| polar | Generate a word cloud from two text sources. The two word clouds will be drawn on the same canvas and will bleed into each other. This is ideal for contrasting two pieces of data such as positive and negative sentiment. |
| layered | Generate multiple word clouds and layer them over one another. This is done by providing layered imaged. The word clouds will be plotted on the non-transparent pixels of the layered images. |
Number of words from data set to draw to word cloud. After the words are sorted by frequency, the words are attempted to be placed in descending order.
| pixel_perfect | (default) When placing text into the word cloud check pixel-by-pixel to determine if text overlaps. |
| rectangle | Perform simple rectangular collision detection when placing text. This is results in faster generation of word clouds but they may not be aesthetically pleasing. |
One ore more input sources. Input sources may be local files or Urls of an image used to define the shape of the word cloud. By default the word cloud is drawn onto a rectangle. The word cloud will place text only in places where background image has non-transparent pixels.
For standard word clouds only the first input source will be used. Multiple input sources are only relevant for layered word clouds. Each background image will be applied to a layer in the order they are listed.
One or more normalizers to apply to words in the word cloud.
| Value | Description |
|-------|-------------|
| uppercase | Convert each character to uppercase. |
| lowercase | Convert each character to lowercase. |
| trim | Trim spaces from words. (Typically a good tokenizer resolves this. |
| upsidedown | Draws the fonts upside down. Only works with alphanumeric characters. e.g. Y -> ʎ |
| bubble | replace alphabet characters with bubble representations. e.g. a -> ⓐ |
| character-stripping | By default this normalizer will remove common punctuation characters. It is programmatically configurable and will eventually be supported in the CLI as well. |
