Add text_rich()

Author: deedy5

README.md

@@ -9,7 +9,8 @@ Python binding to the rust [rust-html2text](https://github.com/jugglerchris/rust
 - [Installation](#installation)
 - [Usage](#usage)
   - [text_markdown()](#1-text_markdown)
-  - [text_plain()](#2-text_plain) 
+  - [text_plain()](#2-text_plain)
+  - [text_rich()](#3-text_rich)
 
 ## Installation
 
@@ -29,12 +30,12 @@ def text_markdown(html: str, width: int = 100):
 
     """
 ```
-*example*
+example:
 ```python
 import html2text_rs
 import requests
 
-resp = requests.get("https://tmz.com")
+resp = requests.get("https://en.wikipedia.org/wiki/AGM-88_HARM")
 
 text_markdown = html2text_rs.text_markdown(resp.text)
 print(text_markdown)
@@ -50,13 +51,34 @@ def text_plain(html: str, width: int = 100):
 
     """
 ```
-*example*
+example:
 ```python
 import html2text_rs
 import requests
 
-resp = requests.get("https://tmz.com")
+resp = requests.get("https://en.wikipedia.org/wiki/AGM-88_HARM")
 
 text_plain = html2text_rs.text_plain(resp.text)
 print(text_plain)
 ```
+### 3. text_rich()
+```python
+def text_rich(html: str, width: int = 100):
+    """Convert HTML to rich text.
+
+    Args:
+        html (str): input html text.
+        width (int): wrap text to width columns. Default is 100.
+
+    """
+```
+example:
+```python
+import html2text_rs
+import requests
+
+resp = requests.get("https://en.wikipedia.org/wiki/AGM-88_HARM")
+
+text_rich = html2text_rs.text_rich(resp.text)
+print(text_rich)
+```

src/lib.rs

@@ -1,4 +1,7 @@
-use html2text::{from_read, from_read_with_decorator, render::text_renderer::TrivialDecorator};
+use html2text::{
+    from_read, from_read_with_decorator,
+    render::text_renderer::{RichDecorator, TrivialDecorator},
+};
 use pyo3::prelude::*;
 
 /// Convert HTML to markdown text
@@ -19,9 +22,19 @@ fn text_plain(html: String, width: usize, py: Python) -> PyResult<String> {
     Ok(text)
 }
 
+/// Convert HTML to rich text
+#[pyfunction]
+#[pyo3(signature=(html, width=100))]
+fn text_rich(html: String, width: usize, py: Python) -> PyResult<String> {
+    let text =
+        py.allow_threads(|| from_read_with_decorator(html.as_bytes(), width, RichDecorator::new()));
+    Ok(text)
+}
+
 #[pymodule]
 fn html2text_rs(_py: Python, m: &Bound<'_, PyModule>) -> PyResult<()> {
-    m.add_function(wrap_pyfunction!(text_plain, m)?)?;
     m.add_function(wrap_pyfunction!(text_markdown, m)?)?;
+    m.add_function(wrap_pyfunction!(text_plain, m)?)?;
+    m.add_function(wrap_pyfunction!(text_rich, m)?)?;
     Ok(())
 }

tests/test_html2text_rs.py

@@ -1,6 +1,4 @@
-import pytest
-
-import html2text_rs
+import html2text_rs  # type: ignore
 
 
 def test_text_markdown():
@@ -19,3 +17,12 @@ def test_text_plain():
     assert (
         result == expected_output
     ), f"\nExpected:\n {expected_output} \nGot:\n {result}"
+
+
+def test_text_rich():
+    html = "<h1>Hello World</h1><p>This is a test.</p>"
+    expected_output = "# Hello World\n\nThis is a test.\n"
+    result = html2text_rs.text_rich(html, width=80)
+    assert (
+        result == expected_output
+    ), f"\nExpected:\n {expected_output} \nGot:\n {result}"