What is the DataHub API?
The DataHub API is an official API provided by Machbase DataHub that enables easy extraction, manipulation, and transformation of replicated data.
By utilizing this API, you can access data in any table and tag within Machbase in a standardized way. Moreover, since programs written in various programming languages can access it in the same manner, you can develop applications based on a transparent and standardized interface.
We plan to continuously add new types of APIs and functionalities. The official APIs are currently located in the datahub folder at:
https://github.com/machbase/datahub/tree/main/api/v2
All APIs are broadly categorized into five areas.
1. EDA(Exploratory Data Analysis)
The process of understanding the characteristics of the data is supported. It helps in better understanding the data through an overall overview, visual, and statistical analysis. This helps shorten data analysis time and improve prediction accuracy.
https://github.com/machbase/datahub/tree/main/api/v2/eda
Name | Function | Description |
---|---|---|
Data_Info | Display basic information about the data | Link |
Visualize_EDA | Display visual information about the data | Link |
statistical_test | Statistical test for the data | Link |
2. Model
It supports calling and using various AI models. The use of models is simple, and flexible utilization is possible through multiple options. By simplifying the complex model implementation process, it enables faster and more effective AI modeling.
https://github.com/machbase/datahub/tree/main/api/v2/model
Name | Function | Description |
---|---|---|
BiLSTM | Time Series model | Link |
DLinear | Time Series Forecasting model | Link |
PatchMixer | Time Series Forecasting model | Link |
SparseTSF | Time Series Forecasting model | Link |
TCN | Time Series model | Link |
TimeLinear | Time Series Forecasting model | Link |
LSTMAE | Time Series model | Link |
ResNet1d | Time Series model | Link |
3. Preprocessing
It supports the data preprocessing process. Based on the characteristics of the data identified during EDA, it helps to easily perform preprocessing tasks.
https://github.com/machbase/datahub/tree/main/api/v2/Preprocessing
Name | Function | Description |
---|---|---|
TimeFeatureGenerator | Creates time features using the time index | Link |
MinMaxScaler | Normalization function using the maximum and minimum values | Link |
FFT(Fast Fourier Transform) | Apply Hanning window and FFT transformation | Link |
4. TQL
It supports the integration between Machbase Neo and Python, a feature from the previous API v1.0. This enables efficient data integration.
https://github.com/machbase/datahub/tree/main/api/v2/tql
Name | Function | Description |
---|---|---|
select-rawdata | Select rawdata in Machbase Neo | Link |
select-rollup | Select Statistics values in Machbase Neo | Link |
select-scale | Select Minimum and maximum values in Machbase Neo | Link |
get-tag-names | Select Tag Table's Tag Name in Machbase Neo | Link |
5. Util
It supports a variety of auxiliary functions that can be useful in the AI development process. It helps efficiently handle tasks such as data loading and result visualization. As a result, developers can automate or simplify additional tasks that would otherwise need to be manually handled.
https://github.com/machbase/datahub/tree/main/api/v2/util
Name | Function | Description |
---|---|---|
show_column | Display Tag Table's Tag Name using get-tag-names.tql | Link |
data_load | Load Tag Table data in Machbase Neo | Link |
compare_graph | Display graph the comparison between the model's predictions and the actual values | Link |
set_minmax_value | Set Minimum and maximum values using select-scale.tql | Link |