v1.1.0

new

• added rate-limiting to protect the API from abuse, and result-caching to improve performance of common database lookups
• added an API endpoint to re-send registration confirmation emails, in case the original was not acted upon and subsequently lost
• added an API endpoint to deduplicate studies, without requiring a new citation file import
• allowed reviews to have multiple "owners", and made it manageable by the owner(s)
• added support for specifying the numbers of reviewers per study for a percentage of studies in a given review, rather than always requiring the same number for all studies; and added support for filtering studies by the number of reviewers
• added functionality for admin users to add other admin users

improved

• across-the-board improvements in performance and stability of the system
• streamlined and standardized the underlying data models powering the application
• improved scalability and correctness of parsing citations from uploaded files
• improved speed and accuracy of text extraction from uploaded study PDFs
• re-wrote study deduplication system using a different approach, which is faster and more scalable
• updated study ranker model and its interface, and implemented its usage in a more efficient way
• ported data extraction functionality (including ML models) from a side-car Scala system into the main Python system
• refactored, standardized, and upgraded user authentication and authorization systems, for better application security
• increased study tags' and screening reasons' maximum lengths, from 25 to 64 characters, to allow for more descriptive values
• moved several admin-only endpoints into a separate API path
• improved styling and content of user emails

fixed

• actually forbid changes to "frozen" reviews
• prevented non-admin users from granting themselves admin privileges

development

• containerized all back-end services for more convenient, consistent builds and deployments, on whatever system
• repackaged colandr to follow current best practice, including consolidated configuration in pyproject.toml, adding a license file, and managing the package/environment with uv
• added an extensive unit testing suite, and set up CI workflows to automatically build, test, lint, and type-check source code
• migrated the API framework from flask-restplus to flask-restx to apiflask
• upgraded all 3rd-party packages to current (or close enough) versions, pulling in performance improvements, security fixes, and other benefits
• upgraded the database system from Postgres v9 to v17, and minimum Python from v3.7 to v3.11
• generalized filesystem storage functionality to allow storing data in Google Cloud Storage in addition to just local disk
• removed lots of "cruft", from HTML template files to outdated configurations to Scala side-car systems