Release highlights

See details: issues, plugin changelogs, JS API compatibility

1.24 || 2025-Feb

Visualization and analysis

Grid: Rows and columns now resize together when you change row heights. This creates a zoom effect, where cell content adapts for optimal viewing. Zooming out shows the big picture. Zooming in shows details. This works especially well with dynamic content like proteins, user profiles, forms, and JIRA tickets.
Grid: We now support right-click panning. Hold the right mouse button to drag and move around your data. This feature is particularly useful when exploring extensive datasets in the zoomed-out mode.
Grid: You can now create column groups
Grid: You can now easily (un)hide columns
Cell renderers: Tags cell renderer now works with boolean columns (previously, it required comma-separated values)
Cell renderers: Sparklines now support multiple normalization options - by row, column, or global values
Cell renderers: We now support adaptive rendering of files in cells

Developer updates

JS API changes:

• New features:
  • Classes: 0
  • Functions: 10
• Breaking changes

1.23 || 2024-Dec

New tools and apps

Pyodide: Run Python in the browser
Admetica: Predict, visualize, and analyze 20+ ADMET properties using Datagrok's open-source ADMET predictor Learn more

Data access and management

New Visual Query Builder: Seamlessly join tables, aggregate, and pivot from one user interface

Visualization and analysis

Grid: Group multiple columns together and apply style for multiple columns at once
Scatterplot: Interactively explore tens of millions of data points Customize how labels appear and behave Define columns for scatterplot whiskers to highlight data distribution

Developer updates

JS API changes:

• New features:
  • Classes: 1
  • Functions: 0
• Breaking changes

1.22 || 2024-Nov

Data access and management

Custom identifier registration: Register custom identifiers (e.g., compound IDs) to search, link, and analyze entity data across the platform Video demo Developers: Learn more
Search integrated functions: Annotate queries and other functions with search patterns to display results when users search with matching terms Learn more
Plugin databases	You can now ship a Postgres database (such as chemical registration system) with your plugin

Visualization and analysis

Scatterplot

New data series connection lines let you track datapoint relationship over time Learn more

Platform performance

• Pyodide integration: you can now run Python functions, including data transformation steps, directly in the browser
• Improved automated data cleanup: keeps the server lean by deleting old data and logs (could be configured)

1.21 || 2024-Sep

warning

This version requires a database migration that cannot be rolled back. Once you upgrade to version 1.21.0, you cannot downgrade. If you need assistance, contact Datagrok Support.

Data transformation

We improved the calculated column feature (requires a PowerPack package):

• Usability: Drag-and-drop, search, and function suggestions in Add New Column dialog
• Formula assistance: Real-time validation, syntax highlighting, autocompletion hints
• Extended functions: Added package-specific function support

Video walkthrough.

ML & modeling

We continue to improve ML capabilities within the platform. We added:

• New interactive model training dashboard
  • Binary classification support with prediction threshold control
  • AUC-ROC visualization
  • Model comparison tools
  • Automatic model selection
• New algorithms (EDA package):
  • Softmax classifier
  • XGBoost tools
  • Partial least squares regression

Other

• File caching is now enabled by default. It works automatically without the need for configuration.

1.20 || 2024-Jul

New tools and apps

Tool / App	Purpose	Requirements	Learn more
Dimensionality reduction using WebGPU	Run dimensionality reduction on 100K molecules in less than 10 seconds	EDA package	Video walkthrough Wiki page
Diff Studio	App that turns differential equations into interactive visual models	Diff Studio package	Launch app Wiki page

ML & modeling

We continue to improve ML capabilities within the platform:

Improved in-platform model training:

• Automatic input substitution for models
• Interactive hyperparameter tuning
• New interactive visualization widgets for model performance evaluation

MLflow integration (learn more):

• Import and apply MLflow predictive models directly to Datagrok datasets
• Streamlined model management:
  • Automatic MLflow model fetching
  • Input annotation via MLflow tags
• MLflow models inference support

Platform performance

We've introduced caching for file shares (including S3, Azure, and local storage):

• Faster file retrieval and display in Browse
• Entirely browser-based cache, eliminating server requests
• Configurable caching for individual files and file shares
• Automatic cache invalidation when files are modified within the platform

Learn more.

1.19 || 2024-Jun

Visualization and analysis

Organize and filter hierarchical data: Group and navigate categories using a tree Toggle selections with checkboxes or shortcuts Customize hierarchy by rearranging, adding, or removing columns
Create custom filters using expressions: Use column-specific operations like >, contains, or regex Combine multiple conditions with AND/OR logic
Highlight text matches in cells with free-text mode

Other

• Added new connectors: Azure Blob, SharePoint
• Enhanced CSV export options for dataframes and views:
  • Preserve row and column order when exporting
  • Export molecules as SMILES strings
  • Selectively export visible columns only
  • Configure column qualifiers for improved data formatting

Developer updates

You can now integrate with Datagrok using Datagrok's REST API:

• Supports operations on files, tables, dashboards, and functions
• Allows for programmatic data management, dashboard creation, and function calls
• Python client library available for simplified API interaction

1.18 || 2024-Mar

New tools and apps

Tool / App	Purpose	Requirements	Learn more
Matched Molecular Pairs	Tool for analyzing how structural changes affect molecular properties and activity	Chem package	Interactive demo Wiki page
Docking	Tool for screening ligand libraries against AutoDock-prepared targets	Docking package	Interactive demo GitHub
Hit Design	App for collaborative hit design	HitTriage package	Launch app GitHub
Hit Triage	App for collaborative hit assessment and prioritization	HitTriage package	Launch app GitHub

Navigation and usability

New Browse view: Navigate the tree to access, preview, and manage anything in Datagrok, all from one convenient location. Learn more

Data access and management

Add custom metadata to anything - molecules, experiments, or users Learn more

Visualization and analysis

Cell renderers

New cell renderers: Dropdown MultipleChoice Tags

Trellis plot

Trellis plot now supports sparklines

Developer updates

• Computation queue
• Harmonized inputs and properties
• Major scripting Platform performance improvements
• CVM: CUDA support
• Push notifications
• Refactor package update mechanism
• JS API for viewer events

1.17 || 2023-Oct

Visualization and analysis

Grid: Pinned rows

Pin rows and save pinned view to the layout or project Learn more

Grid: Column selection

Hold Shift and drag the mouse over the column headers

Grid: Column navigation

In the Column Manager: Hover over a column to automatically scroll to it and see its details Filter columns based on data or semantic type

Grid: Summary columns

Smart forms show values from multiple columns in one cell Adaptive rendering to fit the cell size Values inherit color-coding from source columns Cell renderers for different data and semantic types (From PowerGrid package)

Grid: Linked tables in cells

Link tables and show data from one linked table in another

Line chart: Split by category

Split line charts by multiple categorical columns

New info plane: Plots

Automatically visualizes selected columns for quick profiling (Context Panel > Plots) Learn more

New info pane: Content

Displays details for the selected rows in a tabular format (Context Panel > Content)

Platform performance

We now support client and server-side caching for function results. This feature is particularly useful for functions that produce consistent outputs, like queries and scripts. The client-side cache, limited to 100 MB or 100,000 records, speeds up data access and improves network efficiency. The unrestricted server-side cache improves response times and overall server performance. Together, these caching mechanisms provide a smoother and more responsive platform experience.

Platform administration

• New installation wizard for platform configuration during deployment
• Authorization with Okta
• Amazon CloudWatch log export improvements

Developer updates

Auto-fill function parameters

Auto populate function input parameters based on a chosen key, such as selecting a car model to instantly fill in details like its mileage and engine size Learn more