Improve PyPI packaging metadata and README

[marevol]

Summary

Polishes the package for PyPI publishing. The project ships the GitHub README.md directly as the PyPI long description (readme = "README.md" in pyproject.toml), so this PR makes that file render and read well on the PyPI project page while keeping a single source of truth, and fills in the standard distribution metadata that was missing. It also bumps the version for the next development cycle.

pyproject.toml

• Bump version 0.2.0 → 0.3.0.dev0 to open the next development cycle (PEP 440 canonical dev form).
• Add keywords for discoverability.
• Add Trove classifiers (development status, console environment, intended audience, Python 3 / 3.13, OS, search/utilities topics). No license classifier is added on purpose — the project uses the SPDX license = "Apache-2.0" expression (setuptools ≥ 77, emitted as License-Expression), which is mutually exclusive with the deprecated License :: classifier.
• Add [project.urls] (Homepage, Repository, Issues, Documentation, Changelog) so the links appear in the PyPI sidebar.

README.md

• Add badges (PyPI version, supported Python versions, license) using absolute URLs so they render on PyPI.
• Feature pip install as Method 1 (with pipx / uv tool variants), since PyPI is the primary distribution channel; Docker and from-source methods follow.
• Fix the placeholder repository URL your-org → codelibs in the git clone examples.
• Fix the relative LICENSE link (which 404s on PyPI) to an absolute GitHub URL, both in the license badge and the License section.
• Normalize the FESS_VERSION example from 15.6.1 to 15.7.0 to match the code default and every other example.
• Refresh the Docker image tag from the stale 0.1.0 to 0.2.0.

Verification

• uv build produces fessctl-0.3.0.dev0 sdist + wheel.
• twine check dist/* → PASSED for both artifacts (confirms the README renders as the PyPI long description).
• Inspected the built wheel METADATA: version 0.3.0.dev0, License-Expression: Apache-2.0 (no license classifier), all Project-URLs, keywords, classifiers, and Description-Content-Type: text/markdown present.