84 lines
3.3 KiB
Markdown
84 lines
3.3 KiB
Markdown
|
|
# Documentation
|
|
- Class name: LLMHtmlComposer
|
|
- Category: SALT/Language Toolkit/Tools/HTML
|
|
- Output node: False
|
|
|
|
LLMHtmlComposer节点旨在根据给定的文本输入、分类器列表和可选的额外指令生成HTML内容。它利用语言模型来组合完整的HTML页面文档或HTML片段，具体取决于指定的模式，确保输出结构良好且有效。
|
|
|
|
# Input types
|
|
## Required
|
|
- llm_model
|
|
- 表示用于生成HTML内容的语言模型。它对于解释输入文本和分类器以产生所需的HTML输出至关重要。
|
|
- Comfy dtype: LLM_MODEL
|
|
- Python dtype: dict
|
|
- text_input
|
|
- 用于生成HTML内容的主要文本输入。这段文本作为composition过程的基础。
|
|
- Comfy dtype: STRING
|
|
- Python dtype: str
|
|
- classifier_list
|
|
- 以逗号分隔的字符串形式提供的分类器列表。这些分类器指导语言模型生成符合特定标准或类别的HTML内容。
|
|
- Comfy dtype: STRING
|
|
- Python dtype: str
|
|
## Optional
|
|
- extra_directions
|
|
- 可选的附加指令，用于进一步指导HTML生成过程，允许生成更加定制化的输出。
|
|
- Comfy dtype: STRING
|
|
- Python dtype: str
|
|
- composer_mode
|
|
- 指定HTML composition的模式，可以是完整的HTML页面文档或HTML片段，影响输出中HTML元素的结构和包含。
|
|
- Comfy dtype: COMBO[STRING]
|
|
- Python dtype: str
|
|
|
|
# Output types
|
|
- html_output
|
|
- 生成的HTML内容，可以是完整的HTML页面或HTML片段，基于输入文本、分类器和可选指令。
|
|
- Comfy dtype: STRING
|
|
- Python dtype: str
|
|
|
|
|
|
## Usage tips
|
|
- Infra type: `GPU`
|
|
- Common nodes: unknown
|
|
|
|
|
|
## Source code
|
|
```python
|
|
class LLMHtmlComposer:
|
|
@classmethod
|
|
def INPUT_TYPES(cls):
|
|
return {
|
|
"required": {
|
|
"llm_model": ("LLM_MODEL",),
|
|
"text_input": ("STRING", {"multiline": True, "dynamicPrompts": False, "placeholder": "Data..."}),
|
|
"classifier_list": ("STRING", {"multiline": False, "dynamicPrompts": False}),
|
|
},
|
|
"optional": {
|
|
"extra_directions": ("STRING", {"multiline": True, "dynamicPrompts": False, "placeholder": "Extra directions for the LLM to follow..."}),
|
|
"composer_mode": (["full_markup", "blocked_markup"],)
|
|
}
|
|
}
|
|
|
|
RETURN_TYPES = ("STRING",)
|
|
RETURN_NAMES = ("html_output",)
|
|
|
|
FUNCTION = "compose_html"
|
|
CATEGORY = f"{MENU_NAME}/{SUB_MENU_NAME}/Tools/HTML"
|
|
|
|
def compose_html(self, llm_model, text_input, classifier_list, extra_directions="", composer_mode="full_markup"):
|
|
classifier_list = [item.strip() for item in classifier_list.split(",") if item.strip()]
|
|
markup_style = "full HTML page document" if composer_mode == "full_markup" else "HTML snippet (without html, head, body or any extraneous containers)"
|
|
prompt = (
|
|
f"{text_input}\n\n###\n\n"
|
|
f"Given the above text, create a valid {markup_style} utilizing *all* of the data, intact; "
|
|
f"using the following classifiers: {classifier_list}.\n\n"
|
|
f"{extra_directions}\n\n"
|
|
"Please ensure the HTML output is well-structured, valid, and does not omit any data."
|
|
)
|
|
|
|
response = llm_model['llm'].complete(prompt)
|
|
|
|
return (response.text,)
|
|
|
|
```
|