Video Transcript
Welcome. This session covers document extraction — the process that automatically populates much of the document information in Adept. We'll cover what extraction does, the five categories of information it can produce, and the two fields Adept provides to track extraction activity. Let's get started.
Adept extracts information from supported documents and stores it in the database, where it supports Search, document relationships, metadata management, previews, reporting, and more. Here's a good way to think about it: extraction transforms a file into a managed document — one whose content, properties, and relationships can be understood and used by Adept, not just stored as a blob of bytes. Depending on the document type, Adept can extract five categories of information: system metadata, document properties, document relationships, full-text content, and thumbnail preview images. We're going to go through each of these five in turn.
The first category is system metadata. Adept automatically extracts basic information about each document straight from the operating system — no configuration needed. Common system metadata fields include Filename, Ext, File Size, and File Date, which is the file date and time reported by the operating system. This information is what helps Adept identify, organize, and manage documents within the vault, and it's captured automatically the moment a file is added.
The second category is document properties. Many document types contain properties or attributes baked into the file itself — things like CAD attributes, custom properties, file properties, and other application-specific metadata — and Adept can extract these values and store them in database fields. Property extraction allows users to search for documents using information contained within the file, rather than relying solely on file names or folder locations. So even if a file was never manually tagged, its own embedded properties can make it findable.
The third category is document relationships — and this is one of the more powerful things extraction does. Adept can identify and store relationships between documents by analyzing references contained within supported file types, including AutoCAD Xrefs, Autodesk Inventor assemblies and subassemblies, drawing-to-model relationships, and document references and underlays. By maintaining these relationships, Adept helps users understand how documents are connected and how a change to one document may affect others downstream. It's also worth noting that some document types, like AutoCAD drawings, can contain multiple logical entities stored as extended records, letting Adept track those sub-entities individually.
Categories four and five go hand in hand: full-text content and thumbnail previews. For libraries configured to support Full-Text Search — FTS for short — Adept can extract searchable text from supported document types and store that content in a search index, letting users search for words and phrases within documents even when those values don't appear in properties or file names. On the thumbnail side, many supported document types contain embedded preview images that Adept can extract and store, and those thumbnails show up in the Document Dashboard and in Data Cards, giving users a quick visual reference without having to open the file at all. Supported file types, requirements, and limitations are covered in the separate Full-Text Search Extraction article.
Adept provides two fields to track extraction activity on every document. Extract Date records the date and time the document was last extracted, and Extraction State indicates the current status: 'In Progress' means extraction is processing or a previous attempt failed, 'Extracted' means it completed successfully, and a blank value means no extraction status is available. Both fields can be added to grids, searches, and Data Cards just like any other field. Depending on document type and system configuration, extraction may happen automatically as part of standard document management operations or be initiated manually — and users can always use Refresh Extracted Data to re-run extraction or Clear Extracted Data to remove existing extracted information.
Extraction is what turns a file into a managed document — one whose content, properties, and relationships Adept can actually understand and use. Depending on the file type, that can mean up to five categories of information: system metadata, document properties, document relationships, full-text content, and thumbnail previews. You can always track extraction activity with Extract Date and Extraction State, with Refresh Extracted Data and Clear Extracted Data available whenever you need to re-run or remove it. Thanks — happy to take any questions.
