This desktop app was made for a private project, so it cannot be found on my github repository. Instead, I took some pieces of code as examples of its logic. It was made with the purpose of creating semi-random signatures, from which addon images were merge to create a concrete sole image. Then, those images were exposed to be classified by hand. Finally, a series of attributes, like a text, name and such, were assigned to those filtered. It does not imply a lot of complex logic, but a lot of things are taken into account to make it work, so the code became quite long.

Token section's event manager

After setting a number of signatures to be created and letting the app do its job, the user can enter this Tinder-like section, where the images are shown and the user can swipe left or right.

while True:
    if len(list(rawSignatures.keys())) > 0:
        nextKey = list(rawSignatures.keys())[-1]
    else:
        nextKey = None
    if nextKey != None:
        nextImgPath = os.path.join(pathTokensSmall, (str(nextKey) + ".png"))
        window["token"].Update(filename = nextImgPath)
        window.refresh()
    else:
        window["token"].Update(filename = (fileTokenPH))
        
    event, values = window.read()
    if event == sg.WIN_CLOSED:
        break
    elif event == "-BACK_BUTTON-" or event in ["a", "A", "Left:37"]:
        logs = fg.loadJSON(logsFile)
        if (len(logs) > 0):
            rawSignatures, discardedSignatures, superLikeSignatures, approvedSignatures = (
                revert_change(rawSignatures, discardedSignatures, superLikeSignatures, approvedSignatures)
                )
    [...]
    elif  event == '-RETURN-':
        window.close()
        main.main()
        break
    download_signatures(rawSignatures, discardedSignatures, superLikeSignatures, approvedSignatures)
    updateNumbers(window)
window.close()

Profiles management

The app was made with PySimpleGUI and it has a very simple interface. It is divided into three big sections: the first one is the image merger, and the second one is the image classifier and there's a third part in which the images are assigned properties to form profiles, which can be overviewed. The app is in Spanish, but I will translate the code to English so it can be understood.

Profiles' body texts are obtained through some CSV extraction and loaded into a single dictionary.

def loadCSVTexts():
    """ Loads all the phrases' CSVs in a dictionary, downloads 
    it to a JSON, and returns the dictionary in the process """
    texts = {}
    filesNames = list(os.listdir(path_texts))
    #All CSVs are loaded into a single dictionary
    for file in filesNames:
        name = file.split(sep=".")[0]
        data = open(os.path.join(path_texts, file), "r")
        csvreader = csv.reader(data, delimiter='|')
        header = next(csvreader)
        texts[name] = {}
        id = 1
        for i in csvreader:
            texts[name][str(id)] = {header[0] : str(i[0]),
                header[1] : str(i[1]),
            }
            id += 1
        data.close()
    return texts

Results summary

In addition, a test summary was added in the latest versions to check that no addon image was being left aside by mistake, and that occurrency chances where being respected.