laurentmih

import asyncio
from pyppeteer import launch
from PyPDF2 import PdfFileMerger
from pathlib import Path

merge_files = True

urls = [
    'https://pytorch.org/tutorials/beginner/blitz/tensor_tutorial.html',
    'https://pytorch.org/tutorials/beginner/blitz/autograd_tutorial.html#sphx-glr-beginner-blitz-autograd-tutorial-py',

laurentmih

import torch
from torch.utils.data import TensorDataset, DataLoader
from fastai.basic_data import DataBunch

# This is a typical setup for images
# x images have 3 channels, 128x128
# y has 1 channel because I needed to
# predict segmentation
dataset_size = 4
x = torch.ones(dataset_size,3,128,128)#.cuda()

laurentmih

def replace_exif(file, datetime_object):
	""" 
	Replaces the createdTime Exif data of image with date parsed from the datetime_object

	This function replaces the following three properties of the Exif data of 
	the image file that is passed to it:
		Exif.Image.DateTime, at offset 306
		Exif.Photo.DateTimeOriginal at offset 36867
		Exif.Photo.DateTimeDigitized at offset 36868
	This function expects that both PIL and piexif are imported

laurentmih

^account/ ^register/closed/$ [name='registration_disallowed']
^account/ ^register/$ [name='registration_register']
^account/ ^login/$ [name='auth_login']
^account/ ^logout/$ [name='auth_logout']
^account/ ^password/change/$ [name='auth_password_change']
^account/ ^password/change/done/$ [name='auth_password_change_done']
^account/ ^password/reset/$ [name='auth_password_reset']
^account/ ^password/reset/complete/$ [name='auth_password_reset_complete']
^account/ ^password/reset/done/$ [name='auth_password_reset_done']
^account/ ^password/reset/confirm/(?P<uidb64>[0-9A-Za-z_\-]+)/(?P<token>.+)/$ [name='auth_password_reset_confirm']

laurentmih

// ChartJS charts
$(document).ready(function(){
  // Coding/language barchart
  var codingChart = $("#codinglevel");
  var codingChartData = {
    labels: ["Python   ", "Octave/MATLAB   ", "JavaScript   "],
    datasets: [{
      data: [2.7, 2, 2],
      backgroundColor: ["rgba(255, 99, 132, 0.2)", "rgba(255, 159, 64, 0.2)", "rgba(255, 205, 86, 0.2)"],
      borderColor: ["rgb(255, 99, 132)", "rgb(255, 159, 64)", "rgb(255, 205, 86)"],

laurentmih

I've been working on sklearn with the Kaggle Titanic set, and just discovered sklearn Pipelines and FeatureUnions. While following most of the content out there on how to work with the Titanic dataset, I started to rewrite all the usual functions into Transformers, that can be used in the FeatureUnion/Pipeline context.

Usually, you use FeatureUnions for (axis=1) concatenation of transformers output. Here, I created a UnityTransformer that effectively does nothing to the input, that way I could keep the original column, and concatenate it with a transformed version of itself.

The 'trick' here is in the transform method: just multiply the output by 1, a unity operation. I haven't extensively tested it, but it seems to work fine for strings as well as normal np.arrays. The whole thing feels a little hacky though, so I'll be happy to have feedback.

from sklearn.base import BaseEstimator, TransformerMixin

class IdentityTransformer(BaseEstimator, TransformerMixin):

laurentmih

# This quick Python script should help you detect whether you have overlapping entities in your training data.
# These errors thrown by MITIE usually look like:
# Invalid range given to ner_training_instance.overlaps_any_entity(). It overlaps an entity given to a previous call to add_entity().

# The script assumes the form of the trainingdata is a JSON-encoded object.
# For an example of such a trainingset, see https://raw.githubusercontent.com/golastmile/rasa_nlu/master/data/examples/rasa/demo-rasa.json
# It should be trivial to adapt for other forms though.

# Tested on Python3