Nested Knowledge uses Semantic Versioning for labelling releases. Major version bumps (e.g. 1.x.x to 2.x.x) typically signal that an existing workflow will break & all impacted users will be notified in advance. Minor & patch version bumps (e.g. x.x.1 to x.x.2) correspond to new features, iteration on an existing features, and/or bug fixes. Users typically need not pay attention to these changes to continue using NK as they did before (although they may be missing out on awesome new things)!

Features

• Toggle between Sunburst and Hierarchy in embedded QLS

Bug Fixes

• Add extra publication date fields to RIS files (PY, Y1)
• Disable double click deletion of searches
• Disable creation of searches with empty queries

Bug Fixes

• Inspector Full Text filter on specific source
• Bulk FT import failing when one or more PDFs are encrypted
• EuropePMC searches failing when records have null keywords

Features

• Generalizing Risk of Bias module to Critical Appraisal
• Adding Newcastle-Ottawa Scale for Critical Appraisal
• New bibliographic Inspector columns: volume, issue, pages, and NCT ID

Bug Fixes

• Unable to add columns to tables of tags in Manuscript
• Added explanation of dangerous admin settings

Features

• Bigger and faster bulk full text upload and matching

Features

• Screening Model Improvements:
  • OpenAI embeddings
  • Scite tallies
  • Oversampling
• Advancement probability for two pass screening
• Robot Screener for abstract screening

Bug Fixes

• Deprioritize Robot Screener in Auto-Adjudication
• Numerical instability in QNS NMA for multi-arm studies with similar standard errors
• Relaxed PDF upload validations
• FT Blinding reappearing upon PDF upload

Bug Fixes

• Queues and progress bars for dual two pass screening properly handle switches between screening modes and off-workflow actions in inspector

Bug Fixes

• Correct font not appearing on all clients
• API search not executing through Literature Search UI
• Unable to view duplicate review bibliographic data

Features

• Allocation

Bug Fixes

• Database security patches
• Number of reviewers displayed for dual standard screening

Features

• User specified ordering of exclusion reasons
• Color coding user keywords
• Form Based Tagging question search & status indicator
• Display number of prior reviewers for dual screening modes
• Sticky AB/FT view in Inspector
• Font change

Bug Fixes

• Configuring Synthesis Access is laggy
• Tag import option is obscured by visual
• Double whitespacing in certain PDF annotations

Bug Fixes

• Downloading CSVs from QLS with protected Synthesis
• Added CSP
• Added HSTS

Bug Fixes

• Redirect to new nest upon creation
• Search exploration n-grams functionality returned

Features

• Blinding of full text status for dual screening modes
• Convert doc(x) study supplements to viewable PDFs on upload
• Hover text for sunburst is directly over the label
• Share button on QLS-specfic embeds

Bug Fixes

• Don't show 0 unscreened record counts in PRISMA diagram
• Site-wide load time improvements

Features

• Tag Table Contents

Bug Fixes

• Remove module completion selector from sequential module queues

Bug Fixes

• Updating tagging completion status for single records

Features

• Mark modules complete for individual records from Inspector
• Continuously adjustable module box widths (Screening, Tagging, Extraction, RoB)
• Export Comments from Inspector
• Jump to records in Inspector from Literature Search

• Create tags in-process for Form Based Tagging
• PDF annotation breaking for certain PDFs (UTF-8 null characters)

• Tag selection menu not properly closing in Form Based Tagging

• Reorder tags
• Avoid orphaning tags on merge and delete

• Improved visualization of overlapped tags in large hierarchies
• Sheet based search import method showing API

• Export studies and tags from QLS
• Inspector filter for Tag Contents
• Reorder columns in Table Builder export

• Clarifying Insight Embed UI
• Highlighting the hovered root tag in tag config

• Sync updated bibliographic data: per-record and bulk action

• Broken RoB Visuals
• Tag config styling problems

• Hiding tags on Synthesis
• Updated tag hierarchy visualization in AutoLit
• Export through Inspector
  • Moved Extraction and RoB from Export page
  • Added Screening & Tagging export
• Organizational owner may receive emails on public Synthesis change

• Bypassing invalid DOIs in sheet-based literature search import

• Form Based Tagging dropping excerpt on any click
• Improved arm creation, read-only cells, and status marking in Dual Extraction Adjudication
• Making overflow of concepts scrollable in Search Exploration

• Abstract annotation for tagging
• Import lit searches as spreadsheets of PMID/DOI

• Update PICO highlighting when abstract is edited
• Correcting IRR counts (exclude/exclude counted as include/include)

• Organizational Folders on AutoLit home

• Improved layout of floating elements in QLS & embeds
• Tag tip positioning in extraction is consistently visible

• Nest Backup
• Return of the QLS Dendrogram

• Truncating display of long nest names
• Incorrectly positioned tool tips on PDF viewer
• FT Import volume issues
• Log in / Sign in wording consistency

• Search Exploration v4, featuring concept generalization & improved workflow

• Negations in tag queries against all studies, not just studies with 1 or more tags

• Full text search performance improvements

• Support long API-based queries (>8k characters)

• Scrolling issues in Form-based Tagging
• Increased timeout on a number of requests
• Auto-refreshing background jobs table

• Manual FT import broken; improperly attributing source of the file

• Bulk import of Full Text PDFs with auto-matching
• Background job (searches, FT matching, model training) status tracking
• Performance improvements to FT import

• Restyle of QNS and zoom out
• Consistent scroll bars returned to the site

• Recurring searches causing server overload
• Overflow on nest home

• Import of tagging data with invalid UTF-8 characters
• Nests not sticking in folders on AutoLit home

• Folder organization of nests on AutoLit
• Heterogeneity statistics on QNS
• Funnel plots on QNS

• Radio button style/a11y improvements
• Maintain Inspector filters when responsively switching modes

• PRISMA diagram showing 0 counts for all categories

• Import previously gathered screening, tagging from sheets

• Protected Synthesis

• Bulk Full Text downloads properly unnzipping on windows
• Increase in Manuscript reference list font size (accessibility)

• DOAJ Updates
  • Correct newlines & other whitespace problems in searches
  • Truncate searches to top 1000 records, per DOAJ limit
  • Added DOAJ documentation for search creation

• Tag edit opening after saving Data Elements
• Tag question type not saving upon tag creation
• Typos/improved descriptions for FBT

• Form-based Tagging
• Improvements to related reports, Synthesis author ordering UX

• Allowing text selection in study abstract
• Added nest materials to nest copy

• Expert recommendation by bulk PDF upload (bibliographic data mined + FT attached)
• Nest Copy includes complete data (most notably Manuscript, Insights)
• Reworked layout for AutoLit Menu & Nest Home
  • User-level setting are inlined to respective modules
  • Protocol is edited on Nest Home directly
  • One Study Inspector link

• Embed/share all Synthesis types
• Display Synthesis authors with abstract
  • Improved author ordering UI
• Capability to share Synthesis for a private nest
• Improved placeholders & directives for building Synthesis when empty

• Tag import message for nests without an organizational owner
• Proper edit icon color for exclusion reason configuration

• Dual Two Pass Screening

• Proper layout in organizational tag template import
• Unscreening option in inspector typeahead

• Supporting Annual Subscriptions

• Full text import Bulk Action modal not closing on completion

• Improved Bulk Actions UI
• New Bulk Actions
  • Mark Full Text Review
  • Reviewer-level control over dual mode decisions
• Better wiki linking for file-based search import

• Proper labelling of screening form for dual screening
• Checkbox-based selection of RoB Data Elements

• Added comments to RoB export
• Improved a11y labels in Synthesis

• Disabling selection of empty exclusion reasons
• Enabling button-based addition of user-keywords

• Synthesis home restyle and updates
• New Synthesis embeds
  • Full Synthesis
  • All pages (QNS, PRISMA, RoB, Manuscript) embeddable

• New, distinct Sign Up page
  • Gather agreement to terms and user name in advance of credentials
• Bibliomining upgrades
  • 5-10x speed improvements with to improved CrossRef querying
  • Maintaining order of references from bibliography

• Tagging Recommendations

• Screening form UI improvements
  • Hover focus on exclusion reasons
  • Removed jitter when iterating records
• More precise wording of several Inspector filters
• Removed support for TLS 1.0 & 1.1
• DMARC support via DKIM/SPF: more robust email

• Insights on Synthesis
• One click exclusions
• Edit nest & study level comments

• Improved tag export options
  • “Exact” tag export
  • Adding all (exact) tags as columns
• Indicating tagged interventions in Extraction
• Consistent color schemes within Extraction

• Mark duplicate records manually
• Maintain multiple, related reports of a study
• Bibliomine from ClinicalTrials.gov using NCT ID

• Site-wide UI improvements, especially focused on a11y, improved contrast, and reduced color distraction

• Fixing RoboPICO backfill for large searches

• Corrected broken links in emails from mentions in nest-level comments

• Tagging closed by default for non-tagging modules
• Ability to enter discount codes and tax ids at subscription checkout

• Faster import of large searches (RoboPICOs are backfilled when > 500 records imported)
• Preventing duplicate reviewer-level dual screening decisions
• No sizing jitter for Full Text PDF viewer
• Preferring Support page linkout over support@ mailto link

• scite.ai integration: citation badges
• Duplicate information in CER Builder

• Display applied tags with no associated user (e.g. bulk actions, user account deleted) in AutoLit

• Robot Screener
• External ID in Table Builder
• Taxes included in Stripe checkout

• Performance improvement in Table Builder and Manuscript editor
• Removing jitter in tooltip display

• ClinicalTrials.gov integration
  • NCT Number added as a bibliographic field
• Europe PMC integration
• Query writing guidance for integrated databases
• Improved Qualitative Synthesis zoom mode
• Site-wide improvement of scrollbar appearance (especially on Windows platform)
• Adding time value & units to extraction export
• Export of categorical Data Elements in table builder
• Last modified time & user for tags & RoB forms
• Embase ID columns & filters in inspector

• Treating .txt as RIS in file-based search upload
• Explicit allowed file types in search upload
• Color correction of tabs in extraction configuration

• PMC ID filter in Study Inspector properly executes, and handles non-numeric inputs

• Making nest keywords editable from Screening Configuration

• Import full affiliation for EMBASE RIS imports
• Study inspector sorting all records even when top 100 are displayed
• Adding back progress bar for Screening Adjudication
• Corrected PRISMA report/record wording
• Properly display public Synthesis for non-logged-in users

• CER Builder

• Dual Extraction

• Full Text Search in Inspector
• Reduced dependency on DOI in deduplication (specifically, solving conference abstracts)

• No redirect to AutoLit Home upon login from originally requested page, unless on at landing page
• User Keywords do not appear editable from Synthesis

• Custom Search Engines
• Unscreening
• Download bibliomined PDFs
• About link in navbar

• Importing full author affiliations from RIS files
• Searching user name, in addition to email, in user mentions

• Two pass screening (separate screening queues for title/abstract & full text)
• Author affiliation search filter in inspector
• Full text review filter in inspector

• Improved inspector performance for very large (>10k records) nests
• Navigate to next record in dual screening adjudication when overriding determination
• Increase in abstract text line height

• Manage subscriptions & payment on the site, on an individual & organizational basis
• Modify bibliographic data directly on Add Individual References
• New Search Engines
• PDF Annotations included in tag merge

• Supporting lit search vennn diagrams for very large nests
• Correct intersection of multiple literature search filters on Inspector

• Nest deletion button disabled for non-admins
• Increased upload timeout on large files
• Nest invite email corrections

• Displaying size of search venn diagram pieces

• Faster upload and processing of large (5000+ records) file-based ris/nbib imports
• Consistent PICO predictions upon upload
  • Backfill for all records without PICO predictions

• User Invites
  • Adding users to nests & organizations now requires a complete email (privacy measure)
  • Invited user receives an email & an option to accept or decline the invite
  • Org admins receive email from users requesting to join their org
  • Emails without an account may inherit their invited permissions upon account creation
• Full citation export in tables
• Tag description made available in tagging
• Auto-selection of exclusion reasons upon on-the-fly creation

• Loading a mentioned study when not logged in no longer covers the login button
• Supporting PubMed queries over 4,000 characters long

• Duplicate Queue Improvements:
  • Ability to clear the queue, dropping or keeping all records for screening
  • Displaying number of duplicate records

• Proper attribution of triplicate+ records in the Duplicate Queue
• Defaulting data extraction configurationn mode to Data Elements (no hot click to reassign intervention root)
• Defaulting file-based import dates

• Full suite of PDF functions
  • Rotate
  • Zoom
  • Download
  • Delete
• Improved switching between reviewer-level (“You”) and adjudicating screening modes in dual-configured nests
• New home page
• New support page

• Improved display of extraction on horizontally constrained data elements

• Differentiaing inspector screening status for nests with dual screening configured in:
  • filters
  • column options
• Ability to filter to specific bibliomined files in inspector

• Properly deleting bibliomine attempts

• Synthesis RoB visuals only display included studies
• Synthesis RoB visuals only display domains where risk is relevant

• Preparation for new PubMed API (achieveing parity with PubMed web ~June 2022)

• Allowing tagging with an initially empty hierarchy
• Truncating overflowing PICO pills in Search Exploration
• Allowing deletion of tags corresponding to the configured intervention root

• Display a login prompt when visiting AutoLit home not logged in
• Properly handling email preferences from account management
• Reduced app size

• Extraction improvements:
  • Transitioned from study-level timepoints to data element-level timepoints
  • Ability to extract arbitrary many more timepoints than just Baseline/Outcome timepoints
  • Variable units in timepoint specification
  • Displaying continuous data element units in central tendency header
  • One-click metric deletion
• Tag export
• New site navbar
• Improved language around tag deletion

• Merge & edit bibliographic fields in duplication review
• New “plain” cursor mode in PDF viewer
• Improved labeling and ordering of comments & mentions
• Updates to citation string format

• Clearing an annotation in the PDF viewer clears any excerpt text as well

• Timezone attached to publication dates causing incorrect bibliographic details

• Retain tag descriptions in tag template creation & import
• Fix sporadic crashes of PDF Viewer for large files
• Fix tag selection drop jittering on vertically constrained pages

• Bibliographic data editing fixes - PMID & publication dates < 1970

• Exploded tag searching in Inspector
• New user improvements:
  • Account management updates
  • Guided nest creation wording & UX
  • New nest tips auto-open
• Added pricing info to free tier paywall

• Notes for literature searches

• Remove “Full Text” linkouts from rich text citation strings when DOI is not present for record

• Templated Nest Protocols
• Copy record-level rich text citations to clipboard
• Run searches immediately upon creation

• Missing spaces in some PDF tagging selections

• Enabling larger searches (up to 10k records)
• Emails for new user accounts

• Broken PMC linkout
• Author first/last name ordering in exports
• Clearer language for email verification
• Syncing NK email preferences with release & marketing email preferences

• Tag deletion restored
• Do not attempt scroll-to in Extraction when DE tag is not applied
• Deduplicating references within bibliomining attempts

• AutoLit Tips
• Tag Templating at the Organization level
• Automatic retraining of Inclusion Models
• Ability to hide Inclusion Predictions (while maintaining ordering)
• Perputation of Org Admin status to Nests

• Properly displaying bibliomine attempts in Inspector Literature Search Filter
• Improved guidance for Inspector typeahead

• Search deletion crashing when studies in duplicate queue

• Organization Management
  • User listing
  • Nest listing
  • Exclusion Reason Templates
• Exclusion reason import
• Persisting comment edits in AutoLit across page navigation & reloads

• Sorting within pinned nests on AutoLit Home

• Site-wide improvements in page load times
• Re-enabling copy to clipboard in PDF viewer

• Scroll-to annotated tags in extraction

• Replace missing icons in Inspector
• Nest Home tabbing display issues

• PDF Annotation-based Tagging
• Within PDF searching
• Nest Demos & Initialization for new users
• Predicted Search sizes

• Differentiating sign in & sign up links
• Tagged DE excerpt orientation
• Reduced client app size (faster loads)

• New Inspector Filters
  • Comments
  • Data Elements
  • Interventions
  • Other bibliographic data
• Titling manuscript and protocol downloads with the nest name
• Persisting inspector filters

• Tagging form reset on study iteration
• Displaying a user icon for users without a profile image
• Cannot enter N=0 in extraction for dichotomous metrics
• Full text export zipping for windows

• Tags suggest Data Elements for Extraction
• Copying comments on Nest copy
• Apply tag after on-the-fly creation
• Democratizing certain nest settings, notably:
  • Protocol editor
  • User keywords
• Editor stlying updates

• Dashboard scrolling
• Vulnerability patches

• Exploration visuals in inspector
• Bibliomining from PDFs
• Bulk expert recommendation from multiple identifiers
• Bulk Full Text & Supplemental material export
• Direct Google Scholar link out
• PMID parsing from mixed text RIS fields

• Dashboard hover text overflow
• Improved Safari compatibility
• Leaking key listener breaking arrow navigation

• Quantitative Synthesis updates
  • Random/fixed effects toggle
  • Plots of univariate & marginal distribution
  • Appropriate axis/coloring scaling/transformations
  • Styling improvements to fit wider site
  • Responsive/Mobile
• RIS export from Inspector
• Apply tags from extracted Data Elements using a Bulk Action
• View and edit external ID from bibliographic data

• Network diagram inaccuracies
• Ability to completely clear an exclusion reason name

• Enter key to apply tags
• Client side error logging
• Updated privacy policy

• SIGN RoB display

• JBI repeated question (case seres), and missing question (quasi-experimental & cohort)
• No caching login credentials following account deletion

• Account deletion
• Numbered citation export in Manuscript

• JBI case series RoB form

• Tag DE merge messaging correction

• Bulk action history + screening undo
• Ordering Exclusion Reasons by frequency

• Citation copy/paste
• Tag, DE, and intervention merging
• Pooling means with 0 SD in updatable tables

• Sticky AutoLit Menu (no more hover)
• Directed Wiki links from AutoLit
• Default Inspector filters
• Improved home page
• General AutoLit style assimilation

• Inspector RoB filtering crashes when no RoB configured
• RoB typeahead high risk filter correctly filters to high risk

• Risk of Bias - forms, Inspector filtering, export, Synthesis (domain distributions & traffic light). Cochrane RoB 2, JBI, and SIGN.
• Exporting study size in table export

• Mobile AutoLit!

• QNS Summary Table empty when only 1 intervention
• Various Search Exploration navigation

• Query Builder

• Disappearing Mention Selection
• “null” displaying for empty last names in user search

Inspector module switcher not present on tagging and extraction.

• Study supplemental materials
• Nest Materials
• QLS excerpt highlighting
• Inspector improvements
  • Updating screening decision does not close study detail
  • Author search typeahead
• User KW highlighting scheme improvements
• Typing '@' in comment box auto-opens mentions

• Proper error message when ID of expert recommendation cannot be found
• Handling unpaywall returning a link to non-PDF content types (most commonly due to DDoS protection measures)

• AutoLit landing page improvements:
  • Search first ~100 words of protocol from AL home
  • Sort columns
  • Display number of records in each nest
• Grouped PICOs displayed as individual pills when PICO collector is collapsed in Search Exploration

• Comment deletion option only displays to owner
• When unpaywall is down, full-text tab no longer crashes
• Site-wide CSRF protections
• Various style fixes, especially around overflow behavior in AutoLit

• Updated Search Exploration workflow
  • Research question specification on nest creation
    • Auto tag creation, keyword creation
  • PICO exploration
• Restyled Literature Search, with added features:
  • Fuzzy term intersections displayed
  • Displaying number of total & included records being deleted when searches are deleted

• Unfocusing gathered tag upon deletion

• On the fly tag creation no longer causing crashes

• Email notification on mention in comments
• Editing gathered tags

• Study Inspector cursor refocusing
• Protocol download filename
• Improved behavior of Data Element configuration overflow
• Ability to add an author when none are present

• Records without authors were causing crashes in screening, tagging, and extraction

• Synthesis components may be embedded on external sites
• Citations to included studies may be added in Manuscript
• Inclusion probability predictions as a column & sort in Study Inspector

• Comments with multiple mentioned users are no longer repeated multiple times in “All Mentions” on Nest Home
• When merging tags that are interventions, study arms of the source intervention are updated to the target intervention