- Great Britain flags added to better identify the English parts in
README.md - ChatGPT image added in the introduction of the Chimie Théorique et Modélisation : modèles simples notebook
- new getting started with Jupyter notebooks and Jupyter-lab section in the "Python in the Physical Chemistry Lab (PPCL) in a nutshell" notebook
- the
CITATION.cfffile has been rewritten due to an unresolved synchronization issue with Zenodo caused by the previous version
README.mdthoroughly updated: now includes an installation process for a Python distribution based onminiconda- new
requirements4pyPhysChem-LastVersion.txt, akarequirements4pyPhysChem-20241105.txt, useful to install withpipall the Python libraries required by pyPhysChem (``pip install -r requirements4pyPhysChem-LastVersion.txt`) - new 'getting started with Jupyter-lab'
- in the tight-binding notebook, the deprecated
scipy.integrate.simpsis now replaced withscipy.integrate.simpson. Involved to changesimps(y,x)withsimpson(y,x=x)
- Molecular representations notebook:
- the
jsmolwidget is now replaced with thepy3Dmolwidget - a
molViewclass has been created to simplify the use ofpy3Dmol - the deprecated
AllChem.GetMorganFingerprintAsBitVectis now replaced withMorganGenerator
- the
- new application to RDF profiles in the
ML-autoencodernotebook (see alsoML-1DCNN.ipynbnotebook for a 1D CNN classification of RDF profiles)
QCpy-0(FR)concepts_informatique_quantique.ipynbpreliminary Jupyter Notebook for the "Quantum Computing with python" sectionQCpy-additionalFiles/illustrations/Bloch_sphere.pngHD Bloch sphere illustration (original by Smite-Meister : CC BY-SA 3.0)QCpy-additionalFiles/illustrations/BlochSphere.pngLower definition Bloch sphere illustration (original by Smite-Meister : CC BY-SA 3.0)QCpy-additionalFiles/illustrations/circ_paire_bell.pngillustration of a basic quantum circuit for creating bell pairsQCpy-additionalFiles/theoryReminderSheets(FR)/Concepts_cles_mecanique-quantique.pdfquick course on the basics of quantum mechanics
(author: Matthieu Charnay, with I.C. Gerber)
- inpainting of an image added to the
ML-2DCNN.ipynbnotebook
ML-1DCNN.ipynbnotebook finalized
- RDF classification introduced in the
ML-1DCNN.ipynbnotebook (RDF profiles created with a in-housepyNanoMatBuilderlibrary, still under developmment and not publicly available) - new variant for the recurrent neural network (RNN) theme:
RNN-RNN_2.ipynb
- previous
ML-1DCNN.ipynbversion renamed asML-1DCNN-v0-dev.ipynb(private version only)
ML-2DCNN.ipynbupdated by I. C. Gerber, with the classification of squares/triangles/circles, in relation with the shape of nanoparticles
- previous
ML-2DCNN.ipynbversion renamed asML-2DCNN-v0-dev.ipynb(private version only)
QCpy-1(FR)-Modele_d-Ising.ipynbfirst Jupyter Notebook of the "Quantum Computing with python"QCpy-1(FR)appendix-Modele_d-Ising.ipynbappendix to this first Jupyter NotebookQCpy-1(FR)correction-Modele_d-Ising.ipynbcorrection to this first Jupyter NotebookQCpy-additionalFiles/available_gates.ipynblist of available quantum gates by default in the myQLM library for pythonQCpy-additionalFiles/theoryReminderSheets(FR)/Transformation_de_Bogoliubov.pdfquick course on the Bogoliubov transformationQCpy-additionalFiles/theoryReminderSheets(FR)/Transformation_de_Jordan-Wigner.pdfquick course on the Jordan-Wigner transformationQCpy-additionalFiles/theoryReminderSheets(FR)/Transformees_de_Fourier_Quantiques.pdfquick course on different variations of the Quantum Fourier Transform
all notebooks developped by a new co-author, Matthieu Charnay, with I.C. Gerber
- ML-autoencoder notebook. MNIST fashion dataset: latent space reduced to 28 instead of 30 (because imgs are 28x28)
- new Autoencoders applied to SMILES section in the ML-autoencoder notebook (co-author: Nans Bernard)
- new meaning of Morgan fingerprints bits sub-subsection in
MolecularRepresentations.ipynb ignoreLargeFiles.sh, to be run before agit add --allcommand
- in-house
pipManagement.pytool - last
requirements.txtfiles. Such files containing a list of libraries to be installed using pip install, or - even better - usingpipManagement.py. See also the pip documentation
ML-1DCNN.ipynb,ML-2DCNN.ipynb,ML-RNN.ipynbmade available, prior to new updatesCHANGE.mdrenamed asChangeLog.md, and new structuring with added/changed/fixed subsections
- discussion about decision boundaries introduced in the XAI iris4 notebook
- new Physical Chemistry Problems and Solutions notebook
first topic: rotational spectroscopy
- new explainability topic introduced in a new iris4 notebook
- TOC changed accordingly
- short exercises and new comments added to the iris1 notebook
- modification of the CO2 solubility with ANN notebook
- one model removed (with full standardization, GC descriptors included)
- some comments are added to facilitate the self-learning
- rdkit is now used to display the cations and anions
- the last part was commented on because of a persistent bug (unimportant part - will be solved later)
- Tight binding (TB) notebook updated with DOS plots (implementation of useful functions to enhance the capabilities of pybinding, including a Monkhorst-Pack mesh to plot DOS)
- Tight binding (TB) notebook completed with graphene
- Tight binding (TB) notebook completed with:
- Peierls distortion of the linear chain of hydrogen atoms
- 2×2 square lattice
- 2×2 rectangular lattice
- illustrations taken from R. Hoffmann (1991), Solids And Surfaces. A Chemist's View of Bonding in Extended Structures, Wiley-VCH, with © Wiley 1991
- new harmonic oscillator project, by Th. Leininger, @LCPQ (CNRS-UT3 lab)
- new contributor: I. C. Gerber, @LPCNO (CNRS-UT3-INSA lab)
- new Tight binding (TB) theme , adapted from IC Gerber's notebook on this topic
- ML notebooks dev versions still under heavy development and not made available
- Basics of Artificial Neural Networks (ANNs) for supervised learning notebook is now richly commented, and adapted to self-learning and talktorial format
- TOC updated
- new 1.7.3 version released after a careful check that new pyPhysChem rebranding did not introduce major bugs
- ML-CNN-dev.ipynb notebook renamed as ML-1DCNN-dev.ipynb. dev version not uploaded
- new 2D Convolutional Neural Network notebook. dev version not uploaded
- new Autoencoders notebook. dev version not uploaded
- new AdvancedPython notebook. First topic: reshaping NumPy arrays
- TOC updated
- new BrainHalfHalf style added to the css file
- project renamed "pyPhysChem" instead of "pytChem"
- new Convolutional Neural Network notebook. dev version not uploaded
- new Neural Network notebook: combining regression and classification with Keras. dev version not uploaded
- new Recurrent Neural Network notebook. dev version not uploaded
- new contributor: F. Jolibois, @LPCNO (CNRS-UT3-INSA lab)
- new Velocity Verlet project (author: Franck Jolibois)
2023.09.29. release 1.7.0 10.5281/zenodo.8396813
- "talktorial" transformation of the ML part
- pdf support splitted into a couple of pdf files
- some slides are inserted in notebooks as jpeg images
- MLChem.pdf updated with new cited studies
- copy of part of the DS4B-appendix.ipynb as DS4B-BasicsOfANN.ipynb
- H atom notebook converted into a Schrödinger model of the Hydrogen atom exercise notebook
- Unsupervised ML applied to the Iris dataset notebook now includes K-means in addition to PCA
- Molecules notebook renamed MolecularRepresentations.ipynb
- better contextualization of the scripts
- similarities calculated and plotted
- fingerprint plots
- selection of molecules that match a pattern
2023.06.22. v1.6.1 10.5281/zenodo.8069824
- various bug fix
- "Iris3" = Régression logistique is now also available in English (Supervised Machine Learning applied to classification)
- "Iris1" Exercise and Solution to exercise is now available in English
- "Iris1" = Lecture et analyse de la base de données par la librairie pandas is now also available in English (Reading and analyzis of the "iris" database with the pandas library)
- Iris1:
- default value for
corr()changed in Pandas version 2.0.0: The default value ofnumeric_onlyis nowFalse=> script changed withdfi.corr(numeric_only=True)
- default value for
- "Iris2" = Statistiques et régression is now also available in English (Statistics and regression)
- new contributor: S. Christodoulou, @LPCNO (CNRS-UT3-INSA lab)
- New Molecules notebook (in English), mainly based on the
RDKitlibrary and on thejupyter_jsmolextension - New ML-SVR notebook (in English): Prediction of the solubility of CO2 in ionic liquids with the Support Vector Regression (SVR) method. Illustration of an optimal hyperparameters search (author: Stella Christodoulou)
- New ML-SVR exercise, with the solution (author: Stella Christodoulou)
- The need for the rdkit and jupyter-jsmol libraries is now indicated in the README.md document
- Install instructions more detailed in README.md
- "Python in the Physical Chemistry Lab (PPCL) in a nutshell" project completed, prior to a peer review
- New Python in the Physical Chemistry Lab (PPCL) in a nutshell notebook (in English)
- New or modified css classes (ex, app, rq - formerly warn), intro -formerly rq)
- New "Puits de potentiel infiniment profond" project (in French, aka "Particle in a box" project)
- New release with a decent NMR notebook (although work is still in progress) and of the associated White Paper
- TOC updated (the Huckel.ipynb and InfiniteSquareWell.ipynb are not yet mentioned)
- New 2nd-order NMR notebook (work in progress)
- New Infinite Square Well notebook (work in progress)
- New Huckel notebook (work in progress)
- New Hydrogen atom notebook (work in progress)
- New Equations différentielles notebook
- Requires sympy module,as already needed for the Dérivées et Intégrales notebook
- Version 0. Strong improvements are needed as well as the implementation of numerical solution of ODE
- New Constantes Physiques et Mathématiques notebook
- New Dérivées et Intégrales notebook
- Requires sympy module
- Version 0. Strong improvements are needed
- Reorganization of the folder(s) and file renaming (
DS4B*prefix added to all Data Science for Beginners files and folders) - Exercises and solution to exercises of the Data Science for Beginners topic entirely rewritten:
- Comments added to the PCA notebook
- New Principal Component Analysis (PCA) topic
Minimalistic comments (so far)
Requires modules: keras, tensorflow, scikit-learn, pandas, plotly