def make_predictions(year, features, sex):
import predict
PREDICTIVE_INDICES = [0, 25, 43, 65]
print 'make_predictions(year=%d)' % year
X, y, keys = getXy_by_features(year, features, sex)
X = X[:,PREDICTIVE_INDICES]
keys = [keys[i] for i in PREDICTIVE_INDICES]
predict.classify('sex=%s_year=%d' % (sex,year), X, y, keys)
def uploadPC():
print("in upload")
app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER
if request.method == 'POST':
# check if the post request has the file part
if 'file' not in request.files:
print(request.files)
d = Response("Error", status=201, mimetype='application/json')
return d
file = request.files['file']
# if user does not select file, browser also
# submit a empty part without filename
if file.filename == '':
d = Response("No filename",
status=201,
mimetype='application/json')
return d
if file:
filename = secure_filename(file.filename)
file.save(os.path.join(app.config['UPLOAD_FOLDER'], filename))
category, confidence = pred.classify(filename, response=False)
print(category, confidence)
confidence = '%.2f' % (confidence * 100)
if category == "malignant":
return render_template("malignant.html", confidence=confidence)
else:
return render_template("benign.html", confidence=confidence)
def sentiment_classification():
"""Run definition quality classification given text.
---
parameters:
- name: body
in: body
schema:
id: text
required:
- text
properties:
text:
type: string
description: the required text for POST method
required: true
definitions:
SentimentResponse:
Project:
properties:
status:
type: string
ml-result:
type: object
responses:
40x:
description: Client error
200:
description: Text Classification Response
examples:
[
{
"status": "success",
"category": "good"
},
{
"status": "error",
"message": "Exception caught"
},
]
"""
json_request = request.get_json()
if not json_request:
return Response("No json provided.", status=400)
text = json_request['text']
if text is None:
return Response("No text provided.", status=400)
else:
label = classify(text)
return flask.jsonify({"status": "success", "label": label})
def upload():
app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER
if request.method == 'POST':
# check if the post request has the file part
if 'file' not in request.files:
d = Response("Error", status=201, mimetype='application/json')
return d
file = request.files['file']
# if user does not select file, browser also
# submit a empty part without filename
if file.filename == '':
d = Response("No filename",
status=201,
mimetype='application/json')
return d
if file:
filename = secure_filename(file.filename)
file.save(os.path.join(app.config['UPLOAD_FOLDER'], filename))
return pred.classify(filename)
return render_template("index.html")
def run_classification(self, image_path):
label = predict.classify(image_path)
return label
def _predict():
text = request.json["text"]
labels = classify(text)
return jsonify(labels)
def main(alert_mode: bool, sound: bool, save_image: bool) -> None:
if sound:
if not Path(SOUND_ROOT_PATH).exists():
raise OheyaObeyaError(
'音源フォルダが見つかりません。(再配布不可の音源を使用しているため、GitHub上に音源ファイルは置いていません。使用する場合は、自身で用意してください)'
)
pygame.mixer.init()
# 実際には以下のサイトから素材を借りた
# 効果音ラボ: https://soundeffect-lab.info/sound/button/
# PANICPUMPKIN: http://pansound.com/panicpumpkin/index.html
# ※ 再配布不可なのでGitHub上には音源ファイルはuploadしていない
# 使用する際は指定のパスに適当な音源を配置すること
pygame.mixer.music.load(str(Path(SOUND_ROOT_PATH) / 'start.mp3'))
pygame.mixer.music.play(1)
camera_id = check_expected_camera_id()
messy_flag = False
messy_count = 0
not_messy_count = 0
cap = cv2.VideoCapture(camera_id)
pygame.init()
pygame.display.set_caption("OheyaObeya Classification Demo")
screen = pygame.display.set_mode()
status_font = pygame.font.Font(None, 100)
status_color = {
'messy': (255, 0, 0),
'so-so': (255, 165, 0),
'clean': (181, 255, 20)
}
sub_font = pygame.font.Font(None, 60)
if save_image:
for label_name in ['messy', 'so-so', 'clean']:
dir_path = Path(DST_IMAGE_PATH) / label_name
dir_path.mkdir(parents=True, exist_ok=True)
while True:
# Capture
ret, image = cap.read()
path = Path('now.jpg')
cv2.imwrite(str(path), image)
# Predict
result_dict = classify(path)
display_prediction_result(result_dict)
result = result_dict['prediction']
# 保存モードがONの場合は、日時_推測結果.jpgで画像を保存する
if save_image:
file_name = datetime.now().strftime(
"%Y%m%d_%H%M%S_{}.jpg".format(result))
path = Path(DST_IMAGE_PATH) / result / file_name
cv2.imwrite(str(path), image)
# TODO: 連続した回数ではなく、過去n回分のm%分で判断させる
if result == 'messy':
messy_count += 1
not_messy_count = 0
logger.debug('messy_count = {}'.format(messy_count))
else:
not_messy_count += 1
messy_count = 0
logger.debug('not_messy_count = {}'.format(not_messy_count))
# スクリーンを初期化
screen.fill([0, 0, 0])
# スクリーンにカメラの映像を表示する
frame = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
frame = frame[:, ::-1]
frame = np.rot90(frame)
# frame = frame[:, ::-1] # 上下を逆にする
frame = pygame.surfarray.make_surface(frame)
screen.blit(frame, (0, 0))
# スクリーンに判定結果を表示する
status_text = status_font.render(result, True, status_color[result])
screen.blit(status_text, [20, 20])
# スクリーンにmessyの確率を表示する
messy_prob = [
x['probability'] for x in result_dict['predictions']
if x['label'] == 'messy'
][0]
messy_prob_text = sub_font.render('{:.2f}'.format(messy_prob), True,
(255, 0, 0))
screen.blit(messy_prob_text, [20, 100])
if messy_flag and alert_mode:
# スクリーンに警報の文字を表示する
messy_alarm = '!!! Obeya Alarm!!!'
messy_alarm_text = sub_font.render(messy_alarm, True, (255, 0, 0))
screen.blit(messy_alarm_text, [20, 150])
# 警報中は、messyじゃない状態が続いたら緑のバーが伸びていく
not_messy_bar = '{}: {}'.format(not_messy_count,
'*' * not_messy_count)
not_messy_bar_text = sub_font.render(not_messy_bar, True,
(181, 255, 20))
screen.blit(not_messy_bar_text, [150, 100])
else:
# messyの状態が続いたら赤いバーが伸びていく
messy_bar = '{}: {}'.format(messy_count, '*' * messy_count)
messy_bar_text = sub_font.render(messy_bar, True, (255, 0, 0))
screen.blit(messy_bar_text, [150, 100])
pygame.display.update()
# 汚部屋警報のON/OFFが切り替わったか判定
if messy_count > 10 and not messy_flag and alert_mode:
messy_flag = True
alert_obeya(on=True, sound=sound)
elif not_messy_count > 10 and messy_flag:
messy_flag = False
alert_obeya(on=False, sound=sound)
for event in pygame.event.get():
pass
cap.release()
from sklearn.naive_bayes import MultinomialNB from predict import classify # Making final predictions with Multinomial NB multi_NB = MultinomialNB(alpha=0.25) # Calling classify function from predict module classify(multi_NB)