lxml 5.2 removed `lxml.html.clean`, and requests-html still reaches for it, so the import raises this explicit error. Install the split-out package with `python -m pip install lxml_html_clean` and the import works. This is the officially blessed migration path, not a hack.

## The error

```text
ImportError: lxml.html.clean module is now a separate project lxml_html_clean
```

## Fix it

1. **Install the split-out package**

```
python -m pip install lxml_html_clean
```
Expected: `Successfully installed lxml_html_clean-...`.

2. **Confirm**

```
python -c "from requests_html import HTMLSession; print('ok')"
```
Expected: prints `ok`.

3. **Alternative: downgrade lxml below the split**

```
python -m pip install 'lxml<5.2'
```
Expected: the old `lxml.html.clean` path exists again. Prefer step 1; old lxml forfeits security fixes.

## When this applies

- the ImportError text names the lxml_html_clean split
- importing requests-html, facebook-scraper, or similar fails right after an lxml upgrade
- you are on lxml 5.2+

## When it does NOT apply

- `ModuleNotFoundError: No module named 'lxml_html_clean'` is the same split seen as a missing module (same install fixes it)
- `No module named 'lxml'` means lxml itself is absent

## Compatibility

lxml 5.2.0+. The shim package lxml_html_clean tracks lxml 5.x. requests-html 0.10.0 is the version most people hit this on.

## Why it happens

The lxml maintainers extracted the HTML-cleaning code into its own distribution because it was a repeated CVE source. lxml 5.2 raises this descriptive ImportError (instead of a bare ModuleNotFoundError) from a stub, precisely to tell you the new package name. Installing it restores the exact module the old code imports.

## Edge cases

- If pip installs the shim but the error persists, you have two lxml installs (e.g. system and venv). The `python -m pip` spelling targets the right one.
- facebook-scraper and pymoviedb hit the identical error through the same lxml upgrade; the fix is the same single install.