Add custom nodes, Civitai loras (LFS), and vast.ai setup script

Includes 30 custom nodes committed directly, 7 Civitai-exclusive
loras stored via Git LFS, and a setup script that installs all
dependencies and downloads HuggingFace-hosted models on vast.ai.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

custom_nodes/rgthree-comfy/py/power_prompt.py

@@ -0,0 +1,95 @@
+import os
+
+from .log import log_node_warn, log_node_info, log_node_success
+
+from .constants import get_category, get_name
+from .power_prompt_utils import get_and_strip_loras
+from nodes import LoraLoader, CLIPTextEncode
+import folder_paths
+
+NODE_NAME = get_name('Power Prompt')
+
+
+class RgthreePowerPrompt:
+
+  NAME = NODE_NAME
+  CATEGORY = get_category()
+
+  @classmethod
+  def INPUT_TYPES(cls):  # pylint: disable = invalid-name, missing-function-docstring
+    # Removed Saved Prompts feature; No sure it worked any longer. UI should fail gracefully,
+    # TODO: Rip out saved prompt input data
+    SAVED_PROMPTS_FILES=[]
+    SAVED_PROMPTS_CONTENT=[]
+    return {
+      'required': {
+        'prompt': ('STRING', {
+          'multiline': True,
+          'dynamicPrompts': True
+        }),
+      },
+      'optional': {
+        "opt_model": ("MODEL",),
+        "opt_clip": ("CLIP",),
+        'insert_lora': (['CHOOSE', 'DISABLE LORAS'] +
+                        [os.path.splitext(x)[0] for x in folder_paths.get_filename_list('loras')],),
+        'insert_embedding': ([
+          'CHOOSE',
+        ] + [os.path.splitext(x)[0] for x in folder_paths.get_filename_list('embeddings')],),
+        'insert_saved': ([
+          'CHOOSE',
+        ] + SAVED_PROMPTS_FILES,),
+      },
+      'hidden': {
+        'values_insert_saved': (['CHOOSE'] + SAVED_PROMPTS_CONTENT,),
+      }
+    }
+
+  RETURN_TYPES = (
+    'CONDITIONING',
+    'MODEL',
+    'CLIP',
+    'STRING',
+  )
+  RETURN_NAMES = (
+    'CONDITIONING',
+    'MODEL',
+    'CLIP',
+    'TEXT',
+  )
+  FUNCTION = 'main'
+
+  def main(self,
+           prompt,
+           opt_model=None,
+           opt_clip=None,
+           insert_lora=None,
+           insert_embedding=None,
+           insert_saved=None,
+           values_insert_saved=None):
+    if insert_lora == 'DISABLE LORAS':
+      prompt, loras, skipped, unfound = get_and_strip_loras(prompt, log_node=NODE_NAME, silent=True)
+      log_node_info(
+        NODE_NAME,
+        f'Disabling all found loras ({len(loras)}) and stripping lora tags for TEXT output.')
+    elif opt_model is not None and opt_clip is not None:
+      prompt, loras, skipped, unfound = get_and_strip_loras(prompt, log_node=NODE_NAME)
+      if len(loras) > 0:
+        for lora in loras:
+          opt_model, opt_clip = LoraLoader().load_lora(opt_model, opt_clip, lora['lora'],
+                                                       lora['strength'], lora['strength'])
+          log_node_success(NODE_NAME, f'Loaded "{lora["lora"]}" from prompt')
+        log_node_info(NODE_NAME, f'{len(loras)} Loras processed; stripping tags for TEXT output.')
+    elif '<lora:' in prompt:
+      prompt, loras, skipped, unfound = get_and_strip_loras(prompt, log_node=NODE_NAME, silent=True)
+      total_loras = len(loras) + len(skipped) + len(unfound)
+      if total_loras:
+        log_node_warn(
+          NODE_NAME, f'Found {len(loras)} lora tags in prompt but model & clip were not supplied!')
+        log_node_info(NODE_NAME, 'Loras not processed, keeping for TEXT output.')
+
+    conditioning = None
+    if opt_clip is not None:
+      conditioning = CLIPTextEncode().encode(opt_clip, prompt)[0]
+
+    return (conditioning, opt_model, opt_clip, prompt)