Releases: Psychoanalytic-Electronic-Publishing/OpenPubArchive-Content-Server
v.2020-09-17 Server with updated opasDataLoader and solrUpdateViews apps
Downloads for MostCited and MostViewed are limited to 500 rows in this version, since large downloads were breaking the Solr data return, at least through Solrpy. Looking at alternative means.
v.2020-09-15 Server with opasDataLoader and solrUpdateViews apps
Alpha 2020-09-15.1 release of the OpenPubArchive Server (OPAS) with first official alpha of AWS based opasDataLoader and solrUpdateViews. These additional apps load from S3 storage to the configured Solr and RDS instances.
v.2020-08-26.Alpha
Two new endpoints: v2/Reports and v2/Client/Config and bugfixes and code cleanup per https://github.com/Psychoanalytic-Electronic-Publishing/OpenPubArchive-Content-Server/wiki/CHANGELOG.md
v2019.1205.1.alpha (preview)
This is an early alpha of the server that's still under development and not feature complete. It's depends on fairly PEP-specific components at this point. The solrXMLLoad data load software is written to load documents written to the PEP A1v1 DTD only. The solr and MySQL schemas are based on PEP requirements, as is of course the included features and endpoints. This is considered acceptible since it is the PEP-Web project that's currently the only driver for development.
The alpha server implements the same v1 API that's used by PEP-Web and can be accessed using any v1 client. The only such client at this point is PEP-Easy. A slightly modified version of PEP-Easy was set up for use with this release--besides the non PEP URL for the server, there are 2 or 3 small tweaks that have been made in the return data to workaround things that weren't implemented as planned in the v1 implementation on PEP-Web.
The API release can be tested right in the OpenAPI documentation interface, online at http://api.pep-web.rocks/docs#
You can try it with the prelease PEP-Easy 2021 build at http://pepeasy.pep-web.rocks/
These demos are on a lower end server configuration and of course the code and configuration has not been performance tuned. Both the API server and PEP-Easy links can be tried without a username or password. However, you cannot view full-text documents without one, and many features in the data (e.g., links, glossary) are not available when not logged in.
Release notes
- Our devops have not yet setup some of the folder dependencies, so
- image in documents and journal banners are not currently displayed
- PDF original downloads are not available
- error handling is done differently in this server, and PEP-Easy hasn't been designed to recognize that. In the old system a "HTTP 200" response with an error message (see the ErrorReturn schema in the API) is returned from the server even when there's an error. The FastAPI/RESTful way of doing things in contrast is to return a standard HTTP error message. I'm still debating whether to accommodate the old way, or the more "standard" way. Both approaches are valid.
- Currently, searches are using a Solr field without thesaurus expansion, though it's implemented in other fields.