برنامه‌نویسی به زبان پایتون |‌ Python Package Management System

Technologies & Concepts Covered (Python Package Management System)

Verifying Installation of “Kanda” (likely Anaconda/conda)

• Run the following command in the terminal:
  • kanda --version
• If it displays version output, the installation likely succeeded.
• If you see an error like “command not found”, installation likely failed—then re-check your steps or ask the community/group.

What This Session Is About

This session focuses on system/package management concepts before any hands-on work. Later sessions will cover:

• Using the package manager (e.g., conda)
• Installing Python
• Installing packages

Package Repositories (Where Packages Come From)

A package repository is described as a “hub” (like YouTube) / storage location where people host packages.

• There isn’t just one repository for everything—many repositories exist.
• Examples mentioned:
  • PyPI (Python Package Index) — the most famous repository used with pip
  • Conda channels — multiple sources used by conda, with conda-forge highlighted

---

Pip vs Kanda/Conda (Key Differences)

Core Scope

• pip
  • Installs/removes packages (and uses repositories)
  • Does not manage environments
• conda (Kanda)
  • Can create environments
  • Manages more of the system-level package workflow

Environment Management

• conda can create and manage isolated environments.
• pip generally can’t function as a full environment manager in the same way (as presented in the subtitles).

Non-Python Dependencies

• conda can handle non-Python dependencies (e.g., system/native libraries) when supported.
• pip is framed as focusing mainly on Python packages.

Dependency Resolution / Compatibility

• conda is presented as better at resolving complex dependencies to keep versions compatible.
• pip relies more on what package authors provide, so conflicts are more likely in large projects.

---

Repositories/Channels Used by Each Tool

pip → PyPI

• PyPI is described as the default Python package repository for pip.
• The speaker mentions the scale (tens of thousands of projects, increasing over time).

conda → conda-forge (Default Channel)

• conda uses channels (multiple repositories).
• The default channel emphasized is conda-forge.
• conda-forge is described as community-driven rather than tied to isolated individuals.

---

PyPI-only vs conda-only Packages (What to Do)

If a package exists only on PyPI and not on conda-forge:

• The suggested practical solution is to install it with pip, but inside an environment created by conda.
• The subtitles mention later you’ll see both options:
  • pip install <package>
  • conda install <package> (for packages available via conda)

General guidance:

• Prefer conda when the package is available in conda channels (often smoother dependency handling).
• Use pip when the package is only on PyPI or conda can’t provide it.
• Alternative approaches (like virtual machines or building packages from scratch) are mentioned but noted as rarely needed.

---

Big “System Diagram” Explained (Mental Model)

The subtitles explain package management end-to-end:

1. Package manager (e.g., conda, pip, others like Poetry)
2. Creates an environment
3. Installs packages into the environment
4. Installed packages have dependencies, which may include other packages/libraries
5. The environment includes (or is tied to) a Python interpreter
   • Python versions mentioned: 3.7, 3.8, 3.9, 3.10, 3.11
   • Possibly 3.12 referenced as being worked on
6. Package managers download packages from repositories/channels
   • conda-forge for conda, PyPI for pip
7. Understanding this structure helps with debugging when errors occur.

---

Tutorial/Guide Emphasis

This session acts as a conceptual prerequisite for later hands-on package/environment management.

The speaker encourages:

• Re-reading/re-watching definitions: Python environment, interpreter, packages/dependencies, package manager, repository
• Actively asking questions and explaining concepts to yourself

---

Main Speakers/Sources

• Speaker: The course instructor/lecturer (name not given in subtitles)
• Tools/Primary Sources Discussed: conda (called “Kanda” in subtitles), pip, PyPI, conda-forge, and a reference to Anaconda