⚡️ Speed up method MetadataToStringInvocation.invoke by 25%

[codeflash-ai]

📄 25% (0.25x) speedup for MetadataToStringInvocation.invoke in invokeai/app/invocations/metadata_linked.py

⏱️ Runtime : 10.9 microseconds → 8.66 microseconds (best of 52 runs)

📝 Explanation and details

The optimization achieves a 25% speedup by eliminating unnecessary dictionary allocation when metadata is None, which appears to be a common case based on the test results.

Key optimizations:

1. Early return for None metadata: Instead of creating an empty dictionary {} when self.metadata is None, the code now returns immediately with the default value. This avoids the overhead of dictionary creation and the subsequent .get() call.

2. Reduced conditional evaluation: The key construction logic is moved earlier and stored in a variable, avoiding redundant string conversion and conditional checks in the dictionary lookup path.

Performance impact:
The line profiler shows that in the None metadata case, the optimized version executes fewer operations - it bypasses the dictionary allocation entirely and goes straight to returning the StringOutput. The test results demonstrate consistent improvements:

• test_edge_none_metadata_returns_default: 30.1% faster (5.25μs → 4.04μs)
• test_edge_metadata_is_none: 21.4% faster (5.62μs → 4.63μs)

Why this works:
In Python, dictionary allocation has overhead even for empty dictionaries. When metadata is None (which appears to be a frequent case in metadata processing workflows), the original code unnecessarily allocated an empty dict just to call .get() on it with a default value. The optimization recognizes that {}.get(key, default) always returns default, so it shortcuts directly to that result.

This optimization is particularly valuable for metadata processing pipelines where None metadata may be common, and the function could be called repeatedly during image generation workflows.

✅ Correctness verification report:

Test	Status
⚙️ Existing Unit Tests	🔘 None Found
🌀 Generated Regression Tests	✅ 6 Passed
⏪ Replay Tests	🔘 None Found
🔎 Concolic Coverage Tests	🔘 None Found
📊 Tests Coverage	100.0%

🌀 Generated Regression Tests and Runtime

from typing import Any, Dict, Literal, Optional # imports import pytest # used for our unit tests from invokeai.app.invocations.metadata_linked import MetadataToStringInvocation # function to test CUSTOM_LABEL: str = "* CUSTOM LABEL *" # Minimal stub for StringOutput class StringOutput: def __init__(self, value: str): self.value = value # Minimal stub for InvocationContext class InvocationContext: pass # Minimal stub for WithMetadata class MetadataStub: def __init__(self, root: Dict[str, Any]): self.root = root class WithMetadata: def __init__(self, metadata=None): self.metadata = metadata # Function to validate custom label (from Pydantic model_validator) def validate_custom_label(self): # If label is CUSTOM_LABEL, custom_label must be set and non-empty if getattr(self, "label", None) == CUSTOM_LABEL: if not getattr(self, "custom_label", None): raise ValueError("custom_label must be set when label is CUSTOM_LABEL") return self from invokeai.app.invocations.metadata_linked import MetadataToStringInvocation # ----------------- UNIT TESTS ----------------- # Basic Test Cases def test_edge_none_metadata_returns_default(): """Test with None metadata, should return default value.""" inv = MetadataToStringInvocation(label="positive_prompt", default_value="Default", metadata=None) codeflash_output = inv.invoke(InvocationContext()); result = codeflash_output # 5.25μs -> 4.04μs (30.1% faster) def test_edge_custom_label_none_key(): """Test with custom label as None (should raise ValueError).""" metadata = MetadataStub({None: "None Key Value"}) with pytest.raises(ValueError): MetadataToStringInvocation(label=CUSTOM_LABEL, custom_label=None, default_value="Default", metadata=metadata) def test_edge_custom_label_missing_and_label_is_custom(): """Test with label CUSTOM_LABEL but missing custom_label (should raise ValueError).""" metadata = MetadataStub({"foo": "bar"}) with pytest.raises(ValueError): MetadataToStringInvocation(label=CUSTOM_LABEL, custom_label=None, default_value="Default", metadata=metadata) #------------------------------------------------ from typing import Any, Dict, Optional # imports import pytest # used for our unit tests from invokeai.app.invocations.metadata_linked import MetadataToStringInvocation # --- Minimal stubs to allow testing without full app context --- # Simulate the output class class StringOutput: def __init__(self, value: str): self.value = value # Simulate the InvocationContext and Metadata class FakeMetadata: def __init__(self, root: Dict[str, Any]): self.root = root class InvocationContext: pass # Not used in logic, just a placeholder # Constants from the original code CUSTOM_LABEL: str = "* CUSTOM LABEL *" CORE_LABELS = [ CUSTOM_LABEL, "positive_prompt", "positive_style_prompt", "negative_prompt", "negative_style_prompt", ] from invokeai.app.invocations.metadata_linked import MetadataToStringInvocation # --- Unit tests --- # Basic Test Cases def test_edge_metadata_is_none(): """Test behavior when metadata is None.""" invocation = MetadataToStringInvocation( label="positive_prompt", default_value="default value", metadata=None ) codeflash_output = invocation.invoke(InvocationContext()); result = codeflash_output # 5.62μs -> 4.63μs (21.4% faster)

To edit these changes git checkout codeflash/optimize-MetadataToStringInvocation.invoke-mhlk3gkk and push.