Request for ClipTextEncode HiDream node

### Feature Idea

Similar to the node you have for SD3, Flux, and others:

Please add a ClipTextEncode node specific for HiDream, so we can separately prompt all 4 encoders.

If possible, please also allow weighting the encoders (as testing with the HiDream wrappers exposed that Clip-L could be underweighted and Llama could be overweighted for better results).

Also if possible, please allow full token length of each encoder (currently anything over 77 throws a warning as that is the lowest common denominator). So in theory we could prompt the llama text encoder at (I think) 2048 tokens if we wanted to get wild.

### Existing Solutions

For reference, this is from the excellent wrapper by SanDiegoDude https://github.com/SanDiegoDude/ComfyUI-HiDream-Sampler . This allows prompting per encoder, and setting the weighting of each encoder, and allows setting max token length for each:


### Other

_No response_