Format code with current Black

Author: luochen211

deepeval/cli/inspect.py

@@ -14,7 +14,6 @@
 import typer
 from rich import print
 
-
 _INSTALL_HINT = (
     "[bold red]deepeval inspect[/bold red] requires extras that are not "
     "installed.\n"

deepeval/inspect/widgets/_styling.py

@@ -12,7 +12,6 @@
 
 from deepeval.inspect.types import Trace, TraceOrSpan
 
-
 # `(glyph, tag, rich style)` per span type. Tags are full words rather
 # than abbreviations because the tree pane is wide enough to spell them
 # out, and "RETRIEVER" reads instantly while "RET" trips users into

deepeval/inspect/widgets/details.py

@@ -30,7 +30,6 @@
     type_prefix,
 )
 
-
 # Matches the TRACE tag so the eye learns "cyan = structure markers".
 _HEADER_ACCENT = "#8be9fd"
 _CTA_ACCENT = "#bd93f9"

deepeval/inspect/widgets/help_modal.py

@@ -9,7 +9,6 @@
 from textual.screen import ModalScreen
 from textual.widgets import Static
 
-
 _HELP_ROWS = [
     ("↑ ↓ / k j", "move selection in the tree"),
     ("h / l", "go to parent / select child in the tree"),

deepeval/inspect/widgets/span_tree.py

@@ -25,7 +25,6 @@
     type_prefix,
 )
 
-
 # Minimum gap (in cells) between the left content (name + metric badge +
 # optional ERRORED pill) and the right-aligned duration. Below this the
 # right column gives up trying to right-align and just leaves the

deepeval/metrics/contextual_relevancy/contextual_relevancy.py

@@ -30,12 +30,10 @@ def _contextual_relevancy_verdict_kwargs(multimodal: bool) -> Dict[str, str]:
     context_type = "context (image or string)" if multimodal else "context"
     statement_or_image = "statement or image" if multimodal else "statement"
     if multimodal:
-        extraction_instructions = textwrap.dedent(
-            """
+        extraction_instructions = textwrap.dedent("""
             If the context is textual, you should first extract the statements found in the context if the context, which are high level information found in the context, before deciding on a verdict and optionally a reason for each statement.
             If the context is an image, `statement` should be a description of the image. Do not assume any information not visibly available.
-            """
-        ).strip()
+            """).strip()
         empty_context_instruction = ""
     else:
         extraction_instructions = (

deepeval/metrics/dag/serialization/registry.py

@@ -15,7 +15,6 @@
 
 from .types import NodeType
 
-
 NODE_CLASSES: Dict[bool, Dict[NodeType, Type]] = {
     False: {
         NodeType.TASK: TaskNode,

deepeval/metrics/dag/serialization/serialization.py

@@ -61,7 +61,6 @@
 from .registry import CLASS_TO_NODE_TYPE, NODE_CLASSES
 from .types import ChildType, NodeType
 
-
 # ----------------------------------------------------------------------------
 # Public API
 # ----------------------------------------------------------------------------

deepeval/metrics/multimodal_metrics/image_editing/image_editing.py

@@ -287,15 +287,13 @@ def is_successful(self) -> bool:
     def _generate_reason(
         self,
     ) -> str:
-        return textwrap.dedent(
-            f"""
+        return textwrap.dedent(f"""
             The overall score is {self.score:.2f} because the lowest score from semantic consistency was {min(self.SC_scores)} 
             and the lowest score from perceptual quality was {min(self.PQ_scores)}. These scores were combined to reflect the 
             overall effectiveness and quality of the AI-generated image(s).
             Reason for Semantic Consistency score: {self.SC_reasoning}
             Reason for Perceptual Quality score: {self.PQ_reasoning}
-        """
-        )
+        """)
 
     @property
     def __name__(self):

deepeval/metrics/multimodal_metrics/text_to_image/text_to_image.py

@@ -289,15 +289,13 @@ def is_successful(self) -> bool:
         return self.success
 
     def _generate_reason(self) -> str:
-        return textwrap.dedent(
-            f"""
+        return textwrap.dedent(f"""
             The overall score is {self.score:.2f} because the lowest score from semantic consistency was {min(self.SC_scores)} 
             and the lowest score from perceptual quality was {min(self.PQ_scores)}. These scores were combined to reflect the 
             overall effectiveness and quality of the AI-generated image(s).
             Reason for Semantic Consistency score: {self.SC_reasoning}
             Reason for Perceptual Quality score: {self.PQ_reasoning}
-        """
-        )
+        """)
 
     @property
     def __name__(self):