Document metadata
Office documents, PDFs, and EPUBs return bibliographic and structural fields. None use column statistics.
docx_metadata / xlsx_metadata
DOCX and XLSX share DocumentMetadata (serialized as docx_metadata for both):
| Field | Meaning |
|---|---|
page_count | Pages (DOCX when available) |
word_count, character_count, character_count_no_spaces, paragraph_count | Text counts |
sheet_count, sheet_stats | XLSX: sheet count and per-sheet {rows, columns} |
title, author, subject, description | Core properties |
creation_date, modified_date, last_modified_by, revision | Dates and revision |
file_size | Document size in bytes |
pptx_metadata
| Field | Meaning |
|---|---|
slide_count | Number of slides |
title, author | Core properties |
creation_date, modification_date | ISO 8601 dates |
file_size | File size in bytes |
pdf_metadata
| Field | Meaning |
|---|---|
page_count, pdf_version | Structure |
is_encrypted | Encryption flag |
title, author, subject, creator, producer | Info dictionary |
creation_date, modification_date | Document dates |
file_size | File size in bytes |
epub_metadata
| Field | Meaning |
|---|---|
title, author, language, identifier | Dublin Core-style fields |
chapter_count | Spine item count |
file_size | File size in bytes |
DRM-protected EPUBs (encryption.xml) skip body parsing; metadata may be partial. Body text may also feed writing footprint when mined.
Related: Metadata overview.