Examples¶
This page demonstrates various ways to use TextBaker programmatically with visual outputs.
All examples can be found in the examples/ folder.
1. Basic Text Generation¶
Generate text images with default settings using characters from your dataset.
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import DatasetConfig, TransformConfig
config = GeneratorConfig(
seed=42,
dataset=DatasetConfig(dataset_dir="assets/dataset"),
transform=TransformConfig(
rotation_range=(0, 0),
scale_range=(1.0, 1.0),
),
)
generator = TextGenerator(config)
result = generator.generate("12345")
cv2.imwrite("example_basic.png", result.image)
Output:
2. With Transformations¶
Apply rotation, perspective, and scale transformations.
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import (
DatasetConfig, TransformConfig, CharacterConfig
)
config = GeneratorConfig(
seed=123,
dataset=DatasetConfig(dataset_dir="assets/dataset"),
transform=TransformConfig(
rotation_range=(-20, 20),
perspective_range=(0.0, 0.08),
scale_range=(0.9, 1.1),
),
character=CharacterConfig(
font="FONT_HERSHEY_DUPLEX",
font_scale=2.0,
),
)
generator = TextGenerator(config)
result = generator.generate("67890")
cv2.imwrite("example_rotated.png", result.image)
Output:
3. Random Colors¶
Generate text with random colors per character.
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import (
DatasetConfig, ColorConfig, CharacterConfig
)
config = GeneratorConfig(
seed=456,
dataset=DatasetConfig(dataset_dir="assets/dataset"),
color=ColorConfig(
random_color=True,
color_range_r=(100, 255),
color_range_g=(50, 200),
color_range_b=(50, 255),
),
character=CharacterConfig(
font="FONT_HERSHEY_SCRIPT_COMPLEX",
font_scale=2.5,
thickness=2,
),
)
generator = TextGenerator(config)
result = generator.generate("Hello")
cv2.imwrite("example_colored.png", result.image)
Output:
4. With Background Image¶
Place generated text on background images.
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import (
DatasetConfig, BackgroundConfig,
ColorConfig, CharacterConfig
)
config = GeneratorConfig(
seed=789,
dataset=DatasetConfig(dataset_dir="assets/dataset"),
background=BackgroundConfig(
enabled=True,
background_dir="assets/backgrounds",
),
color=ColorConfig(
random_color=False,
fixed_color=(255, 255, 255), # White text
),
character=CharacterConfig(
font="FONT_HERSHEY_TRIPLEX",
font_scale=3.0,
thickness=2,
),
)
generator = TextGenerator(config)
result = generator.generate("World")
cv2.imwrite("example_background.png", result.image)
Output:
5. With Texture Overlay¶
Apply textures to the generated text.
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import (
DatasetConfig, TextureConfig,
TransformConfig, CharacterConfig
)
config = GeneratorConfig(
seed=555,
dataset=DatasetConfig(dataset_dir="assets/dataset"),
texture=TextureConfig(
enabled=True,
texture_dir="assets/textures",
opacity=0.7,
),
transform=TransformConfig(
rotation_range=(-5, 5),
scale_range=(1.0, 1.0),
),
character=CharacterConfig(
font="FONT_HERSHEY_COMPLEX",
font_scale=2.5,
),
)
generator = TextGenerator(config)
result = generator.generate("Baker")
cv2.imwrite("example_texture.png", result.image)
Output:
6. Full Pipeline¶
Combine transforms, colors, textures, and backgrounds.
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import (
DatasetConfig, TransformConfig, ColorConfig,
TextureConfig, BackgroundConfig, CharacterConfig
)
config = GeneratorConfig(
seed=999,
dataset=DatasetConfig(dataset_dir="assets/dataset"),
transform=TransformConfig(
rotation_range=(-15, 15),
perspective_range=(0, 0.05),
scale_range=(0.95, 1.05),
),
color=ColorConfig(
random_color=True,
color_range_r=(150, 255),
color_range_g=(100, 200),
color_range_b=(50, 150),
),
texture=TextureConfig(
enabled=True,
texture_dir="assets/textures",
opacity=0.4,
),
background=BackgroundConfig(
enabled=True,
background_dir="assets/backgrounds",
),
character=CharacterConfig(
font="FONT_HERSHEY_SCRIPT_SIMPLEX",
font_scale=2.0,
thickness=2,
),
)
generator = TextGenerator(config)
result = generator.generate("TextBaker")
cv2.imwrite("example_full.png", result.image)
Output:
Batch Generation¶
Generate many random samples for training:
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import (
DatasetConfig, TransformConfig,
TextureConfig, BackgroundConfig
)
from pathlib import Path
config = GeneratorConfig(
seed=303,
text_length=(4, 8), # Random length between 4 and 8
dataset=DatasetConfig(dataset_dir="assets/dataset"),
transform=TransformConfig(
rotation_range=(-10, 10),
scale_range=(0.9, 1.1),
),
texture=TextureConfig(
enabled=True,
texture_dir="assets/textures",
opacity=0.7,
),
background=BackgroundConfig(
enabled=True,
background_dir="assets/backgrounds",
),
)
generator = TextGenerator(config)
output_dir = Path("batch_output")
output_dir.mkdir(exist_ok=True)
# Generate 100 random samples
for i in range(100):
result = generator.generate_random()
output_path = output_dir / f"{i:04d}_{result.text}.png"
cv2.imwrite(str(output_path), result.image)
print(f"Generated 100 samples in {output_dir}")
Using Config Files¶
Save and load configurations from YAML/JSON:
from textbaker import TextGenerator, GeneratorConfig
from textbaker.core.configs import (
DatasetConfig, TextureConfig, BackgroundConfig
)
# Create and save config
config = GeneratorConfig(
seed=999,
spacing=2,
dataset=DatasetConfig(dataset_dir="assets/dataset"),
texture=TextureConfig(
enabled=True,
texture_dir="assets/textures",
opacity=0.8,
),
background=BackgroundConfig(
enabled=True,
background_dir="assets/backgrounds",
),
)
# Save to YAML
config.to_file("my_config.yaml")
# Later, load and use
loaded_config = GeneratorConfig.from_file("my_config.yaml")
generator = TextGenerator(loaded_config)
result = generator.generate("FromConfig")
CLI Examples¶
Generate images from the command line:
# Basic generation
textbaker generate "hello" "world"
# With transforms
textbaker generate "test" --rotation "-15,15" --perspective "0.0,0.1"
# With background images
textbaker generate "test" -b ./assets/backgrounds
# With textures
textbaker generate "test" -t ./assets/textures --texture-opacity 0.8
# Random colors
textbaker generate "test" --random-color
# Full pipeline
textbaker generate "hello" \
-r "-10,10" \
-p "0,0.05" \
-b ./assets/backgrounds \
-t ./assets/textures \
--random-color
# Generate 50 random samples
textbaker generate -n 50 --seed 42
Generate Example Images¶
To regenerate the example images shown above, run:
Output will be saved to assets/readme/.