Initial Commit: Add algorithms to determine peripheral and farthest nodes in a weighted graph

- Implemented Floyd-Warshall algorithm to calculate all-pairs shortest paths,
  supporting positive edge weights for weighted graphs.
- Added functions to identify peripheral and far nodes:
  - `find_peripheral_node`: Determines the node with maximal eccentricity.
  - `find_far_node`: Finds the node with maximal farness, summing shortest-path distances.
- Included error handling for non-positive edge weights, with appropriate ValueError raising.
- Developed extensive test cases using doctests:
  - Covered single-node, fully connected, sparse, cyclic, directed acyclic, and
    disconnected graphs.
  - Validated handling of edge cases including zero and negative weights.
- Ensured code meets PEP 8 standards and passed all lint checks with `ruff`.
- Provided detailed module-level docstring with explanations of algorithms,
  complexity, and example applications.

This commit provides foundational functionality for graph-theoretical analysis
and sets up for future expansions or integrations.

Author: jeffreyyancey