TextRecognitionDataGenerator

https://textrecognitiondatagenerator.readthedocs.io/en/latest/
Admin
A tool for generating synthetic text data for text recognition with support for multiple languages,...
Tokens:7,111
Snippets:89
Trust Score:9.6
Update:4 months ago
Show doc for...
Tokens:
Raw
### Install Base Dependencies from Source

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/installation.html

Install the required dependencies for TextRecognitionDataGenerator when installing from source. It is recommended to use a virtual environment.

```bash
pip3 install -r requirements.txt
```

--------------------------------

### Install Handwritten Text Dependencies from Source

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/installation.html

Install the additional dependencies required for handwritten text generation when installing from source.

```bash
pip3 install -r requirements-hw.txt
```

--------------------------------

### Generate Text with Tight Crop (Fit)

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/tutorial.md.txt

Generates a single text example and applies a tight crop around the rendered text, removing extra space for accents. Use this to minimize whitespace.

```bash
python3 run.py -c 1  -i texts/test.txt --fit
```

--------------------------------

### Install Official Package

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/installation.html

Install the TextRecognitionDataGenerator package using pip. This command makes the 'trdg' binary available in your system's PATH.

```bash
pip install trdg
```

--------------------------------

### Generate Specific Number of English Examples

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates a specified number of English, single-word images. Use the -c argument to set the count.

```bash
python3 run.py -c 10
```

--------------------------------

### Case Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Generates text in either upper or lowercase only. Example: --case upper.

```python
-ca [CASE], --case [CASE]
                        Generate upper or lowercase only. arguments: upper or
                        lower. Example: --case upper
```

--------------------------------

### TRDG CLI Usage Help

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

This snippet shows the general usage of the TRDG command-line tool. Use the -h flag for detailed help on specific arguments.

```bash
usage: trdg [-h] [--output_dir [OUTPUT_DIR]] [-i [INPUT_FILE]] [-l [LANGUAGE]]
            -c [COUNT] [-rs] [-let] [-num] [-sym] [-w [LENGTH]] [-r]
            [-f [FORMAT]] [-t [THREAD_COUNT]] [-e [EXTENSION]]
            [-k [SKEW_ANGLE]] [-rk] [-wk] [-bl [BLUR]] [-rbl]
            [-b [BACKGROUND]] [-hw] [-na NAME_FORMAT] [-d [DISTORSION]]
            [-do [DISTORSION_ORIENTATION]] [-wd [WIDTH]] [-al [ALIGNMENT]]
            [-or [ORIENTATION]] [-tc [TEXT_COLOR]] [-sw [SPACE_WIDTH]]
            [-cs [CHARACTER_SPACING]] [-m [MARGINS]] [-fi] [-ft [FONT]]
            [-ca [CASE]]

Generate synthetic text data for text recognition.

optional arguments:
  -h, --help            show this help message and exit
  --output_dir [OUTPUT_DIR]
                        The output directory
  -i [INPUT_FILE], --input_file [INPUT_FILE]
                        When set, this argument uses a specified text file as
                        source for the text
  -l [LANGUAGE], --language [LANGUAGE]
                        The language to use, should be fr (French), en
                        (English), es (Spanish), de (German), or cn (Chinese).
  -c [COUNT], --count [COUNT]
                        The number of images to be created.
  -rs, --random_sequences
                        Use random sequences as the source text for the
                        generation. Set '-let','-num','-sym' to use
                        letters/numbers/symbols. If none specified, using all
                        three.
  -let, --include_letters
                        Define if random sequences should contain letters.
                        Only works with -rs
  -num, --include_numbers
                        Define if random sequences should contain numbers.
                        Only works with -rs
  -sym, --include_symbols
                        Define if random sequences should contain symbols.
                        Only works with -rs
  -w [LENGTH], --length [LENGTH]
                        Define how many words should be included in each
                        generated sample. If the text source is Wikipedia,
                        this is the MINIMUM length
  -r, --random          Define if the produced string will have variable word
                        count (with --length being the maximum)
  -f [FORMAT], --format [FORMAT]
                        Define the height of the produced images if
                        horizontal, else the width
  -t [THREAD_COUNT], --thread_count [THREAD_COUNT]
                        Define the number of thread to use for image
                        generation
  -e [EXTENSION], --extension [EXTENSION]
                        Define the extension to save the image with
  -k [SKEW_ANGLE], --skew_angle [SKEW_ANGLE]
                        Define skewing angle of the generated text. In
                        positive degrees
  -rk, --random_skew    When set, the skew angle will be randomized between
                        the value set with -k and it's opposite
  -wk, --use_wikipedia  Use Wikipedia as the source text for the generation, 
                        using this paremeter ignores -r, -n, -s
  -bl [BLUR], --blur [BLUR]
                        Apply gaussian blur to the resulting sample. Should be
                        an integer defining the blur radius
  -rbl, --random_blur   When set, the blur radius will be randomized between 0
                        and -bl.
  -b [BACKGROUND], --background [BACKGROUND]
                        Define what kind of background to use. 0: Gaussian
                        Noise, 1: Plain white, 2: Quasicrystal, 3: Pictures
  -hw, --handwritten    Define if the data will be "handwritten" by an RNN
  -na NAME_FORMAT, --name_format NAME_FORMAT
                        Define how the produced files will be named. 0:
                        [TEXT]_[ID].[EXT], 1: [ID]_[TEXT].[EXT] 2: [ID].[EXT]
                        + one file labels.txt containing id-to-label mappings
  -d [DISTORSION], --distorsion [DISTORSION]
                        Define a distorsion applied to the resulting image. 0:
                        None (Default), 1: Sine wave, 2: Cosine wave, 3:
                        Random
  -do [DISTORSION_ORIENTATION], --distorsion_orientation [DISTORSION_ORIENTATION]
                        Define the distorsion's orientation. Only used if -d
                        is specified. 0: Vertical (Up and down), 1: Horizontal
                        (Left and Right), 2: Both
  -wd [WIDTH], --width [WIDTH]
                        Define the width of the resulting image. If not set it
```

--------------------------------

### Fit Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -fi or --fit option to apply a tight crop around the rendered text.

```bash
  -fi, --fit            Apply a tight crop around the rendered text
```

--------------------------------

### Using Various Generator Types

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/module.html

Demonstrates importing and initializing different generator types: GeneratorFromDicts for bundled dictionaries, GeneratorFromRandom for random strings, and GeneratorFromWikipedia for Wikipedia article content. The generators will continuously produce data until the loop is broken.

```python
from trdg.generators import (
    GeneratorFromDicts,
    GeneratorFromRandom,
    GeneratorFromWikipedia,
)

generator_from_dicts = GeneratorFromDicts()
generator_from_random = GeneratorFromRandom()
generator_from_wikipedia = GeneratorFromWikipedia()

for img, lbl in generator_from_dicts:
    # Do something with the pillow image here.


```

--------------------------------

### TRDG Command-Line Help

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Displays the help message for the TRDG command-line tool, listing all available arguments and their descriptions. Use this to understand the full range of customization options.

```bash
usage: trdg [-h] [--output_dir [OUTPUT_DIR]] [-i [INPUT_FILE]] [-l [LANGUAGE]]
            -c [COUNT] [-rs] [-let] [-num] [-sym] [-w [LENGTH]] [-r]
            [-f [FORMAT]] [-t [THREAD_COUNT]] [-e [EXTENSION]]
            [-k [SKEW_ANGLE]] [-rk] [-wk] [-bl [BLUR]] [-rbl]
            [-b [BACKGROUND]] [-hw] [-na NAME_FORMAT] [-d [DISTORSION]]
            [-do [DISTORSION_ORIENTATION]] [-wd [WIDTH]] [-al [ALIGNMENT]]
            [-or [ORIENTATION]] [-tc [TEXT_COLOR]] [-sw [SPACE_WIDTH]]
            [-cs [CHARACTER_SPACING]] [-m [MARGINS]] [-fi] [-ft [FONT]]
            [-ca [CASE]]

Generate synthetic text data for text recognition.

optional arguments:
  -h, --help            show this help message and exit
  --output_dir [OUTPUT_DIR]
                        The output directory
  -i [INPUT_FILE], --input_file [INPUT_FILE]
                        When set, this argument uses a specified text file as
                        source for the text
  -l [LANGUAGE], --language [LANGUAGE]
                        The language to use, should be fr (French), en
                        (English), es (Spanish), de (German), or cn (Chinese).
  -c [COUNT], --count [COUNT]
                        The number of images to be created.
  -rs, --random_sequences
                        Use random sequences as the source text for the
                        generation. Set '-let','-num','-sym' to use
                        letters/numbers/symbols. If none specified, using all
                        three.
  -let, --include_letters
                        Define if random sequences should contain letters.
                        Only works with -rs
  -num, --include_numbers
                        Define if random sequences should contain numbers.
                        Only works with -rs
  -sym, --include_symbols
                        Define if random sequences should contain symbols.
                        Only works with -rs
  -w [LENGTH], --length [LENGTH]
                        Define how many words should be included in each
                        generated sample. If the text source is Wikipedia, 
                        this is the MINIMUM length
  -r, --random          Define if the produced string will have variable word
                        count (with --length being the maximum)
  -f [FORMAT], --format [FORMAT]
                        Define the height of the produced images if
                        horizontal, else the width
  -t [THREAD_COUNT], --thread_count [THREAD_COUNT]
                        Define the number of thread to use for image
                        generation
  -e [EXTENSION], --extension [EXTENSION]
                        Define the extension to save the image with
  -k [SKEW_ANGLE], --skew_angle [SKEW_ANGLE]
                        Define skewing angle of the generated text. In
                        positive degrees
  -rk, --random_skew    When set, the skew angle will be randomized between
                        the value set with -k and it's opposite
  -wk, --use_wikipedia  Use Wikipedia as the source text for the generation,
                        using this paremeter ignores -r, -n, -s
  -bl [BLUR], --blur [BLUR]
                        Apply gaussian blur to the resulting sample. Should be
                        an integer defining the blur radius
  -rbl, --random_blur   When set, the blur radius will be randomized between 0
                        and -bl.
  -b [BACKGROUND], --background [BACKGROUND]
                        Define what kind of background to use. 0: Gaussian
                        Noise, 1: Plain white, 2: Quasicrystal, 3: Pictures
  -hw, --handwritten    Define if the data will be "handwritten" by an RNN
  -na NAME_FORMAT, --name_format NAME_FORMAT
                        Define how the produced files will be named. 0:
                        [TEXT]_[ID].[EXT], 1: [ID]_[TEXT].[EXT] 2: [ID].[EXT]
                        + one file labels.txt containing id-to-label mappings
  -d [DISTORSION], --distorsion [DISTORSION]
                        Define a distorsion applied to the resulting image. 0:
                        None (Default), 1: Sine wave, 2: Cosine wave, 3:
                        Random
  -do [DISTORSION_ORIENTATION], --distorsion_orientation [DISTORSION_ORIENTATION]
                        Define the distorsion's orientation. Only used if -d
                        is specified. 0: Vertical (Up and down), 1: Horizontal
                        (Left and Right), 2: Both
  -wd [WIDTH], --width [WIDTH]
                        Define the width of the resulting image. If not set it
                        will be the width of the text + 10. If the width of

```

--------------------------------

### Generate Text with All Margins and Tight Crop

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with all margins set to 10px and tight cropping applied.

```bash
python3 run.py -c 1 -i texts/test.txt --margins 10,10,10,10 --fit
```

--------------------------------

### Fit Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Applies a tight crop around the rendered text.

```python
-fi, --fit            Apply a tight crop around the rendered text
```

--------------------------------

### Case Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -ca or --case option to generate text in either uppercase or lowercase.

```bash
  -ca [CASE], --case [CASE]
                        Generate upper or lowercase only. arguments: upper or
                        lower. Example: --case upper
```

--------------------------------

### Advanced Text Generation with Color and Blur

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with skewing, Gaussian blur, and random text colors between black and gray. -k for skew, -rk for random skew, -bl for blur, -rbl for random blur, -tc for text color.

```bash
python3 run.py -c 10 -k 15 -rk -bl 0.5 -rbl -tc '#000000,#888888'
```

--------------------------------

### Generate Text with Zero Margins and Tight Crop

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with zero margins (0,0,0,0) and a tight crop using --fit for absolutely no margins.

```bash
python3 run.py -c 1 -i texts/test.txt --margins 0,0,0,0 --fit
```

--------------------------------

### Margins Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -m or --margins option to set the margins around the text in pixels.

```bash
  -m [MARGINS], --margins [MARGINS]
                        Define the margins around the text when rendered. In
                        pixels
```

--------------------------------

### Font Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -ft or --font option to specify the font to be used for rendering text.

```bash
  -ft [FONT], --font [FONT]
                        Define font to be used
```

--------------------------------

### Initialize Multiple TRDG Generators

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/module.md.txt

Instantiate GeneratorFromDicts, GeneratorFromRandom, and GeneratorFromWikipedia. These generators provide different methods for obtaining text data for image generation.

```python
from trdg.generators import (
    GeneratorFromDicts,
    GeneratorFromRandom,
    GeneratorFromWikipedia,
)

generator_from_dicts = GeneratorFromDicts()
generator_from_random = GeneratorFromRandom()
generator_from_wikipedia = GeneratorFromWikipedia()

for img, lbl in generator_from_dicts:
    # Do something with the pillow image here.
```

--------------------------------

### Clone Repository

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/installation.html

Clone the official TextRecognitionDataGenerator repository from GitHub to use the tool from source.

```bash
git clone https://github.com/Belval/TextRecognitionDataGenerator
```

--------------------------------

### Generate Default English Data

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates 1000 English, single-word images in the 'out/' directory using default settings.

```bash
python3 run.py
```

--------------------------------

### Space Width Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -sw or --space_width option to control the width of spaces between words, with 2.0 indicating twice the normal width.

```bash
  -sw [SPACE_WIDTH], --space_width [SPACE_WIDTH]
                        Define the width of the spaces between words. 2.0
                        means twice the normal space width
```

--------------------------------

### Text Alignment Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -al or --alignment option to define text alignment (left, center, right) when a specific width is set.

```bash
  -al [ALIGNMENT], --alignment [ALIGNMENT]
                        Define the alignment of the text in the image. Only
                        used if the width parameter is set. 0: left, 1:
                        center, 2: right
```

--------------------------------

### Character Spacing Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -cs or --character_spacing option to define the width of spaces between characters in pixels.

```bash
  -cs [CHARACTER_SPACING], --character_spacing [CHARACTER_SPACING]
                        Define the width of the spaces between characters. 2
                        means two pixels
```

--------------------------------

### Generate Text with Default Margins

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with default margins (5, 5, 5, 5) and no fit. Reads text from 'texts/test.txt'.

```bash
python3 run.py -c 1 -i texts/test.txt
```

--------------------------------

### Generate Text with Tight Crop

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with a tight crop around the rendered text using the --fit argument.

```bash
python3 run.py -c 1 -i texts/test.txt --fit
```

--------------------------------

### Text Orientation Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -or or --orientation option to specify text orientation as horizontal (0) or vertical (1).

```bash
  -or [ORIENTATION], --orientation [ORIENTATION]
                        Define the orientation of the text. 0: Horizontal, 1:
                        Vertical
```

--------------------------------

### Generate Data with Sine Wave Distortion

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates data with sine wave distortion applied. Use -d 1 for sine wave distortion.

```bash
python3 run.py -c 5 -w 5 -d 1
```

--------------------------------

### Generate Text with Zero Margins and No Fit

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with zero margins (0,0,0,0) but still has apparent margins due to no fit. Use --margins to set margins.

```bash
python3 run.py -c 1 -i texts/test.txt --margins 0,0,0,0
```

--------------------------------

### Generate Images from Specific Strings

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/module.html

Use GeneratorFromStrings to create image-label pairs from a predefined list of strings. This is useful when you have a specific set of text you need to generate images for.

```python
from trdg.generators import GeneratorFromStrings

generator = GeneratorFromStrings(['Test1', 'Test2', 'Test3'])

for img in generator:
    # Do something with the pillow image here.


```

--------------------------------

### Advanced Text Generation with Increased Resolution

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with increased resolution (height). Use -f to set the output height.

```bash
python3 run.py -c 10 -k 15 -rk -bl 0.5 -rbl -tc '#000000,#888888' -f 64
```

--------------------------------

### Text Color Option

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/_sources/overview.md.txt

Use the -tc or --text_color option to set the text color, accepting a single hex color or a range.

```bash
  -tc [TEXT_COLOR], --text_color [TEXT_COLOR]
                        Define the text's color, should be either a single hex
                        color or a range in the ?,? format.
```

--------------------------------

### Generate Text with Vertical Margins and Tight Crop

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates text with specified vertical margins (10px top/bottom) and tight cropping. Margin values are comma-separated top,left,bottom,right.

```bash
python3 run.py -c 1 -i texts/test.txt --margins 10,0,10,0
```

--------------------------------

### Font Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Defines the font to be used for rendering text.

```python
-ft [FONT], --font [FONT]
                        Define font to be used
```

--------------------------------

### Generate Data with Random Distortion

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates data with random distortion applied. Use -d 3 for random distortion.

```bash
python3 run.py -c 5 -w 5 -d 3
```

--------------------------------

### Generate Chinese Data with Word Length

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates Chinese character images with a specified word length. Use -w to set the length.

```bash
python3 run.py -c 10 -l cn -w 5
```

--------------------------------

### Margins Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Defines the margins around the text when rendered, specified in pixels.

```python
-m [MARGINS], --margins [MARGINS]
                        Define the margins around the text when rendered. In
                        pixels
```

--------------------------------

### Text Orientation Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Defines the orientation of the text. Accepts 0 for horizontal and 1 for vertical orientation.

```python
-or [ORIENTATION], --orientation [ORIENTATION]
                        Define the orientation of the text. 0: Horizontal, 1:
                        Vertical
```

--------------------------------

### Text Alignment Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Defines the alignment of text in an image. Only used if the width parameter is set. Accepts 0 for left, 1 for center, and 2 for right alignment.

```python
-al [ALIGNMENT], --alignment [ALIGNMENT]
                        Define the alignment of the text in the image. Only
                        used if the width parameter is set. 0: left, 1:
                        center, 2: right
```

--------------------------------

### Generate Chinese Data

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates specified number of Chinese character images using the Chinese dictionary. Use -l cn for Chinese.

```bash
python3 run.py -c 10 -l cn
```

--------------------------------

### Space Width Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Defines the width of the spaces between words. A value of 2.0 means twice the normal space width.

```python
-sw [SPACE_WIDTH], --space_width [SPACE_WIDTH]
                        Define the width of the spaces between words. 2.0
                        means twice the normal space width
```

--------------------------------

### Character Spacing Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Defines the width of the spaces between characters. A value of 2 means two pixels.

```python
-cs [CHARACTER_SPACING], --character_spacing [CHARACTER_SPACING]
                        Define the width of the spaces between characters. 2
                        means two pixels
```

--------------------------------

### Generate Vertical Chinese Text

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates Chinese text in a vertical orientation. Use -or 1 for vertical text.

```bash
python3 run.py -c 10 -l cn -w 5 -sw 0 -or 1
```

--------------------------------

### Text Color Argument

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/overview.html

Defines the text's color. Should be either a single hex color or a range in the ?,? format.

```python
-tc [TEXT_COLOR], --text_color [TEXT_COLOR]
                        Define the text's color, should be either a single hex
                        color or a range in the ?,? format.
```

--------------------------------

### Generate Chinese Data with No Spacing

Source: https://textrecognitiondatagenerator.readthedocs.io/en/latest/tutorial.html

Generates Chinese character images with no spacing between characters. Use -sw 0 to remove spacing.

```bash
python3 run.py -c 10 -l cn -w 5 -sw 0
```
Tokens:
Raw