Attributes to add/modify

[zaidalyafeai]

• Partial Attribute for datasets that might not include all attributes. For example, some social media datasets require re-clawring.
• UnitAdd conversations in for tha attribute.
• Source the source of the non-annotated data. ['web pages', 'public datasets', 'books', 'blogs', 'scientific articles', 'wikipedia', 'TV channels', 'LLM']
• Test Split rename to Has_Splits
• Domain is for the domain of the dataset i.e., ['Law', 'Science', 'News', 'General', 'Quran', 'Culture', 'General']
• Collection Style should be either ['crawling', 'manual curation']
• Annotation Style to better represent how the dataset was annotated, and add another option for no annotation, for example, public datasets are inheritly annotated annotated ['Machine', 'Human', 'Inehrited']
• Define a list of venues for easier annotation
• Rename Subsets to Dialect Subsets for better representation
• Preprocessing rather than Tokenized this general attributes define the preprocessning steps for the dataset i.e. ['tokenized','normalized','remove diacritization', 'other']
• Benchmarks are ill-defined in our definition; also, some datasets might include subsets other than dialects

[faizzyhon]

Implemented the requested attribute/schema migration in #664. The PR updates all 1,120 tracked dataset records, the schema, validation, website views/search/statistics, plots, and README documentation. It also separates dialect-specific subsets from generic subsets and documents the conservative migration rules used for topical domains and annotation styles. The local schema validation and JavaScript/Python syntax checks pass; the PR is ready for review and merge.