Anadolu OCR Corpus: Historical Ottoman Turkish and Arabic Text Pages

Anadolu OCR Corpus is an OpenCR export of OCR text and metadata for 52 historical PDF sources in Ottoman Turkish, Turkish, and Arabic. The dataset is provided in two Hugging Face configs: 'pages' and 'documents'. It was authored by fatihburakkaragoz and last updated on 2026-05-12.

Use Cases

• Train or evaluate OCR models based on historical document text.
• Analyze language distribution and script direction based on page-level metadata.
• Perform document-level text analysis using concatenated text and markdown fields.
• Study historical text patterns across Ottoman Turkish, Turkish, and Arabic sources.

Strengths

• Contains OCR text and metadata for 52 distinct historical PDF sources.
• Provides structured data in two configs: page-level and document-level rows.
• Includes validation status, language detection, and script direction metadata.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic, temporal, or source bias inherent to the 52 selected PDFs.

Provenance

Source

huggingface

Collection Method

OpenCR export of OCR text and metadata from PDF sources.

Freshness

Last updated 2026-05-12 17:14:52; freshness should be verified.