feat(scraper): add Azure Foundry scraper for OpenAI and Anthropic models

[pranav671]

Summary

This PR adds a new scraper for the Azure Foundry model deprecation schedule, specifically tracking Azure OpenAI and Anthropic (partner/community) models.

It scrapes the official Azure Foundry OpenAI Model Retirement Schedule and extracts model names, versions, lifecycle statuses, retirement/shutdown dates, and replacement recommendations.

---

Motivation

I wanted to have model deprecations tracked for my application, and this repository has been excellent for that purpose. However, it was missing Azure Foundry (OpenAI & Anthropic) tracking, so I implemented this scraper to bridge the gap.

---

Key Implementation Details

• Target Filtering: Standardizes collection to only process standard Azure OpenAI (h3#azure-openai) and Anthropic (h3#anthropic) models. All other provider tables (Cohere, Mistral, Meta, etc.) are intentionally ignored.
• Exclusions: Explicitly excludes fine-tuned models, OSS models, and community models from other providers.
• Section Termination: Sibling iteration stops immediately at any heading tag to prevent nested tables (like h4#fine-tuned-models) from leaking into standard model results.
• Model Standardisation: Combines the Model and Version columns into the standard Model (Version) format when a version is present.
• Garbled Encoding & Empty-Value Protection: Implements a robust _is_empty_value() helper that normalizes all dash variants (ASCII hyphen, en dash –, em dash —) and strips garbled UTF-8 to Latin-1 decoding artifacts (e.g. â + control characters arising from em dash mis-decoding). This prevents garbage strings from being appended to model names or appearing in the replacement column.
• Status Mapping: Automatically maps Azure's lifecycle terms (GA/Preview to active, Deprecated to deprecated, Retired to retired) and auto-promotes any entry's status to retired if the shutdown date is already in the past.

---

Verification & Tests

• Workflow Run: Successfully executed the automatic README update workflow in GitHub Actions: Workflow Run #27218233737.
• Workflow Commit: Verified that the generated table aligns perfectly with the repository's format: Commit ad94712.
• Scraper Integration Tests: Added TestAzureFoundryScraper in tests/test_scrapers.py covering 8 integration scenarios (standard OpenAI parsing, Anthropic parsing, section bounds, status mapping, and exclusions) utilizing a local HTML fixture (tests/fixtures/azure_foundry.html).
• Encoding & Helper Tests: Added TestIsEmptyValue with 9 unit tests to explicitly verify that various empty, dash, and encoding-corrupted values (like â artifacts) are handled and stripped correctly.

---

Future Thoughts

• Refactoring Scrapers: Planning a future PR to refactor existing scrapers by moving shared helper functions (like date parsing, HTML table column matching, and empty-cell normalization) to a common utilities module under the scraper/ folder to clean up duplication.