OverIQ

Machine Learning Experts You Should Be Following Online

One of the most fast evolving domains in today's tech savvy world is Machine Learning and
Deep Learning. An area expected to continue its vast growth and development also in the coming
years. Getting ourselves up-to-date with the latest on the advancement in technologies, the
trending news, industry patterns, and opinions of compelling figures can appear to be
overpowering.

But one prime approach to remain educated is through Twitter, which has finally proved to be an
useful and engaging approach to remain circled into the insights of a portion of the top
personalities and spearheading scholars in AI, who utilize their records to voice their feelings on
the most recent news and industry rehearses.

If you are someone interested to learn about the latest developments in this field, gain important
perspectives from the best brains and be future proof, then there cannot be a better way other
than following these Machine Learning experts and their successful journey. Beginners like me
have been highly inspired by these heroes of machine learning and can continue to look up to
their research work every day to explore more with their algorithms to increase our knowledge.
To kick you off, I have incorporated a summary of top 10 machine learning experts, profound
learning influencers and specialists whom you ought to follow online. This rundown is a blend of
business leaders, originators, specialists, entrepreneurs, scientists and columnists, every one of
whom welcome a kind interpretation of AI to the table.

1. ANDREW NG #

The most famous name in the book of AI enthusiast, Andrew Yan-Tak Ng is considered as the
most prominent mind in the domain of Machine Learning and Deep Learning in the present time.
The co-founder and chairman of Coursera and deeplearning.ai, is a british-born American
technology entrepreneur. He is presently mentoring in the department of Computer Science as an
Adjunct Professor at Stanford University.

Educator Andrew was the head and helped to establish the Google Brain project and has already
worked as the Chief Scientist at Baidu. He is happily married to fellow AI influencer Carol
Reiley.

His point is to democratize profound learning and give everybody on the planet admittance to top
notch schooling without any cost, the advantage we receive from Coursera. His online seminars
on AI and profound learning are exceptionally pursued by many students and professionals all
around the world. An extraordinary history of a scholarly analyst is what Andrew holds in this
field. He has more than 300 research papers published in the field of machine learning and
advanced mechanics. He is a beneficiary of esteemed honors like ICML Best Paper Award,
IJCAI Computers and Thought grant, ACL Best Paper Award and many more to fill the list.

2. MICHAEL I JORDAN #

Michael I Jordan, one of the most influential figures of the ML world, is presently working as a
Professor at UC Berkeley, and was a former Professor at MIT. His instructions, which are quite
similar to his examination advantages, are divided among Statistics and EECS. Statistics,
artificial intelligence, and deep learning are the areas that attract him most to explore with.

Back in the 1980s, Jordan began creating recurrent neural networks as an intellectual model.
Lately, his work is less determined from a psychological viewpoint and more from the
foundation of conventional measurements. He assisted his work with advocating the utilization
of Bayesian organizations in AI applications, and is frequently credited as one of the guideline
scholars who brought the cover among measurements and AI to well known consideration. He is
an individual of ACM, ISBA, CSS, AAAI, ASA, IEEE, IMS and SIAM. His alumni and postdoc
understudies have also gone on to significantly impact the universe of ML, a few of whom show
up on this rundown Andrew Ng, David Blei and Zoubin Ghahramani. He recently completed an
enlightening Reddit AMA. Moreover, he has also been honored by the Institute of Mathematical
Statistics with the title of Medallion Lecturer and Neyman Lecturer.

Statistics, the subfield of mathematics, is such a subject which plays a dominant role in every
step of Machine Learning and to add to your knowledge, let me tell you a huge number of
Statisticians are present who did really unexceptional experiments in this field. Machine
Learning deals with building different algorithms, systems, and models that learn from structured
data with the help of statistics, instead of normal programming instructions. Statistics has
ultimately proved itself to be the biggest game changer in the growth of a business in this 21st
century. A broad topic to be understood in depth on how the process of Statistical Analysis
actually works for Machine Learning on business case scenarios. To help you out learn more,
check out the blog on Statistical Analysis.

3. KIRILL EREMENKO #

The organizer and CEO of SuperDataScience, an online instructive gateway for Data Scientists,
Kirill Eremenko is a well-known data science entrepreneur whose central goal is to simplify
complex problems with the vision of turning into the greatest learning entrance for Data Science
fans.

Additionally, the online portal highlights many investigation courses going from apparatus based
courses, for example, Python, R Programming, Tableau to complete tutorials for Machine
Learning A-Z and prologue to Data Science. Some of his best course structure education by
Kirill includes Deep Learning, a detailed guide from A-Z and the Hands-On Artificial Neural
Networks. He is also a Udemy Instructor holding the best reviews among all.

This was my first openness to profound learning and trust me, their course is genuinely stand-out
with incredible accentuation on the instinctual level of comprehension with involved coding
instructional exercises on Supervised and Unsupervised Deep Learning. What's more, as of late,
I got the opportunity to peruse his book , Confident Data Skills that assists the readers with
understanding the total information science work process from start till the finish of ventures, all
with zero level of coding.

4. SEBASTIAN THRUN #

Another exceptionally famous name on our rundown, Sebastian Thrun is presently the CEO of
Kitty Hawk Corporation and the Co-founder of the widely used online course platform, Udacity.
Sebastian is an entrepreneur from Germany who established the Google X Lab and Google’s
self-driving group. He drove the task from the beginning and is broadly viewed as a pioneer with
regards to self-sufficient vehicles. Google’s self-driving car project is presently headed by him.
He has fostered various self-governing mechanical frameworks in his profession.
His understanding of AI is that people won’t get replaced by the applications of AI but rather it
would add to our capabilities.

He is well-experienced in the teaching profession too as he is working as the Adjunct Professor
at both Stanford University and Georgia Tech. He is profoundly incorporated into the scholastic
side of machine learning also, as we would anticipate from an individual of Sebastian’s height.
The Popular Science magazine named him as Brilliant 5 back in 2005. He has additionally been
granted the Max-Planck-Research Award in 2011.

5. ANDREAS KRETZ #

At whatever point we talk about information science, the vast majority will in general think that
it's tied in with building cool AI models and making great expectations. In all actuality, building
models is only a piece of the entire work process and information designing (otherwise known as
handyman of information science) is the urgent piece of this work process that upholds
information science projects. Without a steady and strong information designing pipeline and
stage, it'll be a difficult opportunity to try and get information to play out any examination at the primary spot.

Andreas is genuinely a specialist on his own with regards to information designing and
developing large information stages to help information science projects.

He is a data engineer and information science stage designer, building information science stages
that cycle and investigate crazy measures of information consistently. He is an information
designing evangelist situated in Germany, who consistently shares information designing
instruments, methods, and abilities regularly through live recordings that cover themes like the
broadness of occupation alternatives in information science to ongoing Apache Spark coding
meetings.

6. TERRY SEJNOWSKI #

The most popular teacher at the University of California, San Diego, belonging to the Biological
Sciences department as well as the Francis Crick Professor at the Salk Institute. He is also an
Investigator at the Howard Hughes Medical Institute, and one of the pioneers of Neural
Networks. His spearheading commitments to the field of AI incorporate his co-creation of
Boltzmann machines with Geoffrey Hinton.

On account of his work in demonstrating and figuring mind function, he is one of just ten living
researchers to have been chosen for every one of the three public foundations (Medicine, Science
and Engineering) from back in the 1980's. He is presently a counsel to Obama’s $100 million
BRAIN drive, which grows new instruments for planning neural circuits. Terry is known to the
world as the creator of the calculation for Independent Component Analysis that has been
broadly utilized in AI and sign preparation. He is the recipient of the Hebb Prize by the
International Neural Network Society in 1999 for his dedication and commitments to learning
calculations. Indeed, in the year 2002 he was also awarded IEEE’s Neural Network Pioneer
Award. In 2017, he was chosen for the National Academy of Inventors.

7. GEOFFREY HINTON #

It’s unbelievably difficult to summarize the profession of any of these phenomenal personalities
in a couple of sentences, however with Hinton, this demonstrates especially testing. Thirty years
prior, Hinton was at that point positively shaping profound learning, co-creating Boltzmann
machines, backpropagation, and contrastive uniqueness. However, it wasn't until the registering
power figured out how proportional to satisfy the needs of profound discovery that Hinton really
started to get the more extensive acknowledgment outside of the scholarly community he
merited.

In 2004, he helped to establish Neural Computation and Adaptive Perception, a handpicked,
welcome just gathering of specialists from across the fields of physical science, neuroscience and
designing. He additionally established DNNResearch, which was procured by Google a year ago.
From that point forward, he's been dealing with the purported Google "Cerebrum" neural organization venture, and serving to significantly improve Google's picture
acknowledgment and Android's sound acknowledgment abilities.

8. ZOUBIN GHAHRAMANI #

A prominent Professor and researcher of Information Engineering at Cambridge University,
Ghahramani is an distinguished individual belonging to the Adjunct Faculty at the Gatsby
Computational Neuroscience Unit at UCL. He has added to the fields of Bayesian ways to deal
with Artificial Intelligence, man-made reasoning, insights, data recovery, bioinformatics, and
computational engine control. He completed his doctorate from the University of Massachusetts
Institute of Technology at the Department of Brain and Cognitive Sciences , under the
supervision of Tomaso Poggio and Michael I. Jordan.

He started a startup in the year 2014 named as Geometric Intelligence that focuses on item or
situation acknowledgment. In the later years, Uber started implementing Geometric Intelligence
and he finally joined Uber's A.I. Labs in 2016. Publishing more than 250 research papers in the
domain of Machine Learning, he was chosen Fellow of the Royal Society (FRS) in the year
2015.

He was as of late granted $750,000 by Google for The Automatic Statistician, a task driven by
Ghahramani which investigates an open-finished space of conceivable factual models to find a
decent clarification of the information, and afterward creates an itemized report with figures and
normal language text.

9. YANN LECUN #

A French computer scientist working mostly in the field of mobile robotics, computer vision,
machine learning, Yann LeCun and his amigos concocted the convolutional neural networks that
make picture acknowledgment work. In his professional life, Yann holds the position of Director
of AI Research at Facebook.

Yann centers around AI and its applications, including vision, discourse, language, information
mining, and bioinformatics. Likewise, Yann considers PC vision, portable advanced mechanics,
and computational neuroscience.

On following him on Twitter for 45-minute narratives on profound gains from Radio Canada in
French, he says these are astonishing, yet he has his own questions. There are many intriguing
articles about the eventual fate of assembling, for example, that are written in English. Yet,
there's a great deal of French.

10. JAMES CHAM #

I believe that from our viewpoint, the messy mystery around AI right now is that no one
understands what they are doing. James Cham keeps it genuine.

James Cham is a Partner at speculation firm Bloomberg Beta and a beginning phase VC fixated
on the new universe of work. He works really hard scattering a few fantasies about AI and
furthermore collapsing a portion of the publicity.
One can cherish his point about how everybody is centered around the central issues like what
will machine learning mean for the fate of work, yet nobody is zeroing in on the bare essential
execution stuff, which is the thing that's needed for AI to affect the fate of work. It's very much an objective of my blog to help fill in those holes and clarify how organizations can really utilize ML in the present time and place to improve their organizations.

Another extraordinary point he made on the Too Embarrassed to Ask digital recording is that the
second that we sort something out, it simply turns into another element. Sorting out some way to
toast bread was outrageously troublesome, Cham said. Also, individuals discussed a robot that
would really sort out precisely how to get bread perfectly. Also, the second it's feasible, it's a toaster oven, and I imagine that is valid about a great deal of innovation. He doesn't compose or Tweet a ton however when he gives a meeting, it merits your time.

Author Bio #

Senior Data Scientist and Alumnus of IIM- C (Indian Institute of Management - Kolkata) with over 25 years of professional experience Specialized in Data Science, Artificial Intelligence, and Machine
Learning. PMP Certified ITIL Expert certified APMG, PEOPLECERT and EXIN Accredited Trainer for all modules of ITIL till Expert Trained over 3000+ professionals across the globe Currently authoring a book on ITIL "ITIL MADE EASY".

Conducted myriad Project management and ITIL Process consulting engagements in various organizations. Performed maturity assessment, gap analysis and Project management process definition and end to end implementation of Project management best practices Social Profile Links.

Twitter account URL- https://twitter.com/ramtavva?s=09
Facebook Profile URL- https://www.facebook.com/ram.tavva
Linked In Profile URL https://www.linkedin.com/in/ram-tavva/

4 Ways to Prepare for the AP Computer Science A Exam

The AP Computer Science A is a course and exam offered to high schoolers as an option to earn an equivalent of a Computer Science college semester. It essentially covers the basics of object-driven programming paradigm based on the Java programming language, such as problem detection and solution, creating algorithms, and programming concepts. It also introduces data ideas and abstraction procedures.

If you want to jumpstart your college degree in Computer Science, this article will give a few pointers on how to best prepare for the exam.

Training Program for AP Computer Science A Exam #

The training program for the AP Computer Science A exam should be simple, straightforward, and tackle such areas of computer programming as:

Computing: at the very basic level, you need to establish connections between concepts in computing
Abstractions: learn how to apply abstractions, one of the key concepts in object-oriented programming, in modeling and computation
Communication: understand how to communicate ideas on technology and computation Skills application: design a program that solves a real-life problem or performs a task
Analysis: learn to analyze computational work; it’s essential for picking up on someone else's code or mistakes in your own
Collaboration: learn how to work in a team; it’s a must for any programmer but also helps to study and become a professional

You get more information from the article, "Where to Learn and How to Pass the AP Computer Science A Exam"

How to Prepare for the AP Computer Science A Practice Test? #

According to the official data from the College Board, about 70% of students achieved a passing score back in 2020. That’s a good percentage meaning that it’s relatively easy to achieve a passing grade. It all depends on how much you want to get it and how much effort you’re willing to put into the preparation.

Here are a few ideas on how to prepare for the AP Computer Science A exam.

1. Learn the course framework and the exam structure #

You can check out the College Board’s web page to learn about the course and exam structure, framework, and guidelines. It also notes the percentage of the grade you can get for each exam unit.

In brief, the AP Computer Science A exam duration is three hours; it has two sections both taking half of the time frame and being equally judged for the exam results scoring:

Free-response. Here you’ll have just four questions but they need broad answers.
Multiple choice. Here you need to pick the right answer in forty questions.

Note that there’s another one-semester equivalent course and exam – AP Computer Science Principles. This exam has two parts:

Multiple choice. Here you have two hours to answer 70 questions.
Performance task. Here you get at least 12 hours to create a program and give a written and video response.

2. Get familiar with the topics included in both AP Computer Science Exams #

The training program for the AP Computer Science A Exam is quite simple in that it covers the basic skills in programming since the course itself requires only prior knowledge of English and algebra. First of all, you can cover the primary levels of the course which are:

For the A exam:

Writing Classes
Types and Objects
Arrays
Boolean Expressions
Recursion
Inheritance

For the Principles exam:

Algorithms in Programming
Data
Creative Development
Computing Systems and Networks
Impact of Computing

3. Make the preparation process more fun with CodeGym #

Since AP Computer Science includes Java programming, CodeGym is a good option to fit into your preparation routine. The platform offers 1200+ tasks to help you learn object-oriented programming concepts and practice the Java language.

CodeGym can make your training fun and entertaining because of the gamified environment of learning and a lot of practical tasks. In fact, 80% of the Java course is practice-based.

At the primary level, the course covers the following topics:

Boolean expressions, if statements, and iteration
Primitive data types and using objects
Introduction to objects: writing objects, their lifetime, etc
Writing classes and using constructor
Basics of OOP
Arrays and ArrayList

4. Watch AP Computer Science videos on YouTube #

You can seek help on YouTube by watching AP Computer Science tutorial videos. A lot of software engineers and educators have their own channels and gladly share their knowledge.

A few good examples are:

The College Board channel
CS50 lectures
From the Advanced Placement channel
- 33 video lectures on AP Computer Science A
- 5 video lectures on AP Computer Science Principles

Conclusion #

Training for the AP Computer Science Exams can be easier if you have a solid plan – a training program. To build this plan, first learn the knowledge and skills the exams require, i.e. what units the exam consists of and the exam structure.

Then, study the required material and look up some videos on YouTube to learn about the real-world experience and programming skills.

You can also boost your training sessions with a practical course that offers hands-on experience with coding.

Finance Assignment Online Help for the Busy and Tired Students: Get Help from Experts

Handling finance homework can be a burden too heavy to carry for many students. It can be boring. It can contain long collections of problems in one, as well as gradually increase your awareness of the key concepts and ideas throughout the course.

Financing homework also requires maximum attentiveness and good concentration. But here is the point we want to make – even if you consider yourself an absent-minded person and lack the ability to focus, any assignment in finance is solvable on the condition that you have the desire and a little bit of patience for working at a writing task. One of the tried and true ways to get reliable assistance in this is to partner with academic writing services such as APlusEssay.com. On https://aplusessay.com/finance-assignment-help.html, you will get finance assignment help and be able to take something from experts into your own writing.

Now, the pieces of advice that we are going to give here might not be radically different from tips given for solving homework in other disciplines. But still, if you take them seriously and try them out in your own studying routine, fruits will soon appear. Will it be effortlessly? Probably, not. But it can teach you a few valuable lessons.

So, here we go with the tips and tricks to handle college tasks in finance. Check them out!

Tip #1. Make the most of technology. #

How lucky we are today that there is no need for us to count figures in every task in our brains. Now, most of the work is done by mobile applications and software designed to make mathematical operations easier and involve the human only at several stages. Apps such as iMathematics Pro, MathRef, and Digits would be great solutions for much of your finance homework and allow you to save time.

Tip #2. Get a perfect workplace. #

The environment in the place where you usually study determines how successful you will solve all the necessary assignments and how fast you will do it. But here there are no universal formulas since every one of us has got their own preferences and learning styles. Some are used to studying with some background music on, some need total silence. Others try to keep a perfect order on the table while others (‘creative’ people) can tolerate mess. Whatever you choose, try to make this place inspiring, where it will be easy to focus on the assignments and devote all of you to work.

Tip #3. Understand the task. #

Too simple, as some would say? Still, realizing what particularly is required from you in a certain task in finance studies is the good half of success. Specify what you didn’t get at first, consult a teacher, use the help of reliable service for finance assignment help online. This advice has endless benefits – you skip doing tons of extra work and don’t get to lose points, focus on the smaller parts of the learned material, and do not get distracted. So, better get a full knowledge of what to do before you actually write anything.

What concerns professional help, today it’s normal to turn to good services when there is no time for solving all college tasks, especially if the subject is not your major. All you have to do to get finance assignment help is to write a message like ‘Could you do my finance task for me, please?’ and pick the author for writing this academic paper. Yes, the best thing is to solve finance homework on your own but APlusEssay.com, for instance, can be a good Plan B and an online support team that can deal with any type of academic work – coursework, review, research proposal, equations, and much more.

Tip #4. Set the realistic timing. #

There are times when you have all deadlines set in stone. But it often depends on us how soon we are ready to submit the paper and rest in peace. As one of the most effective (yet, most neglected) ways of coping with academic assignments, planning has its special place and has the power to make learning stress-free. Imagine: you are confident of your progress, divide the assignment into do-able parts, and live a blameless student’s life instead of working by fits and starts and still not doing enough.

The skill to manage your time and create plans is a valuable one so why not train it while doing finance homework?

Tip #5. Create a reward stimulus. #

No, this is not about a Twix at every second page read. But it can still work if you create smart rewards – having fun with friends the whole next day or doing practically anything that motivates you. In reality, the fact that you managed to fulfill the promise o do a certain part of the work by the assigned time is already a good reward. But the main point about this system is feeling good about what you have accomplished – creating pleasurable associations and memories will result in even more inner motivation for you to finish it.
As you can see, the given tips aren’t hard to bear. If you haven’t thought of tips to help you deal with finance homework, we graciously give them to you for free and enjoyable use!

Get Finance Assignment Help Today and Never Regret It: Learning from Expert Writers Is Worth It #

As a kind of afterword, we point we wanted to make about using any service for finance assignment help – whether it is APlusEssay.com or any other decent place – is the opportunities for learning you have there. In the end, all the grades at college will be forgotten. Diplomas will be opened rarely in the lifetime. But what will truly remain with you is the skills. The experience.
So, don’t run for cheap academic help – use authentic samples, check the credibility of sources, and take after the best writers in the world such as those who are available on APlusEssay.com.

Top 9 Machine Learning Algorithms for Data Scientists

For most of the beginners, the machine learning algorithms seem to be uninteresting or boring to learn further about. Up to some extent, this stands true but in many cases, you might stumble upon a few-page description for each algorithm and it can turn out to be time-consuming in order to figure out each and every detail. If you have a strong desire to become a Machine Learning expert then you actually need to brush up your knowledge related to it as there is no other way around. In this article, we will be talking about the 8 most common algorithms in simple words including a brief overview along with some useful links. So, let's begin!

#1 Conditional Random Fields (CRFs) #

This algorithm is used to simulate a sequence of an RNN which can be used in conjunction with an RNN algorithm. They can also be used in various tasks of structured prediction as in image segmentation. The algorithm models each element of the sequence so that the neighbors are affecting the component labels in the sequence and not all the labels are independent of each other. You can use CRF for the sequence of sequences in text, image, time series, and DNA.

Useful websites:

sklearn-crfsuite

Detailed Guide:

YouTube playlist on conditional random fields
Introduction to Conditional Random Fields

#2 Convolutional Neural Networks #

In practical, all the modern achievements in the field of machine learning can be achieved by dint of convolutional neural networks that are used for image classification, image segmentation, and object detection. It was invented by Jan Lekun in the early 90s, networks have convolutional layers that act as hierarchical object extractors. You can use it to work with text and even for working with graphics.

Useful websites:

Detailed Guide:

A Beginner’s Guide To Understanding Convolutional Neural Networks

#3 Principal Component Analysis (PCA)/SVD #

This is one of the most important machine learning algorithms which allows you to reduce the dimension of the data by losing the amount of information. The algorithms are used in multiple areas like object recognition, computer vision, data compression and much more. The computational of the principal components are reduced to calculating the eigenvalues and eigenvectors of the covariance matrix of the original data and to the singular decomposition of the data matrix. You can express several signs though one, merge, speak and work already with a simpler model. Of course, it is not possible to avoid information loss but the PCA method will help you to minimize that. SVD is the way to calculate the ordered components.

Useful websites:

scipy.linalg.svd

Detailed Guide:

A Tutorial on Principal Component Analysis

#4 Decision Trees #

This is one of the most common machine learning algorithms which is used in statistics and data analysis for predictive models. The structure represents the leaves and branches. Attributes of the objective function widely depend on the branches of the decision tree where the values of the objective function are stored in the leaves and the remaining nodes contain attributes for which the cases differ. In order to classify a new case, you need to go down the tree to the leaf and give the appropriate value. The ultimate goal is to create a model that predicts the value of the target variable based on several input variables.

Useful websites:

Detailed Guide:

Understanding Random Forests: From Theory to Practice

#5 Feed-Forward Neural Networks (FFNN) #

In general, there are unit-level logistic regression classifiers. Multiple layers of scales are separated by non-linearities like sigmoid, tanh, cool new selu and relu + softmax. They are also known as multilayer perceptrons as FFNN can be used to classify and learn without a guide as autoencoders. The algorithm can be used to train a classifier or extract the functions as autoencoders.

Useful websites:

Detailed Guide:

#6 K-Means Clustering #

This is everyone's favorite uncontrolled clustering algorithm. The k-means algorithms are the simplest but inaccurate clustering method in its classical implementation. The method splits the set of elements of a center space into a previously known number of clusters k. This algorithm minimizes the standard deviation at the points of each cluster. Here, the basic idea is that at each iteration the center of mass is recalculated for each cluster that you get in the previous step. Then, the vectors are divided into the clusters again according to which the new centers were closed in the selected metric.

Useful websites:

sklearn.cluster.KMeans

Detailed Guide:

#7 Logistic Regression #

It is limited to linear regression with non-linearity after applying weights and hence the output limit is close to 1 and 0 in case of the sigmoid. Also, the cross-entropy loss functions are optimized using the gradient descent method. Logistic regression is used for classification and not regression. It is the same as a single-layer neural network and when learned using optimization techniques like gradient descent or L-BFGS, it is called 'maximum entropy classification method'.

Useful websites:

sklearn.linear_model.LogisticRegression

Detailed Guide:

Logistic Regression | Classification

#8 Support Vector Machines (SVM) #

It is a linear model similar to linear and logistic regression. The difference between them is that it has a margin-based loss function which you can optimize the loss function by using optimization methods i.e. SGD or L-BFGS. The most unique thing that SVMs can do is to study classifier classifiers as it can be used to train the classifiers and even regressors.

Useful websites:

sklearn.svm.SVC

Detailed Guide:

#9 Recurrent Neural Networks (RNNs) #

This model sequences by applying the same set of weights at time t and input at time t of the aggregator recursively. Pure recurrent neural networks are rarely used now but it is analogs. For instance - LSTM and GRU are the most up to date in most sequence modeling problems. LSTM is used instead of a simple dense layer in pure RNN. Use this network for any task of text classification, machine translation, and language modeling.

Useful websites:

Detailed Guide:

Deep Learning for Natural Language Processing
RNN Articles

The Road Ahead #

It may sound unpleasant but you have to try a lot of different algorithms and check their effectiveness by applying on each of the test data set to choose the best option. It is an obvious task that you need to choose from the algorithms which match your task. So be patient and do not try to rush things. Keep Learning!</span>

Author Bio:

Ava Barker working as a Technology Consultant at Tatvasoft UK. Coming from a technology background she likes to share her insights about development, design and more. She has also published her author bylines on many different publications online.

Data Science Learning Path or Steps to become a data scientist Final

There is a massive shortage of quality data scientists on the market right now (Source). IBM predicts the market demand for data scientists to rise by 28% increase by 2020. That means there is a lot of room left for data scientists.

We produce 2.5 quintillions of data, and that number will increase with the advancement in IoT. The current world needs more professionals who can play around with data. The corporations do not utilize their collected data due to lack of skilled data science workforce who can extract valuable insights from the massive pool of information.

The cost of wrong data is $3 trillion annually in the United States. On the other hand, the smart use of collected data can significantly enhance the revenue and productivity of an organization (Source).

Image Source: Statista

The average pay of data scientists surpasses $115k annually (Source). Do you want to fill the gap between demand and supply? In this article, we will reveal steps that can help you be a skilled data scientist.

Steps to come out as a data scientist #

The data scientists are responsible for extracting, filtering, and interpreting data in a meaningful way with the use of various statistical and machine learning tools and techniques. Here are the steps that you need to follow to be a data scientist:

Do you meet the basic requirements to apply for a job?

The corporations who are looking to hire data scientists need some degree. Here is one of the posts of data scientist vacancy.

Image Source: Indeed

Most of the jobs on the market require you to at least have a bachelor’s degree in a relevant field like computer science, statistics, applied mathematics, and so on. You will be eligible to apply for the majority of job posts if you have a related degree. If you keep an eye on the pay scale, you will see that even a beginner position surpasses $95k annually.
Understand about various aspects of data science

It is crucial to know what you’re going to experience in advance on the road to becoming a data scientist. Do some research about the duties and responsibilities of data scientists. After that, you need to select the tools and programming language. Currently, R and Python are the two preferred programming languages in the field of data science. I would recommend Python training due to its ease of learning and flexibility.

Image Source: Stack Overflow
Get some knowledge of basic mathematics and statistics

You must have a core knowledge of basic mathematics and statistics to become an outstanding data scientist. It is possible for you to use the tools for performing calculations. However, you will not be able to fulfill your responsibilities efficiently without a strong foundation in both mathematics and statistics.

The data scientists need to frequently use statistic methods (descriptive and inferential analysis) and numerical solutions (linear algebra) in their jobs. You should also be keen to solve problems, applying logic because that is what you are going to do in your career.
Time for some essential machine learning

There is a considerable role of machine learning in data science. The knowledge of machine learning can help you accomplish great things in your profession. At first, you should only get the necessary know-how of machine learning to be able to perform more in-depth applications of it in your job.

The Python programming language is also prevalent for performing machine learning applications. Many professionals prefer implementing machine learning with Python. Due to the full range of applications, it is better to learn Python. You can jump into web applications, game development, and more with Python.
Dig deeper into machine learning applications

After you have a strong foundation of machine learning, you can dig deeper into machine learning applications that you can apply in your job. You can learn some advanced applications, such as boosting algorithm, time series methods, random forest, ensemble learning, and so on. Right after algorithm, you should know how to improve models and bring new predictive models with tricks, such as validation strategies along with feature engineering.

You can use machine learning for producing accurate results and use data-driven models for processing data in real time.
Get some concepts of deep learning

After gaining a good grasp of machine learning, it is a time for you to get into deep learning. Deep learning is an integral part of machine learning. Having a solid knowledge of deep learning can significantly increase your value as a data scientist. At first, you should only get a basic understanding of deep learning, such as a neural network. Also, get your feet wet with Keras.

Image Source: Machine Learning Mastery
Dig deeper into advanced deep learning applications

After you have some strong foundation of machine learning and deep learning, you can now jump into advance neural network frameworks. The neural network frameworks to focus on are convolutional and recurrent neural networks. You will need some time to understand those concepts. I suggest you be patient while learning these advanced deep learning concepts. You may need to invest a few weeks of your time in acquiring them.
Learn computer vision applications

Computer vision application is a trending technology. It has touched different industries like manufacturing, retail, construction, and so on. Some of the examples of these technologies are face detection technologies like MobileNet, object recognition technology like ResNet, and so on.

The data scientists can use these technologies to extract useful insights from images, objects, and so on. You will get more tasks as a data scientist where the computer vision applications will come handy.
Natural Language Processing

Finally, you need to learn about natural language processing to sharpen your skill as a data scientist. You must at least have a basic understanding of the concepts, such as text preprocessing and text classification.

How can I get started learning all these concepts and skills? #

We have broken down the steps to become a data scientist. However, some of you might be confused on how to start learning all the things that I have mentioned above. The task of learning is not difficult in this day and age. You have information in your fingertips. There are plenty of quality resources that you can access for free or for an affordable price.

The market size of the eLearning industry is expected to rise to $325 by 2025. The only thing you need to do is find appropriate articles, videos, and courses to enroll and start learning.

Over to You #

If you have reached this portion of the article, you must be clear about what to do next. You need to follow the steps that I have mentioned to be a data science professional. Every single step takes a lot of time and effort.

Before jumping into the field of data science, you must decide if you want to be a data scientist or not. Do not dive into data science only because of a lucrative pay package. If there is something that you want to add, feel free to leave your opinion in the comment section. I would love to know your thoughts.

Enable Edit Button in Shutter In Linux Mint 19 and Ubuntu 18.04

Shutter is a screen capturing software. It allows you to take screenshot of a specific area, window and even the entire screen.

Shutter also comes with a few editing tools, however these tools are not enabled by default. To enable these tools we'll have to install some additional packages.

All the instructions mentioned in this post is tested on Linux Mint 19 and Ubuntu 18.04.

Let's start!

Enabling the Edit button #

As you can see in the above image the Edit button is disabled. To enable the Edit button, you need following three packages:

libgoo-canvas-perl
libgoocanvas-common
libgoocanvas3

We can download these packages from the following links:

If you prefer command line, you can download all these packages via the wget command as follows:

wget -q http://mirrors.kernel.org/ubuntu/pool/universe/g/goocanvas/libgoocanvas-common_1.0.0-1_all.deb
wget -q http://mirrors.kernel.org/ubuntu/pool/universe/g/goocanvas/libgoocanvas3_1.0.0-1_amd64.deb
wget -q http://mirrors.kernel.org/ubuntu/pool/universe/libg/libgoo-canvas-perl/libgoo-canvas-perl_0.06-2ubuntu3_amd64.deb

Once you have downloaded all the necessary packages, install them using the following command:

sudo dpkg -i libgoocanvas-common_1.0.0-1_all.deb ; sudo apt-get -f install ; sudo dpkg -i libgoocanvas3_1.0.0-1_amd64.deb ; sudo apt-get -f install; sudo dpkg -i libgoo-canvas-perl_0.06-2ubuntu3_amd64.deb ; sudo apt-get -f install

Now kill any any running instance of Shutter and then start it again.

sudo killall shutter

You should now have Edit button enabled.

Enabling Edit button on Linux Mint 17, 18 or Ubuntu 16.04 #

To enable Edit button on Linux Mint 17, 18 or Ubuntu 16.04, the fix is comparatively simple. All you need to do is install libgoo-canvas-perl using the following command:

sudo apt-get install libgoo-canvas-perl

Enabling the Web button #

Just like the Edit button, the Web button is also disabled by default. The Web button allows us to take a screenshot of a website by providing a URL.

To enable the Web button we'll need a package called gnome-web-photo. Download gnome-web-photo from the following link: [

https://launchpad.net/ubuntu/+source/gnome-web-photo/0.10.6-1/+build/5854031](!https://launchpad.net/ubuntu/+source/gnome-web-photo/0.10.6-1/+build/5854031)

Or simply, execute the following command to download the gnome-web-photo package.

wget -q https://launchpad.net/ubuntu/+source/gnome-web-photo/0.10.6-1/+build/5854031/+files/gnome-web-photo_0.10.6-1_amd64.deb

Once the package is downloaded install it by typing:

sudo dpkg -i gnome-web-photo_0.10.6-1_amd64.deb
sudo apt-get -f install

Kill any running instance of Shutter and start the application again. The Web button should be enabled now.

Restart the shutter and now you should be able to take screenshots of websites.

Next, we will see how to enable the Web button on shutter.

Enabling the Web button on Linux Mint 17, 18 or Ubuntu 16.04 #

To install gnome-web-photo on Linux Mint 17, 18 or Ubuntu 17.06, simply execute the following command:

sudo apt-get install gnome-web-photo

That's all there is to it.

Let's us know how this tip worked out for you in comments.

Python 3 time module

The time module provides a number of time-related functions. Most of the functions found in this module call the platform C library functions behind the scenes with the same name.

To import the time module enter the following command:

>>>
>>> import time
>>>

Unix Epoch Time #

Most modern computers store dates and times as a difference of seconds between January 1, 1970 00:00:00 UTC and the time to be stored (also in UTC). We call this difference Unix timestamp or Unix epoch time or simply timestamp. For example, date and time "11 July 2018 22:52:42" in UTC can be represented as timestamp 1531329762.

In the following sections, we discuss some common attributes and functions of the time module.

time() function #

Syntax: time() -> floating point number

The time() function returns the number of seconds passed since January 1, 1970 00:00:00 UTC as a floating point number.

>>> 
>>> time.time()
1531326753.9201949
>>>

localtime() function #

Syntax: localtime([seconds]) -> (tm_year,tm_mon,tm_mday,tm_hour,tm_min, tm_sec,tm_wday,tm_yday,tm_isdst)

Converts the specified Unix timestamp to local time. If the seconds argument not passed in, it uses timestamp returned by the time() function. This function returns a time.struct_time object with the following attributes.

Attribute	Description
`tm_year`	year
`tm_mon`	month of year (1-12)
`tm_mday`	day of month (1-31)
`tm_hour`	hour (1-23)
`tm_min`	minutes (0-59)
`tm_sec`	seconds (0-59)
`tm_wday`	day of week (0-6) (Sunday = 0, Monday = 1 and so on)
`tm_yday`	day of year (0-365), also known as julian day
`tm_isday`	daylight saving flag (0, 1 or -1). A value of `1` indicates that the daylight savings is in effect, 0 if daylight savings is not in effect and `-1` if the information is not available.

The time.struct_time object is also commonly known as timetuple.

>>>
>>> lt1 = time.localtime(1200000000)     # convert the timestamp 1200000000 to local time
>>> lt1
time.struct_time(tm_year=2008, tm_mon=1, tm_mday=11, tm_hour=2, 
tm_min=50, tm_sec=0, tm_wday=4, tm_yday=11, tm_isdst=0)
>>> 
>>>
>>> lt2 = time.localtime()    # convert timestamp returned by time() to local time
>>> lt2
time.struct_time(tm_year=2018, tm_mon=7, tm_mday=11, tm_hour=22, 
tm_min=3, tm_sec=3, tm_wday=2, tm_yday=192, tm_isdst=0)
>>>
>>>

Once you have access to timetuple you can access values inside it either by index or attribute name. For example:

>>>
>>>
>>> lt2.tm_year, lt2.tm_mon, lt2.tm_mday 
(2018, 7, 12)
>>> 
>>> 
>>> lt2[0], lt2[1], lt2[2]   # same as: lt2.tm_year, lt2.tm_mon, lt2.tm_mday
(2018, 7, 12)
>>> 
>>>
>>> lt2.tm_hour, lt2.tm_min, lt2.tm_sec 
(0, 3, 40)
>>> 
>>>
>>> lt2[3], lt2[4], lt2[5]   # same as: lt2.tm_hour, lt2.tm_min, lt2.tm_sec
(0, 3, 40)
>>>
>>>
>>> lt2.tm_wday, lt2.tm_yday, lt2.tm_isdst 
(3, 193, 0)
>>> 
>>> 
>>> lt2[6], lt2[7], lt2[8]    # same as: lt2.tm_wday, lt2.tm_yday, lt2.tm_isdst
(3, 193, 0)
>>>
>>>

gmtime() function #

Syntax: gmtime([seconds]) -> (tm_year,tm_mon,tm_mday,tm_hour,tm_min, tm_sec,tm_wday,tm_yday,tm_isdst)

Works exactly like localtime() but the converts the specified timestamp to UTC instead of local time. If the seconds argument not passed in, it uses timestamp returned by the time() function.

>>>
>>> 
>>> gm1 = time.gmtime(1200000000)   # convert timestamp 1200000000 to UTC time
>>> gm1
time.struct_time(tm_year=2008, tm_mon=1, tm_mday=10, tm_hour=21, 
tm_min=20, tm_sec=0, tm_wday=3, tm_yday=10, tm_isdst=0)
>>> 
>>>
>>> gm2 = time.gmtime()    # convert timestamp returned by time() to UTC time
>>> gm2
time.struct_time(tm_year=2018, tm_mon=7, tm_mday=12, tm_hour=8, 
tm_min=5, tm_sec=38, tm_wday=3, tm_yday=193, tm_isdst=0)
>>> 
>>> 
>>> gm2.tm_year, gm2.tm_mon, gm2.tm_mday, gm2.tm_hour, gm2.tm_min, gm2.tm_sec
(2018, 7, 12, 8, 22, 36)
>>>
>>>
>>> gm1[0], gm1[1], gm1[2], gm1[3], gm1[4], gm1[5]
(2008, 1, 10, 21, 20, 0)
>>> 
>>>

mktime() function #

Syntax: mktime(timetuple) -> floating point number

This is the opposite of localtime() function. It accepts a timetuple representing local time and returns the number of seconds passed since January 1, 1970 00:00:00 UTC.

>>>
>>>
>>> timestamp, timetuple = time.time(), time.localtime()
>>>
>>> 
>>> timestamp, timetuple
(1531384941.615937, time.struct_time(tm_year=2018, tm_mon=7, tm_mday=12, 
tm_hour=14, tm_min=12, tm_sec=21, tm_wday=3, tm_yday=193, tm_isdst=0))
>>>
>>> 
>>> time.mktime(timetuple)
1531384941.0
>>> 
>>>

sleep() function #

Syntax: sleep(seconds)

This function suspends the process execution for the specified number of seconds.

>>>
>>> 
>>> for i in range(5):
...     time.sleep(5)
...     print(time.time())
... 
1531391820.9281697
1531391825.9332898
1531391830.9384072
1531391835.9435263
1531391840.9486341
>>> 
>>>

If you run the preceding code interactively, you will find that the for loop prints the timestamp after every 5 seconds.

ctime() function #

Syntax: ctime([seconds]) -> string

Converts the timestamp to a string representation in local time. If seconds argument not specified or None, it uses timestamp returned by time().

>>>
>>> 
>>> time.ctime()
'Thu Jul 12 16:18:51 2018'
>>>
>>> 
>>> time.ctime(None)
'Thu Jul 12 16:18:54 2018'
>>> 
>>> 
>>> time.ctime(1290000000)
'Wed Nov 17 18:50:00 2010'
>>> 
>>>

asctime() function #

Syntax: asctime([timetuple]) -> string

Converts the specified timetuple to a string representation in local time. If the timetuple is not specified, it uses current time returned by the localtime() function.

>>>
>>>
>>> time.asctime()
'Thu Jul 12 16:57:53 2018'
>>>
>>> 
>>> time.asctime(time.localtime())   # same as time.asctime()
'Thu Jul 12 16:58:03 2018'
>>> 
>>> 
>>> time.asctime(time.localtime(1230000000))
'Tue Dec 23 08:10:00 2008'
>>> 
>>>

strftime() function #

Syntax: strftime(format[, timetuple]) -> string

The strftime() function converts the timetuple to a string representation. It takes two arguments, format and an optional timetuple, and returns string according to the format codes used in the first argument. If timetuple not specified, it uses current time returned by the localtime() function.

The following table lists some common format codes:

Format Code	Description
`%d`	Day of the month
`%H`	hours in 12-hour format
`%I`	hours in 24-hour format
`%M`	minutes
`%S`	seconds
`%p`	A.M. or P.M. (depends upon locale)
`%m`	month number
`%w`	weekday
`%y`	2 digit year
`%Y`	4 digit year
`%Z`	Timezone name
`%x`	Date representation according to locale
`%X`	Time representation according to locale
`%b`	Locale’s abbreviated month name
`%B`	Locale’s full month name
`%a`	Locale’s abbreviated weekday name

>>>
>>> 
>>> time.strftime("%Y/%m/%d")
'2018/07/12'
>>>
>>> 
>>> time.strftime("%H:%m:%d")
'18:07:12'
>>>
>>>  
>>> time.strftime("%a %B %d %Y %H:%I:%S %p")
'Thu July 12 2018 18:06:46 PM'
>>> 
>>> 
>>> time.strftime("%x", time.localtime(1230000000))
'12/23/08'
>>> 
>>> 
>>> time.strftime("%X", time.localtime(1230000000))
'08:10:00'
>>> 
>>>

strptime() function #

Syntax: strptime(string, format) -> struct_time

The strptime() function lets you convert date string to a timetuple according to the format string format. By default, the format argument is set to '%a %b %d %H:%M:%S %Y'. This is the same format as produced by the ctime() function. If the string and format do not match it raises ValueError exception.

>>>
>>>
>>> time.strptime(time.ctime())
time.struct_time(tm_year=2018, tm_mon=7, tm_mday=12, tm_hour=19, 
tm_min=6, tm_sec=3, tm_wday=3, tm_yday=193, tm_isdst=-1)
>>> 
>>> 
>>> time.strptime('11-01-2018', "%d-%m-%Y")
time.struct_time(tm_year=2018, tm_mon=1, tm_mday=11, tm_hour=0, 
tm_min=0, tm_sec=0, tm_wday=3, tm_yday=11, tm_isdst=-1)
>>>
>>>
>>> time.strptime('Mon, 01 Dec 2018', "%a, %m %b %Y")
time.struct_time(tm_year=2018, tm_mon=1, tm_mday=1, tm_hour=0, 
tm_min=0, tm_sec=0, tm_wday=0, tm_yday=1, tm_isdst=-1)
>>>
>>> 
>>> time.strptime('01-Dec-2018', "%m-%b-%Y")
time.struct_time(tm_year=2018, tm_mon=1, tm_mday=1, tm_hour=0, 
tm_min=0, tm_sec=0, tm_wday=0, tm_yday=1, tm_isdst=-1)
>>> 
>>>
>>> time.strptime('01-Dec-2018', "%m-%b-%y")   # string and format do not match, thus ValueError is raised
...
ValueError: unconverted data remains: 18
>>> 
>>>

tzname attribute #

The tzname attribute returns a tuple containing two values: name of the local time zone and name of the local daylight saving time zone (if defined).

>>>
>>> 
>>> time.tzname
('SGT', 'SGT')
>>> 
>>>

As you can see, my default time zone is set to Asia/Singapore. To determine the offset between UTC and local time we use the timezone attribute.

timezone attribute #

The timezone attribute returns the difference between the UTC time and local time in seconds.

>>>
>>> 
>>> time.timezone
-28800
>>>
>>> 
>>> time.timezone/3600    # difference between UTC time and localtime in hours
-8.0
>>>
>>>

This means that the Asia/Singapore time zone is 8 hours (or 28800 seconds) ahead of UTC.

tzset() function #

The tzset() function changes the local time zone to the value stored in the TZ environment variable.

>>>
>>>
>>> time.tzname    # initial time zone
('SGT', 'SGT')
>>> 
>>> 
>>> import os 
>>> os.environ['TZ'] = 'Europe/Moscow'
>>> 
>>>
>>> time.tzset()   # change time zone
>>> 
>>> time.tzname    # local time zone is now changed to Moscow
('MSK', 'MSK')
>>>  
>>>

My default time zone is changed for the session. If you now call the localtime() function, it will return the local time of Moscow.

>>> 
>>>
>>> time.localtime()
time.struct_time(tm_year=2018, tm_mon=7, tm_mday=12, tm_hour=20, 
tm_min=0, tm_sec=27, tm_wday=3, tm_yday=193, tm_isdst=0)
>>> 
>>>

Pygments Tutorial

Pygments is a very popular and robust Python package to highlight code snippets. Many well-known sites like Wikipedia, BitBucket, Read the Docs etc. are using it. Here are some of the features of Pygments:

Supports a wide range of languages and markup formats.
New language and markup can be added easily.
Can use used as library or command line utility
A number of output formats are available, like HTML, RTF, Latex etc.

To install Pygments enter the following command:

$ pip install pygments

Highlighting using Pygments #

To highlight a code snippet using Pygments we follow these steps:

Select lexer.
Select the output format.
Call the highlight() function.

Let's go through each step in detail.

Selecting the lexer #

A lexer is a program which performs the lexical analysis. In other words, it splits the code into tokens (identifier, keyword, literal etc). Every language or markup has its own lexer. To select the lexer use get_lexer_by_name() function from the pygments.lexers package. It takes a single argument named alias which is name of the lexer. For example, to select the lexer for Python do this:

>>>
>>> from pygments import lexers
>>>
>>> lexers.get_lexer_by_name("python")
<pygments.lexers.PythonLexer>
>>>  
>>>

We can see all the available lexers using the get_all_lexers() function:

>>>
>>> list(lexers.get_all_lexers())
[('Transact-SQL', ('tsql', 't-sql'), ('*.sql',), ('text/x-tsql',)), ('Termcap', ('termcap',), ('termcap', 'termcap.src'), ()), ('Snowball', ('snowball',), ('*.sbl',), ()), ('Elm', ('elm',), ('*.elm',), ('text/x-elm',)), ('IRC logs', ('irc',), ('*.weechatlog',), ('text/
...
('pycon',), (), ('text/x-python-doctest',)), ('BBCode', ('bbcode',), (), ('text/x-bbcode',))]
>>>
>>>

Selecting the Formatter #

Once you have selected the lexer, the next step is to select the Formatter. The Formatter's job is to take a token stream from the lexer and writes the output, in the format such as HTML, LaTex, RTF, BBCode etc. To see the available formats use the get_all_formatters() function from the pygments.formatters package.

>>> 
>>> from pygments.formatters import get_all_formatters
>>>
>>> list(get_all_formatters())
[<class 'pygments.formatters.img.JpgImageFormatter'>, 
<class 'pygments.formatters.rtf.RtfFormatter'>, 
...
<class 'pygments.formatters.svg.SvgFormatter'>, 
<class 'pygments.formatters.html.HtmlFormatter'>, 
<class 'pygments.formatters.irc.IRCFormatter'>, 
<class 'pygments.formatters.img.BmpImageFormatter'>]
>>> 
>>>

In our case, we are interested in outputting code snippet in HTML. As a result, we will use, HtmlFormatter (third from the last in preceding output). To use HtmlFormatter, import it from pygments.formatters and then instantiate it as follows:

>>>
>>> from pygments.formatters import HtmlFormatter
>>>
>>> formatter = HtmlFormatter()
>>>

Highlighting Code #

Finally, the last step is to call the highlight() function from the pygments package. The highlight() function puts everything together and returns the highlighted output. It takes three arguments, the code to highlight, lexer and the formatter.

>>>
>>> from pygments import highlight 
>>>
>>> code = """\
... def func():
...     # function body
...     print("hello world!")
... """
>>>
>>> lex = lexers.get_lexer_by_name("python") 
>>>
>>> highlight(code, lex, formatter)
'<div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">func</span><span class="p">():</span>\n\t<span class="c1"># function body</span>\n\t<span class="k">print</span><span class="p">(</span><span class="s2">&quot;hello world!&quot;</span><span class="p">)</span>\n</pre></div>\n'
>>> 
>>>

If you want highlight() function to save the output to a file rather than returning it, use the outfile argument as follows:

>>> 
>>> with open("out.html", "w") as f:
>>>     highlight(code, lex, formatter, outfile=f)
>>> 
>>>

This will create a file named out.html in the current working directory containing the highlighted code.

If open out.html in a browser you will find that code doesn't appear to be highlighted at all. This is because by default HtmlFormatter generates HTML with CSS classes to marks various tokens, but it doesn't generate the actual CSS styles. We will see how to fix this in the next section.

Customizing the Output of HtmlFormatter #

The following is a lists of some common keyword arguments that you can pass to HtmlFormatter() function to customize the generated output:

Argument	Description
`full`	If set to `True`, `HtmlFormatter` generates a self-contained document with inline CSS styles. By default, it is set to `False`
`style`	The color scheme to use. It defaults to `default`.
`linenos`	If set to `True`, tells the formatter to generate output with line numbers. By default, it is set to `False`.
`hl_lines`	Specify the list of lines to be highlighted.
`cssclass`	CSS class for wrapping the code block.

Here is how to use these arguments:

full #

In the previous section we have seen that by default, the output generated by HtmlFormatter() only contains HTML markup and CSS classes only, it doesn't contain the actual CSS styles to format the HTML markup. We can alter this behavior by passing full=True to HtmlFormatter() as follows:

>>>
>>> formatter = HtmlFormatter(full=True)
>>> 
>>> with open("out.html", "w") as f:
>>>     highlight(code, lex, formatter, outfile=f)
>>> 
>>>

This will re-write the existing out.html file. If you now open out.html in a text editor you will find that in addition to HTML markup and CSS classes, it also contains a list of CSS style in the head section of the document.

Open out.html in your browser and it should look like this:

style #

The HtmlFormatter uses a default color scheme named default to highlight the code. Pygments comes with several built-in styles. To view the built-in styles use get_all_styles() function of the pygments.styles package.

>>> 
>>> from pygments.styles import get_all_styles
>>> 
>>> list(get_all_styles())
['borland', 'xcode', 'pastie', 'emacs', 'vim', 'rainbow_dash', 'igor', 
'algol', 'murphy', 'native', 'lovelace', 'algol_nu', 'rrt', 'arduino', 
'paraiso-dark', 'trac', 'fruity', 'colorful', 'friendly', 'vs', 'perldoc',
'autumn', 'monokai', 'abap', 'default', 'paraiso-light', 'tango', 'bw', 'manni']
>>> 
>>>

Note: To get an instant preview of the available styles visit http://pygments.org/demo.

Once you know the style you want to use set it using the style argument as follows:

>>>
>>> formatter = HtmlFormatter(full=True, style='monokai')
>>> 
>>> with open("out.html", "w") as f:
...     highlight(code, lex, formatter, outfile=f)
... 
>>>
>>>

This will highlight the code snippet using monokai color scheme.

Open out.html and it will look like this:

linenos #

>>>
>>> formatter = HtmlFormatter(full=True, linenos=True)
>>> 
>>> with open("out.html", "w") as f:
...     highlight(code, lex, formatter, outfile=f)
... 
>>>

This will add line numbers to the highlighted code:

Open out.html and it will look like this:

hl_lines #

>>> 
>>> formatter = HtmlFormatter(full=True, linenos=True, hl_lines=[1,2])
>>> 
>>> with open("out.html", "w") as f:
...     highlight(code, lex, formatter, outfile=f)
... 
>>>

This will highlight the first and second line of the code.

Open out.html and it will look like this:

cssclass #

By default, HtmlFormatter() generates code block wrapped in <div> tag with
class="highlight". For example:

>>> formatter = HtmlFormatter()
>>> 
>>> highlight(code, lex, formatter)
'<div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">func</span><span class="p">():</span>\n\t<span class="c1"># function body</span>\n\t<span class="k">print</span><span class="p">(</span><span class="s2">&quot;hello world!&quot;</span><span class="p">)</span>\n</pre></div>\n'
>>>

We can specify any other class name using cssclass argument.

>>> 
>>> formatter = HtmlFormatter(cssclass="code-block")
>>> 
>>> highlight(code, lex, formatter)
'<div class="code-block"><pre><span></span><span class="k">def</span> <span class="nf">func</span><span class="p">():</span>\n\t<span class="c1"># function body</span>\n\t<span class="k">print</span><span class="p">(</span><span class="s2">&quot;hello world!&quot;</span><span class="p">)</span>\n</pre></div>\n'
>>> 
>>>

This returns the code block wrapped in <div> with class="code-block".

Generating CSS #

We have seen that passing full=True to HtmlFormatter() creates a self-contained document with CSS styles. Although, this approach works, but it generates a lot of redundant CSS styles. A better approach would be to create an external style sheet containing all the CSS classes used in the output.

We can use get_style_defs() method of HtmlFormatter() to get the style definitions used by the formatter as a string.

>>> 
>>> formatter = HtmlFormatter()
>>>
>>> print(formatter.get_style_defs())
.hll { background-color: #ffffcc }
.c { color: #408080; font-style: italic } /* Comment */
.err { border: 1px solid #FF0000 } /* Error */
.k { color: #008000; font-weight: bold } /* Keyword */
.o { color: #666666 } /* Operator */
.ch { color: #408080; font-style: italic } /* Comment.Hashbang */
.cm { color: #408080; font-style: italic } /* Comment.Multiline */
/* ... */
.bp { color: #008000 } /* Name.Builtin.Pseudo */
.fm { color: #0000FF } /* Name.Function.Magic */
.vc { color: #19177C } /* Name.Variable.Class */
.vg { color: #19177C } /* Name.Variable.Global */
.vi { color: #19177C } /* Name.Variable.Instance */
.vm { color: #19177C } /* Name.Variable.Magic */
.il { color: #666666 } /* Literal.Number.Integer.Long */
>>>

This returns CSS styles for the default style.

To get the CSS styles for monokai, do this:

>>> 
>>> formatter = HtmlFormatter(style='monokai')
>>> 
>>> print(formatter.get_style_defs())
.hll { background-color: #49483e }
.c { color: #75715e } /* Comment */
.err { color: #960050; background-color: #1e0010 } /* Error */
.k { color: #66d9ef } /* Keyword */
.l { color: #ae81ff } /* Literal */
/* ... */
.bp { color: #f8f8f2 } /* Name.Builtin.Pseudo */
.fm { color: #a6e22e } /* Name.Function.Magic */
.vc { color: #f8f8f2 } /* Name.Variable.Class */
.vg { color: #f8f8f2 } /* Name.Variable.Global */
.vi { color: #f8f8f2 } /* Name.Variable.Instance */
.vm { color: #f8f8f2 } /* Name.Variable.Magic */
.il { color: #ae81ff } /* Literal.Number.Integer.Long */

We can also pass additional classes to get_style_defs() method that will be prepended to the class. For example:

>>> 
>>> print(formatter.get_style_defs('.highlight'))
.highlight .hll { background-color: #ffffcc }
.highlight  { background: #f8f8f8; }
.highlight .c { color: #408080; font-style: italic } /* Comment */
.highlight .err { border: 1px solid #FF0000 } /* Error */
.highlight .k { color: #008000; font-weight: bold } /* Keyword */
.highlight .o { color: #666666 } /* Operator */
/* ... */
.highlight .bp { color: #008000 } /* Name.Builtin.Pseudo */
.highlight .fm { color: #0000FF } /* Name.Function.Magic */
.highlight .vc { color: #19177C } /* Name.Variable.Class */
.highlight .vg { color: #19177C } /* Name.Variable.Global */
.highlight .vi { color: #19177C } /* Name.Variable.Instance */
.highlight .vm { color: #19177C } /* Name.Variable.Magic */
.highlight .il { color: #666666 } /* Literal.Number.Integer.Long */
>>>

In case you have specified cssclass argument while creating HtmlFormatter instance then the default selector for get_style_defs() will be this class. For example:

>>> 
>>> formatter = HtmlFormatter(cssclass="code-block")
>>> 
>>> print(formatter.get_style_defs())
.code-block .hll { background-color: #ffffcc }
.code-block  { background: #f8f8f8; }
.code-block .c { color: #408080; font-style: italic } /* Comment */
.code-block .err { border: 1px solid #FF0000 } /* Error */
.code-block .k { color: #008000; font-weight: bold } /* Keyword */
.code-block .o { color: #666666 } /* Operator */
/* ... */
.code-block .bp { color: #008000 } /* Name.Builtin.Pseudo */
.code-block .fm { color: #0000FF } /* Name.Function.Magic */
.code-block .vc { color: #19177C } /* Name.Variable.Class */
.code-block .vg { color: #19177C } /* Name.Variable.Global */
.code-block .vi { color: #19177C } /* Name.Variable.Instance */
.code-block .vm { color: #19177C } /* Name.Variable.Magic */
.code-block .il { color: #666666 } /* Literal.Number.Integer.Long */
>>>

And here is how you can save the generated CSS styles to an external file:

>>> 
>>> formatter = HtmlFormatter(linenos=True, hl_lines=[1,2], css="code-snippet")
>>>
>>> 
>>> with open("out.html", "w") as f:
...     highlight(code, lex, formatter, outfile=f)
... 
>>> 
>>> with open("style.css", "w") as f:
...     f.write(formatter.get_style_defs())
... 
3583
>>>

Now open out.html file in a text editor and add the following line at the top of the it:

<link rel="stylesheet" type="text/css" href="style.css">

Open out.html in a browser and it will look like this:

Another way to create CSS styles is to use pygments as a command line tool. For example:

$ pygmentize -S default -f html > default.css

This command will create the CSS styles for the default style. The -S option specifies the style and -f option specifies the formatter.

To generate CSS styles for monokai, use this:

$ pygmentize -S monokai -f html > monokai.css

We can also specify the top-level class using -a option:

$ pygmentize -S default -f html -a .code-snippet > default.css

How to use Virtualenv?

What is Virtualenv? #

Virtualenv is a handy tool for creating Python Virtual Environments.

So Why do we need Virtual Environments?

You can think of Virtual Environments as a separate Python installation which allows us to work on projects using different versions of the same package without conflicting with one another. Consider the following example:

Let's say we are working on an e-commerce website and a CRM. Our e-commerce website depends upon version 1 of the foo package but our CRM needs version 2. At a time, we can only work with one version of the foo package. We can't have both versions simultaneously on the system. A virtual environment can resolve these kinds of problem easily.

Another excellent use case of Virtualenv is that when you are working on a system and you don't have the privilege to install packages globally. What you can do is create a virtual environment in your home directory and then install all the dependencies of the project inside the virtual environment.

Even if you don't work with conflicting packages and have full access to the system, it is still a good idea to always start your new project in a virtual environment that way if anything goes wrong with the project, your system-wide Python installation will remain intact.

Now you the Why of Virtualenv let's get into the How?

Installing Virtual #

To install Virtualenv type the following command:

pip3 install virtualenv

Creating Virtual Environment #

We now ready to create virtual environments. Create a new directory my-project and change your current working directory to this directory using the cd command:

$ mkdir my-project
$ cd my-project

To create the virtual environment type the following:

$ virtualenv env

Alternatively, you can also use the following command:

$ python -m virtualenv env

This will create a new a new directory named env in your current working directory. The structure of the env directory should look like this:

env/
├── bin/
├── include/
├── lib/     
└── pip-selfcheck.json

So what are these files and folder?

These files and folder constitute a separate Python installation.

The bin/ folder contains all the executables that you would find in an ordinary Python installation:

$ ls env/bin/
activate       activate_this.py  pip     python     python-config
activate.csh   easy_install      pip3    python3    wheel
activate.fish  easy_install-3.5  pip3.5  python3.5

By default, Virtualenv creates virtual environment using the version of Python under which it is installed. In other words, if Virtualenv is installed as a package of Python 3.5 then it will create virtual environment using Python 3.5. We can specify another version of the Python using the -p option.

virtualenv env -p /usr/bin/python2

This will create a virtual environment using Python 2.7 instead of Python 3.5.

Activating the Virtual Environment #

To use a virtual environment, we first have to activate it. Activate virtual environment by typing the following command:

$ source env/bin/activate
(env) $:

If you are on Windows use the following command:

C:\Users\user>env\bin\activate
(env) C:\Users\user>

Our virtual environment is now active. Did you notice (env) in front of the shell prompt? It indicates that the virtual environment named env is up and running.

Activating a virtual environment changes the $PATH environment variable temporarily so that the bin/ directory of the virtual environment will become first in the list. If you now execute python command, Python executable which resides in the env/bin/ directory is executed instead of the globally installed Python.

Once virtual environment is activated, any package you add or remove using the pip will only affect the virtual environment you are working on. Packages installed at system-wide installation will not be affected at all.

Deactivating the Virtual Environment #

Once you are done working with the Virtualenv you can deactivate it using the deactivate command:

$ deactivate

This will remove the bin/ directory from the $PATH environment variable, making globally installed Python accessible again.

Inheriting Packages using --system-site-packages #

On some occasions, you might want to create a virtual environment with the packages from the global Python installation. This can be accomplished using --system-site-packages option.

$ virtualenv env --system-site-packages

Installing MySQL (Windows, Linux and Mac)

MySQL is one of the most used Relational databases on the Web and in this lesson, we will learn how to install it.

This post specifically covers how to install MySQL on Windows, Linux and Mac OS.

Installing MySQL on Windows #

To download MySQL for Windows navigate to https://dev.mysql.com/downloads/windows/installer/. Scroll down a little and you will be displayed download links like this:

The first one is a web installer and the second one is an offline installer. Select whichever you like and hit the Download button.

After the download finishes double-click to start the installer.

Step 1: Accepts Licence Agreement.

Step 2: Select Setup Type.

The default set up installs many additional packages like MySQL workbench, Connectors, plugins, documentation, samples databases and so on. Installing all these packages will take some time and additional space. And the worst part is you won't even use half of these packages. That's why it is better to go with the Custom Installation and only install the packages you need. Don't worry you can always install these packages later by running installer again.

Step 3: Select Product and Features.

In this case, we are only interested in installing MySQL server. Select architecture ( x86 or x64 ) depending upon your OS and click Next.

Step 4: Check Requirements.

In this screen, the installer will download the missing dependencies. You will see this screen only when your system has missing dependencies. Click the "Execute" button to start the process. As you can see, on this system, installers need Microsoft Visual C++ Redistributable Package before it can proceed. Click the Next button and installer will prompt you to install the necessary package.

Once you have installed all the dependencies you will see a green check mark before the product name.

Step 5: Installation.

This screen shows the list of products that will be installed. The list only contains MySQL Server because that's what we are installing here. To start the installation process click the Execute button. After successful installation, you will get the following screen.

Click Next to proceed.

Step 6: Product Configuration.

In this part, the installer will configure MySQL Server. Default options are fine, so just click Next on the next on the several screens. After that, you will be prompted for username and password. Set username and password of your choice. You will use this to login into the MySQL server.

You have made through the hard part of the installation phase. In the next several screens accept the default by clicking Next. Finally, close the installer by clicking the Finish button.

Adding MySQL to PATH Environment Variable in Windows #

MySQL Installer by default doesn't add the mysql.exe to the PATH environment variable. As a result, every time we want to execute mysql.exe we have to specify the full path to it. The mysql.exe resides in the MySQL\MySQL Server 5.7\bin directory. But the path to MySQL directory depends upon the architecture of the application. If you have installed x64 version then it should be in C:\Program Files\. Otherwise, it should be in C:\Program Files (x86). On this system, MySQL is installation is in C:\Program Files\ and the complete path to mysql.exe is C:\Program Files\MySQL\MySQL Server 5.7\bin\mysql.exe.

To add mysql.exe to PATH environment variable follow these steps:

Hit Win+Pause/Break. This will open System Properties Window.
Click on Advanced system settings.
At the bottom of the newly opened window click on Environment Variables.
In the new window Select Path environment variable and click Edit.
Add ; at the end of the current variable value and append C:\Program Files\MySQL\MySQL Server 5.7\bin\mysql.exe (the exact may differ depending upon version and architecture of the application). Click OK to save.

Now start the Command Prompt and enter the following command:

C:\Users\user>mysql --version
mysql  Ver 14.14 Distrib 5.7.20, for Win64 (x86_64)

If you get the MySQL version as shown in the output, then it means everything is working alright and you have successfully installed MySQL Server. To login into the MySQL server use the following command:

C:\Users\user>mysql -u root -p

You will then be asked for a password. On success, on you get the following output:

Welcome to the MySQL monitor.  Commands end with ; or \g.
Your MySQL connection id is 51
Server version: 5.7.20-log MySQL Community Server (GPL)

Copyright (c) 2000, 2017, Oracle and/or its affiliates. All rights reserved.

Oracle is a registered trademark of Oracle Corporation and/or its
affiliates. Other names may be trademarks of their respective
owners.

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.

mysql>

To quite the MySQL shell hit \q.

Install MySQL in Debian/Ubuntu/Mint #

To install MySQL on Debian based system like Ubuntu or Mint type the following command:

$ sudo apt-get install mysql-server

If you are using a Redhat based system like Fedora or CentOS use the following command:

$ sudo yum install mysql-community-server

The installer will automatically add the path to MySQL executable to the PATH environment variable so you don't need to do anything.

Testing the Installation #

To the installation run the following command:

$ mysql --version
mysql  Ver 14.14 Distrib 5.7.19, for Linux (x86_64) using  EditLine wrapper

If you get the version of MySQL then it means you have successfully installed MySQL server.

After the installation MySQL server will start automatically. At anytime, we can check the status of the MySQL server using the following command:

For Debian/Ubuntu/Mint:

$ sudo service mysql status

For Redhat/Fedora/CentOS:

$ sudo service mysqld status

You can stop the server using the following command:

For Debian/Ubuntu/Mint:

$ sudo service mysql stop

For Redhat/Fedora/Cent OS:

$ sudo service mysqld stop

To restart the server use the following command:

For Debian/Ubuntu/Mint:

$ sudo service mysql start

For Redhat/Fedora/Cent OS:

$ sudo service mysqld start

Installing MySQL on Mac OS #

To download MySQL for Mac OS visit https://dev.mysql.com/downloads/mysql/ and download the DMG archive.

Double-click the installer to start and go through the usual installation process. After the successful installation, you will be provided a temporary password to login into MySQL server. The temporary password is not recoverable so make sure to copy and save it in the safe place.

The MySQL installer usually installs the files in the directory named mysql within /usr/local/. Just like Windows, MySQL installer for Mac doesn't add MySQL executable to the PATH environment variable. Consequently, we have to execute MySQL by specifying the full path to it in the terminal.

We are now ready to start the server. Open System Preference and click on the MySQL icon.

This will open MySQL preference pane.

As the screen shows, currently, MySQL server is not running. Click Start MySQL Server button to start the server.

To login to the MySQL server enter the following command in the terminal.

$ /usr/local/mysql/bin/mysql -u root -p

You will then be prompted for a password. Enter the temporary password you copied at the time of installation and hit Enter.
Once you are inside the terminal, change your password using the following command.

ALTER USER 'root'@'localhost' IDENTIFIED BY 'your_new_password'

To quit the MySQL shell enter \q.

Adding MySQL to Path Environment Variable in Mac OS #

To add MySQL executable to the PATH environment variable open .bash_profile and add the following code at the end of the file.

.bash_profile

#...
PATH="/usr/local/mysql/bin:${PATH}"
export PATH

Logoff and then login again for the changes to take effect. To verify the changes try running MySQL with without specifying the full path to it, as follows:

$ mysql --version

If you get the MySQL version then it means you have successfully added MySQL to PATH environment variable. You should now have MySQL server on your Mac OS up and running.

How did this guide work for you? If you are encountering any error let me know in the comments.