cssfinder 0.6.0

Tool for Hilbert-Schmidt distance calculation with Gilbert algorithm.

CSSFinder

Closest separable state finder.

Installing

To install CSSFinder from PyPI, use pip in terminal:

pip install cssfinder

Because CSSFinder is a modular tool, you will have to also install a backend package, which contains concrete implementation of algorithms. Simples way is to just install numpy or rust extras set:

pip install cssfinder[numpy]

pip install cssfinder[rust]

You don't need both, one will be perfectly fine. Alternatively, you may find cssfinder-backend-numpy and cssfinder-backend-rust on PyPI and install them manually, with exact same effect. Backends are dynamically detected from all locations, Python can import modules, thus any valid way of making backend code reachable for interpreter will work.

If you want to use development version, traverse Development and Packaging sections below.

But there is a catch!

CSSFinder can export PDF reports (and other formats too), but it uses weasyprint for that and weasyprint relies on GTK3. Unfortunately it is quite hard to get GTK3 going on windows and weasyprint requires it to work. Therefore you must handle installation yourself. Here you can find official guidelines from weasyprint. This repository may also help. Alternatively you can use WSL to install and run CSSFinder, as its seamless to do that.

Its worth mentioning that other formats are not affected by this issue.

Examples

Fortunately for all newcomers, CSSFinders comes in with some example projects, which may come in handy while starting to describe your first project.

List of examples available with CSSFinder can be obtained with command:

cssfinder examples list

Afterwards you should be presented with table, within console window, similar to this one:

For sake of example, let's assume that name 5qubits_json caught our eye, and now we are willing to spend some of our precious time diving deeper into what's inside. To do that, we have to clone this example somewhere with following command:

cssfinder examples clone --name 5qubits_json

This command, similarly to all other commands of all other CSSFinder commands, can be used with --help flag to inspect possible invocation parameters.

As result, you should find 5qubits_json directory have been created in your current working directory.

Example 5qubits_json relies on numpy backend, make sure to install it, if you haven't done it before.

Now we can proceed with running tasks defined within the project. That can be achieved with following command:

cssfinder project -p ./5qubits_json/ task run

This command will run all tasks, which may take something in between of few seconds and few minutes, depending on your hardware.

Result of calculations can be inspected in output/ directory in project folder (5qubits_json/).

Development

This project uses Python programming language and requires at least python 3.8 for development and distribution. Development dependencies poetry for managing dependencies and distribution building. It is necessary to perform any operations in development environment.

To install poetry globally (preferred way) use pip in terminal:

pip install poetry

Then use

poetry shell

to spawn new shell with virtual environment activated. Virtual environment will be indicated by terminal prompt prefix (cssfinder-py3.10), version indicated in prefix depends on used version of Python interpreter. It is not necessary to use Python 3.10, however at least 3.8 is required.

Within shell with active virtual environment use:

poetry install --sync

To install all dependencies. Every time you perform a git pull or change a branch, you should call this command to make sure you have the correct versions of dependencies.

Last line should contain something like:

Installing the current project: cssfinder (0.1.0)

If no error messages are shown, You are good to go.

Packaging

A Python Wheel is a built package format for Python that can be easily installed and distributed, containing all the files necessary to install a module and can be installed with pip with all dependencies automatically installed too.

To create wheel of cssfinder use poe task in terminal:

poe build

This will create dist/ directory with cssfinder-0.6.0 or alike inside.

Wheel file can be installed with

pip install ./dist/cssfinder-0.6.0

What you expect is

Successfully installed cssfinder-0.6.0

or rather something like

Successfully installed click-8.1.3 contourpy-1.0.7 cssfinder-0.6.0 cycler-0.11.0 dnspython-2.3.0 email-validator-1.3.1 fonttools-4.39.0 idna-3.4 jsonref-1.1.0 kiwisolver-1.4.4 llvmlite-0.39.1 markdown-it-py-2.2.0 matplotlib-3.7.1 mdurl-0.1.2 numba-0.56.4 numpy-1.23.5 packaging-23.0 pandas-1.5.3 pendulum-2.1.2 pillow-9.4.0 pydantic-1.10.5 pygments-2.14.0 pyparsing-3.0.9 python-dateutil-2.8.2 pytz-2022.7.1 pytzdata-2020.1 rich-13.3.2 scipy-1.10.1 six-1.16.0 typing-extensions-4.5.0

But cssfinder-0.6.0 should be included in this list.

Code quality

To ensure that all code follow same style guidelines and code quality rules, multiple static analysis tools are used. For simplicity, all of them are configured as pre-commit (learn about pre-commit) hooks. Most of them however are listed as development dependencies.

• autocopyright: This hook automatically adds copyright headers to files. It is used to ensure that all files in the repository have a consistent copyright notice.

• autoflake: This hook automatically removes unused imports from Python code. It is used to help keep code clean and maintainable by removing unnecessary code.

• docformatter: This hook automatically formats docstrings in Python code. It is used to ensure that docstrings are consistent and easy to read.

• prettier: This hook automatically formats code in a variety of languages, including JavaScript, HTML, CSS, and Markdown. It is used to ensure that code is consistently formatted and easy to read.

• isort: This hook automatically sorts Python imports. It is used to ensure that imports are organized in a consistent and readable way.

• black: This hook automatically formats Python code. It is used to ensure that code is consistently formatted and easy to read.

• check-merge-conflict: This hook checks for merge conflicts. It is used to ensure that code changes do not conflict with other changes in the repository.

• check-case-conflict: This hook checks for case conflicts in file names. It is used to ensure that file names are consistent and do not cause issues on case-sensitive file systems.

• trailing-whitespace: This hook checks for trailing whitespace in files. It is used to ensure that files do not contain unnecessary whitespace.

• end-of-file-fixer: This hook adds a newline to the end of files if one is missing. It is used to ensure that files end with a newline character.

• debug-statements: This hook checks for the presence of debugging statements (e.g., print statements) in code. It is used to ensure that code changes do not contain unnecessary debugging code.

• check-added-large-files: This hook checks for large files that have been added to the repository. It is used to ensure that large files are not accidentally committed to the repository.

• check-toml: This hook checks for syntax errors in TOML files. It is used to ensure that TOML files are well-formed.

• mixed-line-ending: This hook checks for mixed line endings (e.g., a mix of Windows and Unix line endings) in text files. It is used to ensure that text files have consistent line endings.

To run all checks, you must install hooks first with poe

poe install-hooks

After you have once used this command, you wont have to use it in this environment. Then you can use

poe run-hooks

To run checks and automatic fixing. Not all issues can be automatically fixed, some of them will require your intervention.

Successful hooks run should leave no Failed tasks:

Example of failed task:

Those hooks will be run also while you try to commit anything. If any tasks fails, no commit will be created, instead you will be expected to fix errors and add stage fixes. Then you may retry running git commit.

Profiling

To run simple profiling, You can use following command:

python -mcProfile -o "#examples_profile_5qubits_prof.prof" "assets/profiling/5qubits_prof/cssfproject.py"

Then You can view output using snakeviz:

snakeviz "#examples_profile_5qubits_prof.prof"

Command Line Interface

The CSSFinder is a script that finds the closest separable states. The script offers a command-line interface that allows you to execute different tasks related to your CSSFinder project. This documentation will provide a summary of the commands and options available.

Once you have installed CSSFinder, you can use it from the command line:

cssfinder [OPTIONS] COMMAND [ARGS]...

You can run cssfinder without any arguments to display a help message.

Options

The following options are available:

• -v, --verbose: increases the verbosity of logging messages. You can use -v up to -vvv to increase the verbosity level.
• --debug: enables debug mode.
• -V, --version: shows the version number of CSSFinder and exits.

Commands

The following commands are available:

project

This command allows you to interact with a CSSFinder project. You need to provide the path to your project as an argument.

run

This command runs all tasks in a CSSFinder project.

cssfinder project run [OPTIONS]

Options

• -t, --tasks: run specific tasks from the project. You can specify multiple tasks by using this option multiple times.

task-report

This command generates a short report for a single task in a CSSFinder project.

cssfinder project task-report TASK [OPTIONS]

Arguments

• TASK: the name of the task to generate the report for.

Options

• --html, --no-html: include or exclude an HTML report in the generated report.

• --pdf, --no-pdf: include or exclude a PDF report in the generated report.

• --open, --no-open: automatically open report in web browser.