Several components can ideally display content from sources across the web. The goal is to make available tools to enrich Wikimedia projects with then.
Selecting data for the query
- Labels and aliases in all languages
- Not all languages can be handled by an API. The unsuitable ones need to be filtered our in the preprocessing phase.
- If the main languages of the API are known, the query can use labels of that language as a primary option.
- Preprocessing can collect all necessary values and send to the local API. The local component for a specific API can arrange the data suitable for each API, for example concatenate the query strings in different ways, using AND or OR.
- Item's location can be used to narrow down search results or to distinguish from namesakes.
- The dates of the item can be used to narrow down search results. This is especially useful with maps.
- For maps the zoom level or the scale can be calculated using the size of the area the item covers.
- GLAM makes available images and their metadata through their public API. Wikidocumentaries uses many properties from the current topic to query that.
- When reading the data through the Wikidocumentaries API, the metadata is normalized using a different transformation for each GLAM.
- The metadata from different GLAMs is displayed in a uniform format in the Wikidocumentaries metadata display.
- When an image is saved to Wikimedia projects, users can reconcile string values with Wikidata items. String values can be saved as well, they will be available for reconciling later.
- In the Wikidocumentaries metadata interface, the string values are replaced by reconciled Wikidata items. Differences between the source data from the GLAM and Wikidata, such as recent changes in the GLAM's metadata, can also be highlighted.
- Maybe the GLAM could query Wikidocumentaries for changed information?
Consolidate data from different sources for the same item
Especially if Wikidocumentaries decides to store images or their metadata locally, these scenarios become available. Similar images from different sources can be detected. Their metadata can be compared, and the user is asked to verify the correct information. The updated data is saved in the central repository (Wikimedia Commons) with a reference to the source that provided this information. One of the information types that can be compared is the copyright status.
Search criteria for different media types
Image search is ideally based on
- Item label in the current language > must be changed to read the most suitable label.
- Native label
- Aliases > Which languages
- Item location > must be made more contextual
- Item date
- Names of historical administrative entities
- Zoom level / scale
- MediaWiki API
- Wikimedia Commons
- Wikimedia Commons API help - MediaWiki API
- Images from Finnish museums
- Publications from Finnish libraries
- Finna API documentation
- Old but still somewhat useful Finna API documentation
- CC licensing data available, prepared for RightsStatements
- Flick API Docs - Includes useful API Explorer
- Paris Musées
- Creative Commons Search, Openverse
- Internet Archive
- Images from Topotheks submitted by local people
- Linked Data Finland
- Map Warper
|About||Technology||Design||Content modules||Tool pages||Projects|
|Setting up dev environment||
|Visual editor||Central Park Archives|