DOCSEARCHbyAlgolia
Submit Docs

Shengzhou Li's Personal Website

website·Indexed 32 days ago

https://lisz.me

A collection of technical guides and academic reflections by Shengzhou Li, PhD in Information Science from the University of Tsukuba. The site covers machine learning, scientific literature search strategies, and extensive webmaster tutorials on LDAP integration, Docker deployment, Gitlab maintenance, and Jekyll blog configuration.

Tokens
143.2K
Snippets
710
Records
995
Agent score
57%

What's inside Shengzhou Li

  1. Overview of lisz.me site hosting and nodes

    The site is distributed across multiple hosting providers to ensure global accessibility and performance. It utilizes a mix of global CDNs, cloud providers, and regional nodes supporting both IPv4 and IPv6.
    Hosting ProviderDomainIPv4IPv6Note
    Netlifylisz.meYesYesGlobal Node
    Githubblog.lui8.cnYesNoUS Node
    Cloudflarelui.siteYesYesGlobal Node
    Alibaba Cloudlep.siteYesNoChina Node
    Tencent Cloudlui8.cnYesYesChina Node
    Tencent EdgeOneeo.lui8.cnYesNoChina Node
    Vercelzhonger.vercel.appYesNoUS Node
    Azureblog.lisz.inkYesNoHK Node
    4everlandblog.lep.inkYesNoHK Node

  2. Overview of AiiDA Framework

    AiiDA is an open-source framework written in Python designed for designing and managing complex workflows in computational science (e.g., materials science, biology). It ensures reproducibility by automatically tracking inputs, outputs, and metadata in a provenance graph. Key features include high-throughput execution, advanced graph query capabilities, a plugin interface for various computational codes and schedulers (SLURM, PBS Pro, Torque, SGE, LSF), and support for open science via data export to Materials Cloud.

  3. Overview of jekyll-theme-H2O-ac features

    jekyll-theme-H2O-ac is a fork of the H2O Jekyll theme designed specifically for academic researchers and DevOps engineers. Key additions include:

    • Academic-focused page structures: Dedicated templates for Academic Homepages, Archives, and System Logs.
    • UX Improvements: A manual dark/light mode toggle, image enlargement via Fancybox, and an expanded set of social icons (e.g., Google Scholar, ResearchGate, ORCID, SegmentFault, CSDN).
    • Developer Tools: Enhanced code highlighting via Prism.js, a clean code-copying mechanism that bypasses copyright text, and a custom-implemented article pinning feature.

  4. Overview of PICTL image processing and upload tool

    PICTL (Picture Control) is a command-line tool developed in Python that streamlines the process of preparing images for static blogs. Its primary purpose is to compress images, convert them to WebP format, rename them to random strings for uniqueness, and upload them to S3-compatible object storage (such as Cloudflare R2 or AWS S3).

  5. Overview of Kubernetes Network Components

    Kubernetes networking consists of three primary IP types:

    • Node IP: The IP of the physical/virtual node, used for connectivity and exposing ports.
    • Pod IP: A unique internal IP for each Pod, accessible only within the cluster.
    • Service IP: A virtual IP used for load balancing across multiple Pods in a Service; typically only accessible internally.

    Common components mentioned:

    • CNI (Container Network Interface): Calico is used in this implementation.
    • DNS: CoreDNS is the default for providing service discovery and alias resolution.
    • CSI (Container Storage Interface): Used for storage plugins; NFS CSI Driver is used in this implementation.

  6. Overview of Environment Modules

    Environment Modules is a tool used to simplify shell initialization by allowing users to dynamically switch between different software environments (e.g., different versions of Python, GCC, or OpenMPI) during a session. It is particularly useful in high-performance computing (HPC) clusters where multiple users require different versions of the same software without interfering with the global system environment.

  7. Overview of Progressive Web Apps (PWA)

    A Progressive Web App (PWA) is a web application that uses modern web technologies to provide a user experience similar to a native application. Key benefits include faster and more secure user experiences, better SEO (especially for Google), offline access, and the ability to add a shortcut to the mobile home screen. The 'progressive' nature means it enhances the experience for browsers that support PWA features while remaining a functional website for those that do not.

  8. Overview of the Utility vs. Understanding Framework for ML

    This framework evaluates machine learning fields across two dimensions:

    1. Empirical Utility: A composite measure of a method's breadth of applicability, ease of implementation, and real-world usefulness. High utility methods are reliable, predictable, and lack major flaws.
    2. Theoretical Understanding: A composite measure of intuitive interpretability, theoretical explainability, and the depth/completeness of existing literature. High understanding implies formulaic implementation and predictable results, whereas low understanding often relies on heuristics and trial-and-error.

    Fields are mapped into four quadrants: High Understanding/High Utility (Top-Right), Low Understanding/High Utility (Bottom-Right), High Understanding/Low Utility (Top-Left), and Low Understanding/Low Utility (Bottom-Left).

  9. Overview of Static Site Generators and Hosting Options

    Static Site Generators (SSGs) allow developers to write content in Markdown and compile it into static HTML files. Popular options include:

    • Hexo: Written in NodeJS.
    • Jekyll: Written in Ruby (natively supported by GitHub Pages).
    • Hugo: Written in Go.

    Common hosting and deployment platforms include:

    • Netlify: Provides continuous deployment by integrating with GitHub, offering automatic builds and a default *.netlify.app domain.
    • Cloudflare Pages: Offers continuous deployment, free static page hosting, and a *.pages.dev domain, with built-in IPv6 support and CDN capabilities.
    • GitHub Pages: Provides hosting via username.github.io domains.

  10. Overview of CONQUEST first-principles software

    CONQUEST is an open-source first-principles calculation software based on local-orbital density functional theory (DFT). It uses local orbitals to represent Kohn-Sham eigenstates or density matrices and is applicable to atoms, molecules, liquids, and solids. It supports both exact diagonalization and linear scaling methods to find the ground state, making it particularly effective for large-scale systems.

  11. Overview of AiiDA framework

    AiiDA is an open-source framework written in Python designed for designing and managing complex workflows in computational sciences (e.g., materials science, biology). It ensures reproducibility by automatically tracking inputs, outputs, and metadata in a provenance graph. Key features include:

    • Workflows: Connects local and remote computing resources.
    • High Throughput: Event-based engine supporting hundreds of checkpointed routines per hour.
    • Data Provenance: Automatic tracking for full reproducibility.
    • Querying: Fast graph queries across millions of nodes.
    • Plugin Interface: Supports various computing codes, analysis tools, and schedulers via plugins.
    • HPC Integration: Native support for SLURM, PBS Pro, Torque, SGE, and LSF.
    • Open Science: Exportable databases for sharing via Materials Cloud.

  12. Overview of Progressive Web App (PWA) Benefits

    PWA is a philosophy and set of web technologies designed to bridge the gap between native applications and web applications. By using a Web App Manifest and Service Workers, a website can achieve:

    • Faster, more secure user experiences
    • Improved SEO (especially for Google)
    • Offline access and caching
    • Native-like shortcuts on mobile home screens
    • Better performance by using the local device as a cache node