Fix Sphinx documentation build warnings

[xiki-tempula]

Summary

• Resolve all Sphinx build warnings (reduced from 104 to 0 code-related warnings) by fixing duplicate object descriptions, invalid RST markup, inconsistent docstring indentation, and suppressing unresolvable cross-references from inherited scikit-learn docstrings.
• No documentation content is changed — all fixes are structural (Sphinx directives, indentation normalization, label deduplication).

Changes

Duplicate object descriptions from automodule + explicit directives:
Added :no-members: to automodule directives in files that also have explicit autofunction/autoclass directives below. The autodoc_default_options in conf.py sets "members": True globally, which caused automodule to auto-document all members — duplicating the explicit directives that follow.

• parsing/alchemlyb.parsing.{amber,gmx,gomc,lammps,namd,util}.rst
• postprocessors/alchemlyb.postprocessors.units.rst
• preprocessing/alchemlyb.preprocessing.subsampling.rst
• workflows/alchemlyb.workflows.base.rst

Duplicate object descriptions from autosummary + autoclass:
Added :no-index: to autoclass directives in individual API pages. The autosummary tables in parent pages already register these classes in Sphinx's index; the detailed API pages now render the full documentation without creating duplicate index entries.

• estimators/alchemlyb.estimators.{BAR,MBAR,TI,TI_GQ}.rst
• workflows/alchemlyb.workflows.{ABFE,base}.rst
• convergence/alchemlyb.convergence.convergence.rst

RST / docstring fixes:

• parsing/alchemlyb.parsing.lammps.rst: Fixed title underline length (was 1 char too short)
• estimators/alchemlyb.estimators.TI.rst: Renamed label estimators_TI to estimators_TI_api to avoid collision with the same label in estimators-ti.rst
• src/alchemlyb/parsing/namd.py: Replaced invalid :param:fep_filesrole with fep_files ``
• src/alchemlyb/visualisation/convergence.py: Normalized mixed 5-space / 4-space docstring indentation to consistent 4-space for plot_convergence and plot_block_average, fixing "Unexpected section title" errors

Suppress unresolvable sklearn warnings:

• docs/conf.py: Added suppress_warnings = ["ref.ref", "ref.term"] to silence warnings from inherited scikit-learn docstrings that reference internal sklearn labels (metadata_routing) and glossary terms (meta-estimator) we cannot resolve.

Test plan

• make clean html in docs/ produces 0 code-related warnings (only 6 SSL/network warnings from intersphinx on local builds, which do not appear on CI/ReadTheDocs)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.01%. Comparing base (220141f) to head (fed4704).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #454   +/-   ##
=======================================
  Coverage   99.01%   99.01%           
=======================================
  Files          29       29           
  Lines        2446     2447    +1     
  Branches      464      464           
=======================================
+ Hits         2422     2423    +1     
  Misses          2        2           
  Partials       22       22           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[orbeckst]

LGTM, yolo!

[xiki-tempula]

Annoying, this break the hyperlink. I guess if not broken, don't fix it.