Python Tutorial

How to Python in computational neuroscience, data mining, and more

All the tutorials are tested with Python 3.11.x. Older version might have a problem with the new versions of type anotations.

Anna Förster: German Python Videos from the FB1 ComCast Team

Bálint Aradi (FB1): Attic Lectures (English python course)

Flow chart

• Basic Structure of a Computer
• Representation of Numbers in the Computer
• Systematic Programming
• Flow chart symbols
• Examples
• Flow chart for baking bread

Workflow

• Python installation
• VS Code installation
• VS Code configuration
• VS Code Microsoft Tutorials
• VS Code Python Interactive window
• VS Code Markdown
• VS Code Debugging
• VS Code Working remotely via ssh

Important VS Code notes:

• You can mark segments of source node with your mouse (or keyboard) and use the TAB key to increase the level of the indentation or use SHIFT + TAB do decrease the indentation level.
• Linux: CTRL + SHIFT + 7 toggles between comment and normal source code.
• The function key F2 allows you to change variable and function names. VS Code goes through all the node in your Project directory and changes all occurrences of the function or variable name accordingly. The same is true if you change a file name in a project directory.

Python: The Basics of the basics

• Overview
• Hello, Python
  • input, print, string, int, float
• Style Rulez
• Programming Recommendations
  • Assert
• Basic Math Operations
• Truth Value Testing
• Formatted String Literals
• Flow Control Overview
  • Flow Control: if, elif, else
  • Flow Control: for-loop
  • Flow Control: while, pass, break, continue
  • Flow Control: match case
• Sequence Types
  • Lists
    • List Comprehensions​
  • Tuple
    • zip
• Mapping Types
  • Dict​
• Functions
• Type annotations
• Files
• JSON and dict for parameter files
• Creating order via sub-directories: os.makedirs
• Finding files in a directory: glob
• Class
• Exceptions (try / except)
• Importing Modules
• Built-in Functions
• Built-in Keywords

Python: Special topics

• pickle: save and load Python objects
• The Python Standard Library​
• Dataclass
• Organizing parameters: dataclasses and dataconf
• ProcessPoolExecutor: A fast way to implement multiprocessing
• Logging
• Python Scopes and Namespaces
• Collection of distinct hashable objects -- set and frozenset

Numpy: The Basics of the basics

• The N-dimensional array (ndarray)
  • set_printoptions
• Dimensions and shape
  • Broadcasting: Automatic adaption of dimensions​
• Making a new matrix
  • Making a matrix from numerical ranges
• Convert other data into numpy arrays e.g. asarray
• New matrix
• Save and load
• Reshape and flatten
• Slices and Views
• Concatenate Matrices and arrays
• Merging matrices
• Unique
• Where
  • Piecewise
• Extending an existing matrix: tile, repeat, pad
• Boolean matricies and logic functions
• Advanced Indexing
  • Ravel and UnRavel
• Available dtypes
• Constants
• Math functions
• Linear algebra
• Random numbers the non-legacy way
• Statistics
• FFT
  • rfft and spectral power
  • Spectral power and averaging over trials
• Meshgrid
• Flip, rot90, and roll a matrix
• Iterate
  • ndenumerate
  • ndindex
  • flat

Numpy: Special topics

• Einstein summation
• Numba: Numpy just in time compiler -- speeding Numpy up
• Memory layout of Numpy matrices
• Stack and Split, Compress
• Beyond normal np.save
• Trim Zeros of a 1d array
• Iterating over an array / matrix
  • nditer provides many ways to visit all the elements of one or more arrays
  • Replacing an inner for loop with apply_along_axis
  • nested_iters
• Manipulation of integers and their bits
• Numpy and JSON over Pandas
• Resize: Compensation for bad planning? Don't!
• Dealing with the main diagonal / triangles of a matrix

Numpy: Matlab topics

• Dealing with Matlab files
• Matlab is also just a Python package
• Running Python remotely from Matlab

In the case you know Matlab check here: NumPy for MATLAB users

Matplotlib.pyplot

• Simple plot and imshow examples
• Subplot
• subplots and gridspec: A more flexible placement
• Overview of the available functions
• Animation and Slider

Pandas

• Basics

Machine learning: Sci-kit Learn

• Sci-kit Overview
• KMeans
• PCA
• FastICA
• Support Vector Machine
• K Nearest Neighbours (pure numpy)

Machine learning: Pure numpy

• ROC
• K Nearest Neighbours

SciPy

• Remove a common signal from your data with SVD
• scipy.signal: Butterworth low, high and band-pass

Python: Questions of spectral nature

• PyWavelets: Wavelet Transforms in Python
• Instantanious Spectral Coherence
  • Linearize the spectral coherence

Python: Daily little helper

• TQDM: Make your progress visible
• Argh: Organize your command line arguments
• psutil vs os.cpu_count: How many "CPUs" do I have?
• ZeroMQ: Microservices as well as connecting computers via message queue
• Austin: Time and memory profiling

OpenCV2 and Python

• OpenCV2: Play, write, read a video
• How to read a webcam with CV2

PyTorch

• PyTorch lecture
  • Part 0 and 4
  • Part 1
  • Part 2
  • Part 3

• Get CUDA ready!
• Converting the original MNIST files into numpy
• Interfacing Data
• Data augmentation
• Layers
• Creating networks
• Train the network
• Fisher Exact Test: Test if your performance difference is significant
• Write your own layer
  • Replace the automatic autograd with your own torch.autograd.Function
• How to take advantage of an optimizer for your non-Pytorch project
• How to take advantage of a learning rate scheduler for your non-Pytorch project
• Unfold: How to manually calculate the indices for a sliding 2d window

PyBind11

• Expanding Python with C++ modules
• The fast and furious way (CPU and GPU CUDA)
• PyBind11 Stub-Generation

Tensorflow / Keras

• A fast non-introduction

Sympy

• Symbolic Computation

Numerical Integration, Differentiation, and Differential Equations

• Integration and Differentiation
• Differential Equations

Arduino

• Basics with Python / Matlab

Matlab: Basics

• Basic Commands and Variables
• Matlab Graphics
• Vectorization and Vector Calculus in Matlab
• Matlab Data Analysis

Other topics

• Open Source Tools
• Machine Learning Resources
• Datasets

S1: Advanced Programming (last update 16.02.2024)

• S1 Advanced programming and data analysis
• Preperations
• Task 1 -- Classycal neurons: Simulation and Mathematical Anaylsis
• Task 2 -- Collective coherent cortices: Data analysis
• Neuron Models Equations
  • For detailed descriptions, please see:
    • 'Neuronal Dynamics' textbook
    • 'Theoretical Neuroscience' textbook

• Neuron Models Equations -- Rate-based neurons
  • DivInE-model for MT neurons

• Neuron Models Equations -- Spiking neurons without explicit spike generation mechanism
  • Leaky Integrate-and-Fire
  • Exponential Integrate-and-Fire
  • Quadratic Integrate-and-Fire
  • Connor Stevens

• References
• Task 1 -- Pictures for comparision
  • Task 1 --Mandatory Tasks

• Task 2 -- Pictures for comparision
  • Task 2 -- Mandatory Tasks
  • Task 2 -- Voluntary Tasks: Preprocessing
  • Task 2 -- Voluntary Tasks: More Information!

Advanced Studies

• Material from ages past
  • 2022: Preparation -- Python class with and without classes
  • 2022: Deep Networks and Pytorch
  • 2022: Divisive inhibition: a dynamical circuit for change detection
  • 2022: Synchronization and dynamic oscillations in the visual system
  • 2020: Deep Networks and Tensor Flow
  • 2020: Divisive Normalization -- a Universal Concept for Adaptive Dynamics and Function of Cortical Circuits
  • 2020: Recurrent networks: Temporal dynamics and synchronization
  • 2017: Oscillations and information routing: CTC model
  • 2017: Change detection: The DivInE-Model
  • 2017: Contour Integration
  • 2017: Computation Spike by Spike
  • 2017: Natural scenes and sparse coding in visual cortex