28/11/2025
Step 1: Notify platform administrators
Step 4: Export metadata in CSV
This guide describes the process of bulk original content upload of entire collections onto the platform. It requires contact information exchange and initial communication between the collection providers and platform administrators and does not apply to standard, everyday collection management as (will be) available through the platform’s web interface.
This guide applies to platform version 1.1.x (http://multiverse.dcarts.uoa.gr:65352/impulse/api/v1/info).
You will need an FTP client capable of explicit TLS such as FileZilla (https://filezilla-project.org).
Preparing data before the actual collection upload is the most crucial phase of the entire process as it guarantees that the collection will be imported into the platform in a problem-free way and all assets will be available to users as intended.
Before attempting any actual upload, please notify the platform administrators about your intent to upload a collection, providing information on:
A collection id can contain lowercase letters, numbers, underscores and dashes. It is usually a direct transliteration of the collection title. For example, a collection titled “Glass Slides” would be given a collection id of “glass-slides”. This needs to be unique through the entire IMPULSE platform.
The administrators will get back to you with details on any adjustments required before the upload.
First, create a directory for the upload. Use the collection id as the directory name. This will be referred to as the “collection directory” in the rest of this guide.
Currently, only flat (i.e., non-hierarchical) collections are supported. Therefore, all files - assets and previews - should be in the same directory, or subdirectories of it if needed, with each asset file representing a single, individual asset. To proceed accordingly, copy all files into the collection directory. Make sure that no files are copied other than asset and preview files and that no asset file is larger than 25 MB.
Regarding 3D models, please do make sure that each model is optimized for VR rendering and network transmission, appropriately scaled and oriented, and geometrically centered at the global origin.
Use the template provided.
Make sure to fill-in all mandatory fields, that is:
The assetID must be unique for each asset in the collection. It may be derived by the asset filename (e.g., the filename without extension, possibly prefixed by subdirectory), an auto-generated numeric value, a random UUID etc. If no other option is available, simply fill-in the field for the entire collection with an increasing integer after preparing all other metadata. In all cases, please only use lowercase letters, numbers, underscores and dashes.
The assetURI and previewURI must be relative paths to the respective files within the collection directory, that is, just the filenames or relative paths including subdirectories. Please make sure to include filename extensions as this is required for proper handling of the assets and cannot be inferred or generated safely.
The contentType must be one of the MIME types supported by the platform, namely, model/gltf-binary or an image type.
For text fields, e.g., desription, multi-line text is fully supported. Alternatively, you may use “\n” to indicate a line break. There is no need to escape any other character.
For more information, please refer to the Collections and assets schema, discovery and access documentation on the IMPULSE platform wiki, the template provided and the Dublin Core specification.
Use relevant functionality in your spreadsheet editor of choice. For example, Google Documents supports this through “File > Download > Comma Separated Values (.csv)”.
Make sure to use “,” as a field delimiter and double-quote all text fields.
To make the export process easier and compatible with potentially existing infrastructure, the generated CSV may include additional columns not relevant to the IMPULSE platform. These will be ignored and need no further attention. However, the fields relevant to the IMPULSE platform need to be named exactly as specified in the template provided.
Please name the CSV file “metadata.csv” and save it in the collection directory.
Create a ZIP, RAR, 7Z or TAR package containing a single entry, that is, the collection directory. Do use maximum compression as this is meant for network transfer.
The filename of the package should be the collection id.
For example, for the “Glass Slides” collection you should create a “glass-slides.zip” file containing only a “glass-slides” directory which, in turn should contain all asset and preview files and a “metadata.csv” file.
Collection package uploads are to be done via FTP at multiverse.dcarts.uoa.gr:65011.
Use your FTP client of choice to connect via explicit TLS over FTP using the credentials provided by the platform administrators. For Filezilla FTP client in particular, use the following settings:
The remote end may already contain other collection upload packages you may have previously uploaded. Please make sure to manage the remote end’s contents as needed to ensure enough space for your uploads.
Make sure you only upload the package created in Step 5 of the Preparation phase. Please do not upload any other files.
To complete the process, send an email to the platform administrators regarding a completed upload, or to report any errors or issues during any part of the process.
After the upload is verified and imported, you will be notified back to proceed to verifying via the web and/or the virtual world client that the process was successful and that the collection and all its assets are available as intended.
Please make backups of all files involved, including the generated CSV, packages, etc. Files on the remote end may (and probably will) be deleted without notification for filesystem management purposes.
(c) 2025 by Spatial Media Research Group, NKUA