Dataset for testing

For contributer of this library, you can use the default testing images in ../test/img, or you can uncomment the following and download more data you like

# !pip install -q jmd_imagescraper

# from jmd_imagescraper.core import duckduckgo_search, ImgSize

# duckduckgo_search("../test/img", "Nature", "nature", max_results=20)

!du -sh ../test

996K	../test

Labeler

class ImageLabeler[source]

ImageLabeler(image_folder:Path, formats:List[str]=['jpg', 'jpeg', 'png', 'bmp'])

class SingleClassImageLabeler[source]

SingleClassImageLabeler(image_folder:Path) :: ImageLabeler

class MultiClassImageLabeler[source]

MultiClassImageLabeler(image_folder:Path) :: ImageLabeler

Try labeler

slabel = SingleClassImageLabeler("../test")
slabel

load from progress ../test/unpackai_single.json

SingleClassImageLabeler on [../test(20)], see labeler.image_df

slabel()

slabel.progress["data"]

mlabel = MultiClassImageLabeler("../test")
mlabel

load from progress ../test/unpackai_multi.json

MultiClassImageLabeler on [../test(20)], see labeler.image_df

mlabel(labels=["spring", "summer", "autumn", "winter"])

press Command(mac) or Ctrl(win/linux) to select multiple

get_y

get_y = slabel.get_y

Multicategorical label

Create gey_y from a running MultiClassImageLabeler

mlabel.get_y(str(Path('../test/img/Nature/007_9554a747.jpg').resolve()))

Creating dataset with 16 labels

['autumn', 'winter']

Look at the data progress

mlabel.progress["data"]

Create a get_y function from csv file

get_y = MultiClassImageLabeler.gety_from_csv("./progress.csv")

All possible labels:	['summer', 'spring', 'winter', 'autumn']

Create a get_y function from external csv file, please pass in name of the image path column and label column

get_y = MultiClassImageLabeler.gety_from_csv("./progress.csv", path="image_path", label="pepper_types")

get_y('../test/img/Nature/004_26fb347c.jpg')

['summer', 'autumn']

how we label multi categorical?

• please do it as following

pd.read_csv("./progress.csv")["label"]

0            autumn,winter
1            spring,summer
2     spring,summer,autumn
3            summer,autumn
4            spring,winter
5            spring,summer
6            spring,summer
7     summer,autumn,winter
8     spring,summer,autumn
9                   spring
10                     NaN
11                     NaN
12                  spring
13           summer,autumn
14                  spring
15                  summer
Name: label, dtype: object