Support

HISE IDEs Now Support Pixi

HISE now supports Pixi for building and managing IDE environments. Users' Conda-based IDEs continue to work normally, and migration to Pixi is optional but recommended for optimal performance.

Pixi uses the same package sources as Conda but creates environments faster and with fewer failures, resulting in quicker uploads and more reliable builds. A HISE user can either convert an IDE from Conda to Pixi or spin up a new IDE from scratch in either environment. Users can create and manage Conda and Pixi IDEs in a unified dashboard, toggling easily from one environment to the other.

When using the Python SDK to install GitHub packages, Pixi saves build details to capture that environment. Those packages can then be reused in new IDEs. R packages, including large Bioconductor-style packages, can also be installed in the same IDE by using Conda within the Pixi shell.

SDK Performance Improvements

Faster cache_files (Python) | cacheFiles (R)

We optimized this method so it skips unnecessary lookups for file descriptors, since file metadata isn't returned in the output. Absolute file paths are still returned as before.

Faster read_files | readFiles

We improved performance for this method by indexing the file metadata database, which boosts query speed.

Updated Xenium pipeline metadata

The get_file_descriptors() method for the Xenium pipeline now pulls newer specimen metadata. New fields, such as tissueBlockSections, are now formatted correctly and returned in the response. (R SDK updates for this behavior are still in progress.)​

New v2/v3 Conda environments

New versions of Conda environments (for example, minimal_v3, Olink_v2) now appear alongside existing environments.

Pixi-related package updates

With Pixi now available, the SDKs require some additional packages, and all existing Conda environments need to be updated over time. IDEs created before this release will continue to work with minimal impact. If you upgrade to the latest SDK version, you will need to install a few extra required packages to complete the setup.​

Improved sampleId prompt (with skip option)

We updated the text for the sampleId prompt to better explain what to do when you don't have sample references. You can also bypass this prompt by using the do_prompt | doPrompt parameter in upload_files | uploadFiles.

Better logging and validation in read_samples | readSamples()

The read_samples() | readSamples() method now logs which samples were retrieved. Previously, some users downloaded a list of samples and then used that list as sample references in upload_files() | uploadFiles, which sometimes caused validation errors. This workflow now succeeds as expected.

Users can now securely download certain HISE files from Advanced Search to their laptops for use in tools like FlowJo or Xenium Ranger. Files that are eligible are marked Downloadable, and users can select the download icon to save them to their usual download location.

The Admin UI now provides a detailed managed data source view for no-code visualizations, including JSON metadata, associated files, and explicit status chips for completion state, errors, and publication state.

The Xenium pipeline was updated to support the tissue block where a single slide contains multiple tissue specimens. The pipeline matches the sampleId and TissueId for each tissue in a block, including blocks that contain multiple tissues. If this match can’t be made, validation fails, and the misidentified tissue block is prevented from advancing downstream for analysis.

New HISE Advanced Search Supports Shopping-Style Filters

The Advanced Search feature in HISE has been reconceptualized and redesigned. Instead of database-style queries, users now select a FILES, SAMPLES, or SUBJECTS search and add shopping-style filters (for example, file, sample, or subject metadata, lab results, surveys, custom metadata) to find the info they need:

• Context-specific filters are exposed based on the search type and subsequent selections.
• Filter operators available for query refinement include equals, contains term, is one of,  ,  , and others.
• With their final search results, users can create new file sets, add to existing file sets, or analyze results in an IDE. Data visualization is also supported for specified file types.
• Users can save and share queries via a URL or in the VIEW SAVED & SHARED interface. Queries are now unsaved by default. Users can also manage query visibility (private, shared with specific users, or shared with all account users).
• Advanced filters are categorized into logical filter groups (File Metadata, Sample Metadata, Subject Metadata, Lab Results, Surveys, Custom Metadata).

Implementation of Dedicated Public Subnet

Deployed our public‑facing service into a dedicated public subnet that is logically separated from internal application and data subnets. This architecture aligns with NIST SP 800‑171 3.13.5, which requires implementation of subnetworks for publicly accessible system components that are physically or logically separated from internal networks. The public subnet serves as a DMZ‑style network segment for internet‑exposed services, reducing the vulnerability of internal systems.

Upload Files Workflow UI Updates to Support Fast Mode

New UI for the Upload Files Workflow improves visibility into upload files calls for IDE instances associated with studies:

• Added banner on IDE Instances page that displays how many upload files calls you're currently running. It includes a REFRESH button (updates counts and shows time since last refresh but does not retry the call) and a VIEW ALL CALLS button (opens list view).​

• Added list view showing all upload files calls for a given study, including status, created and last modified date/time (with user or service), and a View Details button in each row.​

• Added workflow details view, which is a graphical representation of workflow tasks and their status, metadata showing when the upload files call was added and last modified (with user or service), and task-level tooltips that display logs with human-readable timestamps.

SDK Updates for Python and R

• Introduced new use_fast_mode boolean parameter in upload_files()  | uploadFiles() to shorten uploads by skipping environment builds (reducing build time from ~20 minutes to under 5).
• Added structured logging and improved error diagnostics across all methods.
• To prevent confusion over duplicate results, added ProjectGuid field to outputs of read_samples()(Python) | readSamples() (R), read_files() | readFiles(), and get_file_descriptors() |  getFileDescriptors().
• Added to_df parameter in get_study_spaces() | getStudySpaces().
• Optimized large queries in get_file_descriptors() | getFileDescriptors() so that they finish significantly faster and are less likely to time out.
• To reduce blocking time during large file uploads, made upload_file_to_private_folder() | uploadToPrivateFolder() asynchronous.
• Added prechecks to catch empty environment.yml files before the IDE commit step.
• Added prompt in upload_files() | uploadFiles() that appears when no input_sample_ids are given.

Collaboration Space UX/UI Improvements

• Updated Collaboration Space UX/UI for improved usability and workspace navigation
• Added expandable/collapsible sidebar to study workspace
• Simplified permissions and collaborator management section
• Created enhanced tagging interface with expanded view and edit capabilities
• Enabled editing of study details (name, description, hero image, collaborators, tags, associated projects) from Study Details tab
• Introduced study deletion functionality for study owners
• Improved document, asset, and permission management workflows
• Refined dialogs for creating, editing, and sharing studies/assets

IDE Stacking

• Stacking mechanism groups related IDEs to improve visibility, findability, and provenance tracking.
• Stack actions are available to all collaborators, including editing stack membership, updating trace information, and managing permissions.
• IDE stacks can be autogenerated by using a common IDE name plus a version number (v) or date (YYYY-MM-DD). Alternatively, a user can create a stack manually.
• Edits are supported for IDE trace, stack membership changes, and designation of the top-of-stack IDE.
• Stacks can be cloned, resulting in a duplicate set of IDEs for independent use or analysis.
• Removal, dissolution, and permanent deletion are supported for individual IDEs or entire stacks.

Traceability

• Traceability tab shows standalone assets attached to studies. This tab displays a graph of all data inputs and outputs for the asset.
• Users can click nodes in the traceability graph to view additional connected data assets or navigate to the detail page to view a selected node.
• Users can "walk the graph" to display and explore immediate neighbors (inputs and outputs) one step at a time, instead of rendering the entire CertPro graph at once.
• Backend and frontend now support partial graph fetching and incremental expansion as users interact with the traceability graph.

Elimination of Lag Between File Upload and File Availability in Advanced Search

After calling uploadFiles() or dropping files into a watchfolder, most files are now available in a few seconds, and even large data transfers take just a few minutes to become available in the Advanced Search UI.

update_sdk_version() SDK Method

• Updates your Conda environment to the latest SDK version.

start_training_run() SDK Method

• Launches customizable AI/ML training jobs with advanced resource options, including configuration of CPUs, GPUs, and memory.
• Enables distributed execution, script validation, and detailed error handling.
• Supports the recommended workflow of testing locally, approving or rejecting the run using the review_training_job_run() SDK, and monitoring the job in the AI/ML Hub.

review_training_job_run() SDK Method

• Allows programmatic review and approval or rejection of AI/ML training jobs.
• Returns the complete job context—including audit logs, properties, metadata, output files, and file provenance.
• Facilitates efficient quality control and troubleshooting by providing all relevant job details in an accessible tabbed view.

Quota Override

• Lets HISE admin users increase resource limits beyond the existing billing quota. To request an adjustment, contact Support.

Billing

• On Personal IDE Billing tab, added bar chart comparing user's quota spend to that of the team.
• Throughout Billing section, standardized positioning and spacing of graphs and axes, and improved the consistency of formatting and labeling, including naming conventions, case, percentages, and dollar amounts.

hisepy

• New method: save_custom_conda_environment(env_name : string, description : string, language : list[str])
• New method: update_sdk_version()
  • NOTE: This installs only the latest hisepy SDK version–to install a different version, contact Support.
• Bugfix: survey results attached to output from get_file_descriptors()

R

• New method: saveCustomCondaEnvironment(envName : string, description : string, language : list) 
• New method: updateSDKVersion()
  • NOTE: This installs only the latest R SDK version–to install a different version, contact Support.