Attachments
このコンテンツはまだ日本語訳がありません。
The Attachments page lists the files attached to communications in this project. From here you can open the original file, see which communication it came from, and see what register artefacts the classifier pulled out of it.
The list
Section titled “The list”Each row is one file, newest first. Use the Search filename box to narrow the list. The columns are:
| Column | What it shows |
|---|---|
| File | The filename. Click it to open the detail page. |
| Source email | The communication the file arrived with — click through to the Inbox item it belongs to. (The column and links are labelled “email” even when the source is an uploaded document or a meeting.) |
| Type | The file’s content type — the raw MIME type, for example application/pdf. |
| Size | File size. |
| Artefacts | How many register items were extracted from the file. |
| Tags | Tags applied to the file. You can edit them inline here. |
| Extracted | How the text was read (the extraction method), or a no text badge if none could be. |
| Received | When the source communication was sent, falling back to when the file was ingested. |
Opening the file
Section titled “Opening the file”Under each filename is an ↗ open file link (and an ↗ Open file button on the detail page). This opens the original file in a new tab using a secure, short-lived link — the link is signed, expires after about five minutes, and is scoped to your project, so a member of one project can’t use it to reach another project’s files. If a link has expired, open the file again to get a fresh one.
What was extracted
Section titled “What was extracted”Click a filename to open its detail page, which shows:
- The source communication — subject and sender, linking back to the Inbox item. (The detail page labels this “From email” regardless of the actual source.)
- Inferred from this file — the register artefacts the classifier extracted, each with its type, a link to the artefact, a confidence percentage where one is recorded, and the exact excerpt that supported it. Items you added by hand are marked manual.
- Extracted text — expand this to read the text the classifier actually read. Long files are previewed; open the file to read the rest.
Some files can’t be read (for example scanned images with no text layer). These show a no text badge and won’t have extracted artefacts. A file marked not classified (audit-only) is stored for the record but was never run through classification.