https://github.com/edgeimpulse/transformation-blocks
Repository that contains transformation blocks examples
https://github.com/edgeimpulse/transformation-blocks
Last synced: 12 months ago
JSON representation
Repository that contains transformation blocks examples
- Host: GitHub
- URL: https://github.com/edgeimpulse/transformation-blocks
- Owner: edgeimpulse
- License: bsd-3-clause-clear
- Created: 2023-09-20T16:10:29.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2025-02-10T10:40:01.000Z (over 1 year ago)
- Last Synced: 2025-06-22T02:12:18.030Z (12 months ago)
- Language: Python
- Size: 9.21 MB
- Stars: 4
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Transformation block examples
This repository contains examples of Edge Impulse transformation blocks.
For more information on how to set up a transformation block, please head to Edge Impulse documentation, [Custom blocks -> Transformation blocks](https://docs.edgeimpulse.com/docs/edge-impulse-studio/organizations/custom-blocks/transformation-blocks)
## Examples
Name
Description
Mode
Programming language
Public docker image (optional)
Hello Edge
print hello +name on the transformation job logs
Standalone
Bash
edgeimpulse/ei-transform_hello-edge:latest
Utils - Access data
Utility example to explain how to access data
Standalone | Data item | File
Python
edgeimpulse/ei-transform_utils-access-data:latest
Fetch Kaggle Dataset
Import a dataset hosted on Kaggle to your bucket
Standalone
Python
edgeimpulse/ei_transform_fetch-kaggle-dataset:latest
Create graphs (Standalone)
Generate a helper graph from sensor CSV
Standalone
Python
edgeimpulse/ei-transform_create-graphs-standalone
Create graphs (in-file)
Generate a helper graph from sensor CSV
File
Python
edgeimpulse/ei-transform_create-graphs-standalone-in-file
Check file existance
Add ei_check metadata on file existance
Data item
Python
edgeimpulse/ei-transform_check-files-exist
Merge CSV
Merge CSV files on a given key
Data item
Python
edgeimpulse/ei-transform_merge-csv
Merge Audio and CSV
Merge audio file and time-series csv
Data item
Python
edgeimpulse/ei-transform_merge-audio-time-series-csv
Resample CSV
Upsample or downsample CSV files with a constant frequency
File
Python
edgeimpulse/ei-transform_resample-csv
Split CSV
Split CSV files into smaller CSV files defined by a split interval
File
Python
edgeimpulse/ei-transform_split-csv
Sync SCP Dataset
Fetch files using scp and sync with organization dataset
Standalone
Python
edgeimpulse/ei-transform_sync-scp-dataset
## Bring the examples to your organization
### Using the Docker Image
To use the examples, a README.md instruction file will be provided inside the example repositories. Sometimes, a public docker image will also be provided to quickly test so you don't need to import it and clone the entire repository.
If you want to quickly test the transformation blocks, head to your Edge Impulse organizaton, go to **Custom blocks**->**Transformation** and click on **+ Add new transformation block**:
Replace the **Docker Container** by the one provided in the example above, the operation type and the optional parameters (check the examples README.md instruction to set the parameters):
.
### Using Edge Impulse CLI
To modify the code and push the blocks to your organization, you will need to use [Edge Impulse CLI](https://docs.edgeimpulse.com/docs/tools/edge-impulse-cli).
1. Clone this repository:
```
git clone https://github.com/edgeimpulse/transformation-blocks.git
```
2. Navigate to the desired example:
```
cd transformation-blocks/hello-edge
```
3. Create the transformation block:
```
edge-impulse-block init
```
4. Push the transformation block to your organization:
```
edge-impulse-block push
```
## Contributing to this repository
We welcome contributions to this repository. Both improvements to our own transformation blocks, as well as new and well-tested transformation blocks for other use cases. Make sure to provide a public dataset - or subset of this dataset - so everyone can reproduce your workflow seamlessly.