Lorem-ipsum dolor-sit amet.
`; const docDefinition = { content: htmlToPdfmake(html, { replaceText: (text, nodes) => text.replace(/-/g, '\u2011') }) }; ``` ### Context-Aware Replacement Use the `parents` array to apply transformations only within specific HTML elements. This example masks digits only when they appear inside `` blocks:
```javascript
const html = `
Version: 1.2.3
12345
`;
const docDefinition = {
content: htmlToPdfmake(html, {
replaceText: (text, nodes) => {
const insideCode = nodes.some(node => node.nodeName === 'CODE');
return insideCode ? text.replace(/\d/g, '#') : text;
}
})
};
```
### Advanced Text Transformation
While `replaceText` modifies the string content, you can use it to inject markdown-style markers that PDFMake can interpret, or prepare text for post-processing. This example wraps the word "important" with markers for later styling:
```javascript
const html = `This is important information.
`;
const docDefinition = {
content: htmlToPdfmake(html, {
replaceText: (text) => text.replace(/important/g, '**important**')
})
};
```
## Technical Implementation Details
### The Core Logic in index.js
The `replaceText` integration occurs within the `parseElement` function in [`index.js`](https://github.com/aymkdn/html-to-pdfmake/blob/main/index.js). The library traverses the DOM recursively, and when it identifies a node of type `3` (text node), it immediately checks for the replacement function before applying any default styling or wrapping.
This placement ensures that text modifications happen **before** PDFMake object creation, allowing your custom logic to affect the final output without interfering with style inheritance or layout calculations.
### Return Value Requirements
The `replaceText` function must always return a string value. Returning `undefined`, `null`, or an object will cause the conversion to fail or produce invalid PDFMake definitions. If no modification is required for a particular node, simply return the original `text` parameter unchanged.
The function executes synchronously; asynchronous text processing is not supported within this hook. For complex async transformations, preprocess your HTML before passing it to `html-to-pdfmake`.
## Summary
- The `replaceText` option in `html-to-pdfmake` provides a synchronous hook to modify text content during conversion at [`index.js`](https://github.com/aymkdn/html-to-pdfmake/blob/main/index.js) lines 166–168.
- The function receives the raw `text` string and a `parents` array containing all ancestor DOM elements for context-aware processing.
- Return a modified string to change the output, or return the original string to leave content unchanged; never return `undefined` or non-string values.
- Use this hook for character substitution, sensitive data redaction, or conditional formatting based on parent HTML elements.
## Frequently Asked Questions
### Can I use replaceText to remove text content entirely?
Yes. To remove text content, return an empty string `""` from your `replaceText` function. This will effectively delete that text node from the final PDF output. Be cautious when removing text inside inline elements, as this may affect layout or leave empty parent containers.
### Does replaceText affect text inside tables, lists, and other complex structures?
Yes. The `replaceText` hook processes **every** text node in the HTML document regardless of its parent structure. Text inside ``, ``, ``, or any other element passes through this function. Use the `parents` parameter to detect specific container types if you need to apply different logic for text in tables versus paragraphs.
### Can I change text styling using the replaceText option?
No. The `replaceText` function can only modify the **string content** of text nodes. It cannot directly alter PDFMake style properties like `bold`, `italics`, or `fontSize`. To apply conditional styling, you would need to either use the `customTag` option to handle specific elements, or post-process the resulting PDFMake document definition after conversion.
### What happens if my replaceText function returns undefined or null?
Returning `undefined`, `null`, or any non-string value from `replaceText` will cause the conversion to fail or produce invalid PDFMake document definitions. The library expects a string return value to populate the text content of the PDF node. Always ensure your function returns a string, even if it's just the original `text` parameter unchanged.