Reader and Document Extraction
Edit on GitHubDetect formats and extract normalized documents, chunks, hierarchy, tables, visuals, assets, and ingest results.
Reader is the cross-format path for search, review, ingestion, migration, and AI preparation. Its 13 exported commands expose the normalized OfficeIMO.Reader model without forcing the caller to switch on every native document type.
What Reader returns
Get-OfficeDocumentreturns the normalized document envelope.Get-OfficeDocumentDetectionreports detected format and routing evidence.Get-OfficeDocumentCapabilityexplains the selected adapter's supported output.Get-OfficeDocumentHierarchyexposes section and structural relationships.Get-OfficeDocumentChunkreturns bounded text/content chunks.Get-OfficeDocumentStructuredexposes structured content.Get-OfficeDocumentTable,Get-OfficeDocumentVisual, andGet-OfficeDocumentAssetreturn non-text material explicitly.Get-OfficeDocumentBatchandGet-OfficeDocumentIngestsupport pipeline-scale processing.Search-OfficeDocumentsearches the normalized result.
Choose Reader versus a native family
Use Reader when downstream behavior is common across file formats: indexing, chunking, extraction, discovery, classification, previews, and bulk ingest. Use Word, Excel, PowerPoint, PDF, Visio, or another native family when the script must modify format-specific objects and preserve their native semantics.
Reader adapters are modular. A deployment can carry only the adapters it needs, while broader presets compose multiple local formats. Optional OCR, web, and process-backed integrations have separate dependency and trust boundaries.
Operational pattern
Detect first, inspect capabilities, extract the required surfaces, and record warnings/provenance with the output. Do not silently reduce a document to plain text when tables, visuals, or assets matter to the consumer.
See the document Reader examples and search the command reference for OfficeDocument.