p8n-importer 1.0.2

A tool to import various dataset formats and upload them to the PropulsionAI platform.

p8n-importer

Overview

p8n-importer is a versatile dataset import tool designed to simplify the process of importing various dataset formats into the PropulsionAI platform. This tool supports multiple formats, including VOC, YOLO, COCO, Image Classification, and Tabular data, and enables seamless integration with the PropulsionAI ecosystem. The tool now includes enhanced functionalities such as verbose logging and dataset visualization before uploading.

Explore more about PropulsionAI at PropulsionHQ.

Features

• Multiple Format Support: Easily import datasets in formats like VOC, YOLO, COCO, Image Classification, and Tabular.
• Direct Upload: Upload datasets directly to the PropulsionAI platform with an easy-to-use command-line interface.
• Flexibility: Extendable to support additional dataset formats in the future.
• Secure API Key Handling: API keys are handled securely via environment variables or interactive input.
• Verbose Logging: Get detailed logs of the import process with the --verbose option.
• Visualization: Preview the dataset conversion result before uploading with the --visualize option.

Installation

To install `p

8n-importer, you need Python 3.x and pip installed on your system. You can install p8n-importer` directly from PyPI:

pip install p8n-importer

Usage

Command-Line Interface

p8n-importer can be executed from the command line. Here’s how you can use it:

p8n-importer [format] [source_folder] [--verbose] [--no-upload] [--visualize]

• [format]: The format of your dataset (e.g., voc, yolov8, coco_json, im_classification, tabular).
• [source_folder]: Path to the source dataset folder.
• --verbose (optional): Enable verbose logging for detailed information during the import process.
• --no-upload (optional): Skip uploading the converted dataset to the platform.
• --visualize (optional): Visualize the dataset conversion result before uploading.

Programmatic Use

p8n-importer also supports being used programmatically as shown in the example below:

from P8nImporter import P8nImporter

importer = P8nImporter(
    api_key="YOUR_API_KEY",
    dataset_id="DATASET_ID",
)

image = "path/to/image.jpg"
annotation_path = "path/to/annotation.txt"

with open(annotation_path, "r") as f:
    annotation = f.read() # should be a string (multiline is supported)

label_names = ["label1", "label2"]

importer.import_task(
    format="yolov8",
    image=image,
    annotation=annotation,
    label_names=label_names,
)

Supported Formats

The following table lists the formats currently supported by p8n-importer, along with their respective format codes:

Format Code	Format	Input(s)	Action(s)	Description
voc	VOC	Image	Object Detection	Visual
yolov8	YOLOv8	Image	Object Detection	You Only Look Once, version 8
coco_json	COCO JSON	Image	Object Detection	Common Objects in Context, JSON format
im_classification	Image Classification	Image	Classification	Generic image classification datasets
tabular	Tabular	Tabular	Classification, Regression	Datasets in tabular formats like CSV, Excel, Parquet
iam	IAM	Image	Character Recognition	txt file with format "abc.jpg ABC" on each line. images in "image" folder

More formats are planned for future releases.

Visualization Feature (Only supported on CLI)

When using the --visualize flag, you can preview how the dataset will look after conversion. This feature is particularly useful to verify annotations and dataset integrity before uploading it to the PropulsionAI platform.

API Key and Dataset ID

The tool will prompt you for the API key and dataset ID. The API key can also be set as an environment variable PROPULSIONAI_API_KEY.

Contributing

Contributions to p8n-importer are welcome! If you're looking to contribute, please read our Contributing Guidelines.

License

p8n-importer is available under the MIT license. See the LICENSE file for more info.

Contact

For support or any questions, feel free to contact us at [e-mail].