* [improvement][semantic-fe] Add model alias setting & Add view permission restrictions to the model permission management tab. [improvement][semantic-fe] Add permission control to the action buttons for the main domain; apply high sensitivity filtering to the authorization of metrics/dimensions. [improvement][semantic-fe] Optimize the editing mode in the dimension/metric/datasource components to use the modelId stored in the database for data, instead of relying on the data from the state manager. * [improvement][semantic-fe] Add time granularity setting in the data source configuration. * [improvement][semantic-fe] Dictionary import for dimension values supported in Q&A visibility * [improvement][semantic-fe] Modification of data source creation prompt wording" * [improvement][semantic-fe] metric market experience optimization * [improvement][semantic-fe] enhance the analysis of metric trends * [improvement][semantic-fe] optimize the presentation of metric trend permissions * [improvement][semantic-fe] add metric trend download functionality * [improvement][semantic-fe] fix the dimension initialization issue in metric correlation * [improvement][semantic-fe] Fix the issue of database changes not taking effect when creating based on an SQL data source. * [improvement][semantic-fe] Optimizing pagination logic and some CSS styles * [improvement][semantic-fe] Fixing the API for the indicator list by changing "current" to "pageNum" * [improvement][semantic-fe] Fixing the default value setting for the indicator list * [improvement][semantic-fe] Adding batch operations for indicators/dimensions/models * [improvement][semantic-fe] Replacing the single status update API for indicators/dimensions with a batch update API * [improvement][semantic-fe] Redesigning the indicator homepage to incorporate trend charts and table functionality for indicators * [improvement][semantic-fe] Optimizing the logic for setting dimension values and editing data sources, and adding system settings functionality * [improvement][semantic-fe] Upgrading antd version to 5.x, extracting the batch operation button component, optimizing the interaction for system settings, and expanding the configuration generation types for list-to-select component. * [improvement][semantic-fe] Adding the ability to filter dimensions based on whether they are tags or not. * [improvement][semantic-fe] Adding the ability to edit relationships between models in the canvas. * [improvement][semantic-fe] Updating the datePicker component to use dayjs instead. * [improvement][semantic-fe] Fixing the issue with passing the model ID for dimensions in the indicator market. * [improvement][semantic-fe] Fixing the abnormal state of the popup when creating a model. * [improvement][semantic-fe] Adding permission logic for bulk operations in the indicator market. * [improvement][semantic-fe] Adding the ability to download and transpose data. * [improvement][semantic-fe] Fixing the initialization issue with the date selection component in the indicator details page when switching time granularity. * [improvement][semantic-fe] Fixing the logic error in the dimension value setting. * [improvement][semantic-fe] Fixing the synchronization issue with the question and answer settings information. * [improvement][semantic-fe] Optimizing the canvas functionality for better performance and user experience. * [improvement][semantic-fe] Optimizing the update process for drawing model relationship edges in the canvas. * [improvement][semantic-fe] Changing the line type for canvas connections. * [improvement][semantic-fe] Replacing the initialization variable from "semantic" to "headless". * [improvement][semantic-fe] Fixing the missing migration issue for default drill-down dimension configuration in model editing. Additionally, optimizing the data retrieval method for initializing fields in the model. * [improvement][semantic-fe] Updating the logic for the fieldName. * [improvement][semantic-fe] Adjusting the position of the metrics tab. * [improvement][semantic-fe] Changing the 字段名称 to 英文名称. * [improvement][semantic-fe] Fix metric measurement deletion. * [improvement][semantic-fe] UI optimization for metric details page. * [improvement][semantic-fe] UI optimization for metric details page. * [improvement][semantic-fe] UI adjustment for metric details page. * [improvement][semantic-fe] The granularity field in the time type of model editing now supports setting it as empty. * [improvement][semantic-fe] Added field type and metric type to the metric creation options. * [improvement][semantic-fe] The organization structure selection feature has been added to the permission management. * [improvement][semantic-fe] Improved user experience for the metric list. * [improvement][semantic-fe] fix update the metric list. * [improvement][headless-fe] Added view management functionality. * [improvement][headless-fe] The view management functionality has been added. This feature allows users to create, edit, and manage different views within the system. * [improvement][headless-fe] Added model editing side effect detection. * [improvement][headless-fe] Fixed the logic error in view editing. * [improvement][headless-fe] Fixed the issue with initializing dimension associations in metric settings. * [improvement][headless-fe] Added the ability to hide the Q&A settings entry point. * [improvement][headless-fe] Fixed the issue with selecting search results in metric field creation. * [improvement][headless-fe] Added search functionality to the field list in model editing. * [improvement][headless-fe] fix the field list in model editing * [improvement][headless-fe] Restructured the data for the dimension value settings interface. * [improvement][headless-fe] Added dynamic variable functionality to model creation based on SQL scripts. * [improvement][headless-fe] Added support for passing dynamic variables as parameters in the executeSql function. * [improvement][headless-fe] Resolved the issue where users were unable to select all options for dimensions, metrics, and fields in the metric generation process. * [improvement][headless-fe] Replaced the term "view" with "dataset" * [improvement][headless-fe] Added the ability to export metrics and dimensions to a specific target. * [improvement][headless-fe] Enhanced dataset creation to support the tag mode. * [improvement][headless-fe] Added tag value setting. * [improvement][headless-fe] Optimized the tag setting system. * [improvement][headless-fe] Optimized the tag setting system. * [improvement][headless-fe] Updated the data initialization for model editing to use API requests instead. * [improvement][headless-fe] Added search functionality to model management. * [improvement][headless-fe] Removed field null validation during model editing. * [improvement][headless-fe] Updated the batch operation button component. * [improvement][headless-fe] Optimized the logic for initializing indicators in dimension value settings. * [improvement][headless-fe] Adjusted the length of the input field for model editing names. * [improvement][headless-fe] Lock the version of the @ant-design/pro-table component and replace it with @ant-design/pro-components. * [improvement][headless-fe] Optimized the style of the metrics market and tags market. * [improvement][headless-fe] The quick creation of model fields now defaults to using the "comment" field for filling. * [improvement][headless-fe] The quick creation of model fields now defaults to using the "comment" field for filling * [improvement][headless-fe] The quick creation of model fields now defaults to using the "comment" field for filling. * [improvement][headless-fe] Fixed the issue where the conditions for metric measurement creation were not being saved correctly. * [improvement][headless-fe] Default value setting for hiding dimensions. * [improvement][headless-fe] Updated the file imports in the project. * [improvement][headless-fe] Adjusted the logic for displaying the tab in the theme domain. * [improvement][headless-fe] Added term management functionality. * [improvement][headless-fe] When creating a model, the current metric operator now allows for clearance. * [improvement][headless-fe] Term management interface transformation * [improvement][headless-fe] Migrating scaffold version to @umi/max * [improvement][headless-fe] remove modle column * [improvement][headless-fe] 1.Added configuration for the large language model in the agent; 2.upgraded React version from 17 to 18; 3.modified some UI effects. * [improvement][headless-fe] Added a simplified mode to the question-answering system.
SuperSonic
SuperSonic is the next-generation BI platform that integrates Chat BI (powered by LLM) and Headless BI (powered by semantic layer) paradigms. This integration ensures that Chat BI has access to the same curated and governed semantic data models as traditional BI. Furthermore, the implementation of both paradigms benefits from the integration:
- Chat BI's Text2SQL gets augmented with context-retrieval from semantic models.
- Headless BI's query interface gets extended with natural language API.
SuperSonic provides a Chat BI interface that empowers users to query data using natural language and visualize the results with suitable charts. To enable such experience, the only thing necessary is to build logical semantic models (definition of metric/dimension/tag, along with their meaning and relationships) through a Headless BI interface. Meanwhile, SuperSonic is designed to be extensible and composable, allowing custom implementations to be added and configured with Java SPI.
Motivation
The emergence of Large Language Model (LLM) like ChatGPT is reshaping the way information is retrieved, leading to a new paradigm in the field of data analytics known as Chat BI. To implement Chat BI, both academia and industry are primarily focused on harnessing the power of LLMs to convert natural language into SQL, commonly referred to as Text2SQL or NL2SQL. While some approaches show promising results, their reliability falls short for large-scale real-world applications.
Meanwhile, another emerging paradigm called Headless BI, which focuses on constructing unified semantic data models, has garnered significant attention. Headless BI is implemented through a universal semantic layer that exposes consistent data semantics via an open API.
From our perspective, the integration of Chat BI and Headless BI has the potential to enhance the Text2SQL generation in two dimensions:
- Incorporate data semantics (such as business terms, column values, etc.) into the prompt, enabling LLM to better understand the semantics and reduce hallucination.
- Offload the generation of advanced SQL syntax (such as join, formula, etc.) from LLM to the semantic layer to reduce complexity.
With these ideas in mind, we develop SuperSonic as a practical reference implementation and use it to power our real-world products. Additionally, to facilitate further development we decide to open source SuperSonic as an extensible framework.
Out-of-the-box Features
- Built-in Chat BI interface for business users to enter natural language queries
- Built-in Headless BI interface for analytics engineers to build semantic data models
- Built-in rule-based semantic parser to improve efficiency in certain scenarios (e.g. demonstration, integration testing)
- Built-in support for input auto-completion, multi-turn conversation as well as post-query recommendation
- Built-in support for three-level data access control: dataset-level, column-level and row-level
Extensible Components
The high-level architecture and main process flow is as follows:
-
Knowledge Base: extracts schema information periodically from the semantic models and build dictionary and index to facilitate schema mapping.
-
Schema Mapper: identifies references to schema elements(metrics/dimensions/entities/values) in user queries. It matches the query text against the knowledge base.
-
Semantic Parser: understands user queries and generates semantic query statement. It consists of a combination of rule-based and model-based parsers, each of which deals with specific scenarios.
-
Semantic Corrector: checks validity of semantic query statement and performs correction and optimization if needed.
-
Semantic Translator: converts semantic query statement into SQL statement that can be executed against physical data models.
-
Chat Plugin: extends functionality with third-party tools. The LLM is going to select the most suitable one, given all configured plugins with function description and sample questions.
Quick Demo
SuperSonic comes with sample semantic models as well as chat conversations that can be used as a starting point. Please follow the steps:
- Download the latest prebuilt binary from the release page
- Run script "assembly/bin/supersonic-daemon.sh start" to start a standalone Java service
- Visit http://localhost:9080 in the browser to start exploration
Build and Development
Please refer to project wiki.
WeChat Contact
Please follow SuperSonic wechat official account: