Janium Collect catalogs the material that the institution leaves in its input sources: a folder, an SFTP server, cloud storage, a mailbox, a web form or the Janium capture app. Each institution has its own cataloging process, which watches its sources and picks up what arrives, and whatever enters through any of them follows the same path until it becomes a record.
For the people who decide or administer the systems, this answers a practical question: the institution can keep storing the material where it already does, and Collect connects to that place instead of asking for everything to be moved somewhere else.
The sources Collect supports
Folders and SFTP servers. This is the most direct way. The scanner, the operator or a system at the institution leaves the files in a folder, and Collect takes them from there. When the folder is on an SFTP server, Collect downloads the new files and can delete them from the server after the download, if the institution configures it that way.
Cloud storage. Collect connects to Google Drive, SharePoint and Zoho WorkDrive. In Google Drive the institution shares the folder with a service account, in SharePoint it registers an application with read permissions, and in WorkDrive it authorizes access to its account. Collect detects new files either through a notification from the service itself or by checking every few minutes. Collect reads the folder without modifying it, and each record stores the link to the original file, so from the catalog you can go back to the document in the cloud.
Email. A dedicated mailbox receives the files as attachments. Collect accepts only messages from the senders the institution authorizes, by address or by domain, and marks each message as read or moves it to another folder once it has been picked up.
Web form. The operator uploads the file from a form and fills in a few fields, which reach the model as context for cataloging that document.
The Janium capture app. The person cataloging photographs the item from a phone, a tablet or the browser, and Collect picks up each submission and tells the app which records are ready. Its two modes of work are described in Cataloging in front of the collection with the Janium capture app.
One channel per material type
Within a source, each folder is tied to a material type and to the way it is processed: which cataloging profile it uses, what kind of record it produces, how many pages it reads and in what format the output comes out. Whatever a folder does not set is inherited from the institution’s general configuration.
A library can therefore have on its server one folder for photos of books, another for scans, another for born-digital documents and another for journal articles. Books come out as monographs, digital documents as electronic resources and articles as analytic records linked to their journal, and the operator only has to leave each file in the folder that corresponds to it.
How the files of one item are put together
An item does not always arrive as a single file, and Collect has three ways of grouping the files that belong together.
Separator sheet. The operator photographs the books one after another and places a sheet with a QR code between each one. Collect splits the batch at each sheet, and the photos between two sheets make up one book.
Matching names. Images that share the same prefix in their name, such as the front and back covers of a record album, are grouped into a single item.
One PDF, one record. In the scan and document channels, each PDF is one item. Collect waits until the batch has gone a few minutes without changes, so it does not pick up a file that is only half copied, and then processes the complete batch.
Repeated files
Collect keeps track of what it has already processed by the location of each file, in all sources, so a file that is still in its place is not cataloged twice. In the PDF channels it also compares the content: a PDF identical to one already delivered, even if it arrives under another name, is set aside in a duplicates folder and does not enter the batch.
What comes next is the same for all of them
From the point of entry, the path is the same for any source: reading the material, cataloging with AI, quality scoring, injection of the fixed data and export to the institution’s format. Adding a new source changes none of that, and a record that arrived by email is reviewed the same way as one that arrived by SFTP. Delivery of the batch to the catalog is explained in Integrating AI cataloging with your catalog (ILS).
What the people who administer the systems need to know
Credentials. Each source is configured for the institution with its own credentials: the SFTP server key, the Google Drive service account, the application registered in SharePoint or the WorkDrive authorization. For Google Drive and SharePoint, read permissions are enough.
The copy of the material. Collect downloads a copy of each file to process it and does not work on the original in the source. If the material cannot leave the institution, the alternative is to process it on a machine installed inside it, as described in Cataloging with AI without the material leaving the institution.
Cadence. Sources that are polled, such as SFTP, the cloud or email, are checked every few minutes, so a file shows up in cataloging a little while after it is dropped.
Limits of the input channels
A moved file is a new file. Outside the PDF channels, the count of what has been processed is kept by location, so a file that is moved or renamed within the source may be cataloged again.
Order depends on the operator. A forgotten separator sheet joins two books into one record, and a file left in the wrong folder is cataloged with the profile of another material. Review brings it to light, and the fix is in the capture: a sheet in its place and each file in its folder.
A new source has to be set up. Connecting another folder, another mailbox or another cloud account requires registering it for the institution, with its material type and its credentials.
To continue the conversation
If your institution’s material already lives on a server or a cloud drive, or arrives by email, and you want to catalog it without moving it from there, write to us at info@janium.com. We would like to know where you keep what is still to be cataloged today and how you organize it.