fix: pad SDXL compel conditionings for long prompts

compel([prompt, negative]) pads internally via
conditioning_provider.empty_z, which the SDXL dual-encoder
EmbeddingsProviderMulti does not expose, so long prompts (whose chunk
count differs from the negative) crashed with AttributeError. Encode the
prompt and negative separately and pad with the empty-string conditioning
passed as precomputed_padding.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

app.py

@@ -342,13 +342,24 @@ def generate_image(
             negative_prompt = ""
         conv_prompt = get_embed_new(prompt, noobxl_pipe, compel, only_convert_string=True)
         conv_negative = get_embed_new(negative_prompt, noobxl_pipe, compel, only_convert_string=True)
-        conditioning, pooled = compel([conv_prompt, conv_negative])
+        # Encode prompt and negative separately, then pad to equal length. The
+        # batched ``compel([prompt, negative])`` call pads internally via
+        # ``conditioning_provider.empty_z``, which the SDXL dual-encoder
+        # ``EmbeddingsProviderMulti`` does not expose (crashes on long prompts
+        # whose chunk count differs from the negative). Supplying the empty-string
+        # conditioning as ``precomputed_padding`` sidesteps that missing attribute.
+        cond_prompt, pooled_prompt = compel(conv_prompt)
+        cond_negative, pooled_negative = compel(conv_negative)
+        empty_padding, _ = compel("")
+        cond_prompt, cond_negative = compel.pad_conditioning_tensors_to_same_length(
+            [cond_prompt, cond_negative], precomputed_padding=empty_padding
+        )
 
         image = noobxl_pipe(
-            prompt_embeds=conditioning[0:1],
-            pooled_prompt_embeds=pooled[0:1],
-            negative_prompt_embeds=conditioning[1:2],
-            negative_pooled_prompt_embeds=pooled[1:2],
+            prompt_embeds=cond_prompt,
+            pooled_prompt_embeds=pooled_prompt,
+            negative_prompt_embeds=cond_negative,
+            negative_pooled_prompt_embeds=pooled_negative,
             width=int(width),
             height=int(height),
             guidance_scale=float(guidance_scale),