if objects were detected for the dashboard
@app.get('/countobjectdetected', response_class=RedirectResponse)
@update(name='Detected') 
def objectdetected():
    return RedirectResponse("https://two-trick-pony-nl-dashboard-photoscavenger-dashboard-jvqrac.streamlitapp.com/")
@app.get('/countobjectnotdetected', response_class=RedirectResponse)
@update(name='NotDetected') 
def objectnotdetected():
    return RedirectResponse("https://two-trick-pony-nl-dashboard-photoscavenger-dashboard-jvqrac.streamlitapp.com/")



@app.get('/', response_class=RedirectResponse)
@update(name='Homepage') 
def read_root(request: Request):
    print("redirect to dashboard")
    return RedirectResponse("https://two-trick-pony-nl-dashboard-photoscavenger-dashboard-jvqrac.streamlitapp.com/")



@app.get('/exampleresponse')
@update(name='ExampleResponse') 
def exampleresponse():
    return {'Searchedfor:': 'boat', 
            'Wasfound': 'NO', 
            'OtherObjectsDetected': ['person', 
                                        'person', 
                                        'person', 
                                        'person', 
                                        'bicycle', 
                                        'motorbike', 
                                        'bicycle', 
                                        'motorbike', 
                                        'bicycle'], 
            'Processed_FileName': 'scanned_image54e46fb8-93f8-43ad-a8ec-99eb83f260af.jpg', 
            'file_url': 'image54e46fb8-93f8-43ad-a8ec-99eb83f260af.jpg', 
            'listofobjectsWithConfidence': 
                            [{'person': 100}, 
                            {'person': 100}, 
                            {'person': 100}, 
                            {'person': 100}, 
                            {'bicycle': 97}, 
                            {'motorbike': 91}, 
                            {'bicycle': 75}, 
                            {'motorbike': 48}, 
                            {'bicycle': 28}]}


# This endpoint triggers for this API call: 127.0.0.1:8000/uploadfile/cat. 
# Add a image to your call in the body, and set a expected object in the URL. 
# The function then runs the AI model to see if the picture contains the expected object
@app.post('/uploadfile/{expectedobject}')
@update(name='uploadfileV1')  # if the name kwarg is not passed it will default to the function name
async def UploadImage(expectedobject, file: bytes = File(...)):
    uniqueid = str(uuid.uuid4())
    with open('image'+uniqueid+'.jpg', 'wb') as image:
        rawimage = 'image'+uniqueid+'.jpg'
        os.chdir('static')
        image.write(file)
        image.close()
        os.chdir('..')
        results = detectorV1(rawimage, expectedobject)
        objectfound = results[1]
        otherobjectsdetected = results[2]
        listofobjectsWithConfidence = results[3]
        filename = results[0]
        # remove uploaded image
        os.remove(rawimage)

    return {'Searchedfor:': expectedobject, 'Wasfound': objectfound, 'OtherObjectsDetected': otherobjectsdetected, 'Processed_FileName': filename, 'file_url': rawimage, 'listofobjectsWithConfidence': listofobjectsWithConfidence}


@app.post("/v2/uploadfile/{expectedobject}")
@update(name='uploadfileV2')
async def UploadImageV2(expectedobject, file: bytes = File(...)):
    # Receiving the image the image
    uniqueid = str(uuid.uuid4())
    with open('image'+uniqueid+'.jpg', 'wb') as image:
        # Saving te raw image
        rawimage = 'image'+uniqueid+'.jpg'
        image.write(file)
        # Calling the detector
        results = detectorV2(rawimage, expectedobject)
        # Parsing the different results
        objectfound = results[1]
        print(objectfound)
        if objectfound == 'NO':
            await objectnotdetected()
        else: 
            await objectdetected()
        filename = results[0]
        otherobjectsdetected = results[2]
        # Removing the image uploaded so we don't store any data
        os.remove(rawimage)
    return {'Searchedfor:': expectedobject, 'Wasfound': objectfound, 'OtherObjectsDetected': otherobjectsdetected, 'Processed_FileName': filename}


@app.get('/newassignment/{score}')
@update(name='NewAssignmentV1')  # if the name kwarg is not passed it will default to the function name
def new_assignment(score: int):
    assignment = new_assignmentV1(score)
    return assignment


@app.get('/v2/newassignment/{score}')
@update(name='NewAssignmentV2')  # if the name kwarg is not passed it will default to the function name
def new_assignment_v2(score: int):
    assignment = new_assignmentV2(score)
    return assignment
  

@app.get('/healthcheck')
@update(name='Healthcheck by AWS')
def healthcheck():
    return 'Server is up'


@app.get('/get_long_term_data/')
def get_long_term_data():
    with open('statshistoric.json', 'r+') as payload:
        payload = json.load(payload)
        return payload
    

def daily_analytics_update():
    # Get the current days stats and add a timestamp
    with open('stats.json', 'r+') as data:
        new_data = json.load(data)
        date = {'timestamp': datetime.today().strftime('%Y-%m-%d %H:%M')}
        new_data.update(date)

    with open('statshistoric.json', 'r+') as file:
        # First we load existing data into a dict.
        file_data = json.load(file)
        # Join new_data with file_data inside emp_details
        file_data["LongtermData"].append(new_data)
        # Sets file's current position at offset.
        file.seek(0)
        # convert back to json.
        json.dump(file_data, file, indent=4)
 
    # reset the daily json to 0 
    for i in new_data:
        new_data[i] = 0
    with open("stats.json", "w") as f:
        json.dump(new_data, f, indent=4)


# This schedular runs every hour making a snapshot of the stats.json 
# file and stores in in the longermdata file
scheduler = BackgroundScheduler()
scheduler.add_job(daily_analytics_update, 'interval', hours=48)
scheduler.start()

daily_analytics_update()