Fixes #289

[vfdev-5]

Fixes #289

Description:
Updated ProgressBar.attach API by adding supplementary args:

• event_name=Events.ITERATION_COMPLETED
• closing_event_name=Events.EPOCH_COMPLETED)
  to handle demanded use-case:

    n_epochs = 10
    loader = [1, 2, 3, 4, 5]
    engine = Engine(update_fn)

    pbar = ProgressBar()
    pbar.attach(engine, event_name=Events.EPOCH_STARTED, closing_event_name=Events.COMPLETED)
    engine.run(loader, max_epochs=n_epochs)
    # expected = u'Epoch: [9/10]  90%|█████████  [00:00<00:00]'

• Custom events are not supported yet !
  • We need to order base events and custom events to prevent the cases where closing event happens before logging event
  • It is not evident how to compute length of progress bar with custom events and number of total custom events...

Check list:

• New tests are added (if a new feature is added)
• New doc strings: description and/or example code are in RST format
• Documentation is updated (if required)

[vfdev-5]

DO NOT MERGE YET. Forgot about #391

EDIT: Now OK

[vfdev-5]

@drozzy you can check out this version if it works better for you.
Keep in mind that tqdm_notebook has some issue with the layout in some cases: tqdm/tqdm#649

[anmolsjoshi]

@vfdev-5 awesome work! The enter and exit ended up fixing the previous incorrect print outs.

• Could you run the TextCNN example again to fix the progress bar display?
• It might be worth it to add a note about bar_format being empty when using with a notebook.

[vfdev-5]

@anmolsjoshi OK, I'll update the notebooks ASAP

[vfdev-5]

@anmolsjoshi I updated TextCNN and docstring. Rendering of progressbar in the github is not better.
Just HBox ... however inside jupyter notebook we can see the widgets etc... We can leave it like this, IMO

[anmolsjoshi]

@vfdev-5 LGTM! I agree that we can ignore the rendering as it works correctly in terminal and notebook.

[drozzy]

@vfdev-5 Sorry I haven't used ignite in a while so I can't actually reproduce it. Is there a minimal working example that I can use to test this?

[vfdev-5]

@drozzy check this one:

import torch
import time

from ignite.engine import Engine, Events
from ignite.contrib.handlers import ProgressBar

def update(engine, batch):
    time.sleep(0.5)

def run(data):
    trainer = Engine(update)
    ProgressBar(persist=True, bar_format="").attach(trainer)
    trainer.run(data, max_epochs=2)

data = [1] * 10
run(data)