rdlab-dataset 0.4.4

No project description provided

📚 Rdlab Dataset

Overview

Rdlab Dataset is a collection of Khmer datasets made for research & development (R&D) in Cambodia. It provides ready-to-use datasets and utilities for text and image processing.

Datasets included:

• Khmer Words
• Khmer Addresses
• Khmer Sentences
• Khmer Combination Locations
• Khmer-English Combined Dataset
• Text-to-Image Generators (with noise)

---

📦 Dataset Statistics (v0.3.0)

Dataset Name	Number of Records
Khmer Word Dataset	784,011
Khmer Address Dataset	20,817
Khmer Sentence Dataset	211,928
Combination Location in Cambodia	200,000
Khmer English Dataset	545,941

---

📥 Installation

Install using pip:

pip install rdlab_dataset

Best example

from rdlab_dataset.module import KhmerDatasetLoader, TextArrayListImageGenerator

# Load datasets
word_loader = KhmerDatasetLoader("word")
address_loader = KhmerDatasetLoader("address")
sentence_loader = KhmerDatasetLoader("sentence")

# Get data
words = word_loader.get_all()
addresses = address_loader.get_all()
sentences = sentence_loader.get_all()

# Combine into one array
combined_texts = words + addresses + sentences
print(f"Total combined items: {len(combined_texts)}")

# Image generator
text_image_gen = TextArrayListImageGenerator(
    customize_font=True,
    folder_limit=10,
    output_count=4,
    num_threads=4
)

# Generate images from address list
text_image_gen.generate_images(
    text_list=addresses,
    font_folder="/home/vitoupro/code/rdlab-dataset/test_font"
)

🧠 Module Overview

• KhmerDatasetLoader — Unified loader for all dataset types
• ATextImageGenerator — Generate noisy images from single text
• TextArrayListImageGenerator — Generate noisy images from a list with annotations

🔍 Usage

📝 KhmerDatasetLoader

from rdlab_dataset.module import KhmerDatasetLoader

# Supported types: "word", "address", "sentence", "location", "khmer_english"
loader = KhmerDatasetLoader("word")

# Access methods
all_data = loader.get_all()
first = loader.get_first()
first_five = loader.get_n_first(5)
exists = loader.find("សេចក្ដី")
length = len(loader)

print(first, first_five, exists, length)

🖼 TextArrayListImageGenerator (Text List with Annotations)

from rdlab_dataset.module import TextArrayListImageGenerator

texts = ["ខ្ញុំស្រឡាញ់វេយ្យាករណ៍", "ភាសាខ្មែរ"]
gen = TextArrayListImageGenerator(customize_font=False)
gen.generate_images(texts, save_as_pickle=False, output_count=3)

✅ Features

• Easy dataset loading (.pkl format)
• Support .ttf font and .jpg background assets
• Built-in noisy image generation (Gaussian, Speckle, Salt & Pepper, etc.)
• Clean data APIs: get all, first, or n-first records
• Lightweight, fast, and ready for machine learning pipelines

🤝 Contributing

• Fork, commit, and submit a pull request. All contributions welcome!

📜 License

Rdlab Community License

• 📲 Telegram: 0964060587