v2.4.0
Mon Jun 24 2024
New Features
AI-Engine Enhancements
- Enhanced Decomposition Outputs: Now supports multiple types of decomposition in the output for more detailed analysis.
- Prior vs. Posterior Analysis: Initial work on differentiating between prior and posterior data, including a new endpoint to access these comparisons.
API Improvements
- Variable Statistics Table: Added a comprehensive table for variable statistics.
- Variable Change Tracking: Monitor changes in variables with new endpoints.
- Attribution Reference Point Fallback: Enhanced the reliability of attribution by adding fallback mechanisms.
- Distinct Project Attribution Endpoint: New endpoint to uniquely identify project attributions.
- Total Profit Calculation: Automatic calculation of total profit.
- Project Management Enhancements: Set a maximum limit on the number of projects per tenant, with new identifiers for each project.
- Scoped Identifiers: Introduced scoped IDs to variables and response curves for better clarity and removed redundant IDs.
Teams and User Management
- Membership Count: Main API now includes membership count.
User Interface Updates
- Project Creation Limits: Disabled the option to create new projects when the limit is reached.
- Scoped ID Integration: Refactored the system to use scoped IDs for better project management.
Bug Fixes
AI-Engine Fixes
- Prediction Function Improvements: Added a function to handle deconditioning in predictions.
- Decomposition Control: Improved setup for decomposition with enhanced control over conditioning parameters.
- Negative Binomial Parameterization: Corrected issues with the negative binomial parameterization.
- Hierarchy Checks: Skipped direct and indirect computations if no hierarchy is present.
- Graph State Handling: Allowed passing of graph states to decomposition processes.
- Test Speed Enhancements: Worked on speeding up test processes.
- Decomposition Fallback Functions: Added fallback functions for reference points in decomposition.
- Variable Base Correction: Removed outdated variable base corrections.
- Seed Management: Ensured positive seeds and protected against excessively large seeds.
- Bulk Prediction: Added functionality to freeze nodes in bulk predictions.
- Progress Updates: Provided proper progress updates for decomposition processes.
- Prior vs. Posterior Interface: Redesigned the prior vs. posterior split for a more reasonable interface.
- Simulation Effect Model: Initial work on a more flexible simulation effect model, including baseline and target control.
- Mathematical Functions: Added various mathematical nodes, including Weibull CDF, log, and log-logistic CDF functions.
API Fixes
- Variable Stats Migration: Migrated variable stats and interpolation configurations.
- Data Handling Improvements: Improved handling of data fetching, including support for slugs and retrospective effects.
- Dashboard Enhancements: Updated attribution and commercial dashboards with new triggers and functions.
- CSV Data Download: Enabled CSV data downloads.
- Error Handling: Enhanced error handling for revenue and profit calculations, handling of missing data, and project limits.
Teams and User Management Fixes
- Member Limit Validation: Added validation for member limits and handled errors when retrieving data for deleted users.
User Interface Fixes
- Console Cleanup: Removed unnecessary console logs.
- Dropdown Improvements: Improved styles and functionality for search dropdowns.
- Project Limit Messages: Updated project limit error messages.
- Search Enhancements: Introduced fuzzy search for model and dataset selection in simulation predictions and optimization.
User Management
- Auth0 Integration: Updated to Auth0 v4 and locked dependency versions for FastAPI and Pydantic with Auth0.
These enhancements and fixes aim to provide a more robust and user-friendly experience for marketing mix modeling, ensuring accurate analysis and seamless project management.