Add lora with metadata and health endpoints & IPA fix instead of explicit flag

__init__.py

@@ -1,4 +1,4 @@
 from .nodes import NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS
+from .server import *
 
-
-__all__ = ['NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS']
+__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS"]

attention_couple.py

@@ -152,10 +152,6 @@ class AttentionCouple:
                 ),
                 "width": ("INT", {"default": 1024, "min": 8, "step": 8}),
                 "height": ("INT", {"default": 1024, "min": 8, "step": 8}),
-                "ip_adapter_active": (
-                    "BOOLEAN",
-                    {"default": False, "tooltip": "Set to true if using IPA."},
-                ),
             },
         }
 
@@ -173,27 +169,21 @@ class AttentionCouple:
         height,
         width,
         regions,
-        ip_adapter_active,
         **kwargs,
     ):
         base_mask = torch.zeros((height, width)).unsqueeze(0)
         global_mask = (torch.ones((height, width)) * global_prompt_weight).unsqueeze(0)
-        ip_even_mask = (torch.ones((height, width)) * 0.1).unsqueeze(0)
 
         new_model = model.clone()
 
         if not isinstance(regions, list):
             regions = [regions]
 
-        num_conds = (
-            len(regions) + 1 + (1 if ip_adapter_active and len(regions) % 2 == 0 else 0)
-        )
+        num_conds = len(regions) + 1
 
         mask = [base_mask] + [
             global_mask
             if i == 0
-            else ip_even_mask
-            if ip_adapter_active and len(regions) % 2 == 0 and i == num_conds - 1
             else F.interpolate(
                 regions[i - 1]["mask"].unsqueeze(0),
                 size=(height, width),
@@ -207,10 +197,7 @@ class AttentionCouple:
         self.mask = mask / mask.sum(dim=0, keepdim=True)
 
         self.conds = [
-            base_prompt[0][0]
-            if i == 0
-            or (ip_adapter_active and len(regions) % 2 == 0 and i == num_conds - 1)
-            else regions[i - 1]["cond"][0][0]
+            base_prompt[0][0] if i == 0 else regions[i - 1]["cond"][0][0]
             for i in range(0, num_conds)
         ]
         num_tokens = [cond.shape[1] for cond in self.conds]
@@ -253,6 +240,13 @@ class AttentionCouple:
             qs = torch.cat(qs, dim=0)
             ks = torch.cat(ks, dim=0).to(k)
 
+            if qs.size(0) % 2 == 1:
+                empty = torch.zeros_like(qs[0]).unsqueeze(0)
+                qs = torch.cat((qs, empty), dim=0)
+
+                empty2 = torch.zeros_like(ks[0]).unsqueeze(0)
+                ks = torch.cat((ks, empty2), dim=0)
+
             return qs, ks, ks
 
         def attn2_output_patch(out, extra_options):

server/__init__.py

@@ -0,0 +1 @@
+from .api import *

server/api.py

@@ -0,0 +1,71 @@
+from server import PromptServer
+from aiohttp import web
+import folder_paths
+import os
+import logging
+
+import struct
+import json
+import pathlib
+
+
+@PromptServer.instance.routes.get("/a8r8/loras")
+async def loras(request):
+    try:
+        loras = [
+            {
+                "path": lora,
+                "name": pathlib.Path(lora).stem,
+                "metadata": get_lora_metadata(lora),
+            }
+            for lora in folder_paths.get_filename_list("loras")
+        ]
+
+        return web.json_response(loras)
+    except Exception as e:
+        logging.error(e)
+        return web.Response(status=400, text=str(e))
+
+
+@PromptServer.instance.routes.get("/a8r8/health")
+async def health(_request):
+    try:
+        return web.json_response(
+            {},
+            status=200,
+        )
+    except Exception as e:
+        logging.error(e)
+        return web.Response(status=400, text=str(e))
+
+
+def get_lora_metadata(lora):
+    try:
+        if ".safetensors" in lora:
+            base_path = folder_paths.folder_names_and_paths["loras"][0][0]
+
+            lora_path = os.path.join(base_path, lora)
+
+            return read_metadata(lora_path)
+        return None
+    except FileNotFoundError as _e:
+        return None
+
+
+def read_metadata(file_path):
+    # Open the file in binary mode
+    with open(file_path, "rb") as io_device:
+        # Read 8 bytes and unpack as little-endian unsigned integer
+        n_bytes = io_device.read(8)
+
+        n = struct.unpack("<Q", n_bytes)[
+            0
+        ]  # '<Q' means little-endian unsigned long long
+
+        # Read n bytes and decode JSON
+        metadata_bytes = io_device.read(n)
+        io_device.close()
+        metadata = json.loads(metadata_bytes.decode("utf-8"))
+
+        # Retrieve the value associated with "__metadata__"
+        return metadata.get("__metadata__")