output-meta

Author: okaris

src/inferencesh/models/__init__.py

@@ -12,6 +12,16 @@
     stream_generate,
     timing_context,
 )
+from .output_meta import (
+    MetaItem,
+    MetaItemType,
+    TextMeta,
+    ImageMeta,
+    VideoMeta,
+    VideoResolution,
+    AudioMeta,
+    OutputMeta,
+)
 
 __all__ = [
     "BaseApp",
@@ -26,4 +36,13 @@
     "build_messages",
     "stream_generate",
     "timing_context",
+    # OutputMeta types
+    "MetaItem",
+    "MetaItemType",
+    "TextMeta",
+    "ImageMeta",
+    "VideoMeta",
+    "VideoResolution",
+    "AudioMeta",
+    "OutputMeta",
 ] 

src/inferencesh/models/base.py

@@ -1,11 +1,11 @@
 from typing import Any, Dict, List, Optional
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel, ConfigDict, Field
 import inspect
 import ast
 import textwrap
 from collections import OrderedDict
 from inferencesh.models.file import File
-from pydantic import Field
+from inferencesh.models.output_meta import OutputMeta
 
 class Metadata(BaseModel):
     app_id: Optional[str] = None
@@ -88,7 +88,10 @@ class BaseAppInput(OrderedSchemaModel):
 
 
 class BaseAppOutput(OrderedSchemaModel):
-    pass
+    output_meta: Optional[OutputMeta] = Field(
+        default=None,
+        description="Structured metadata about inputs/outputs for pricing calculation"
+    )
 
 
 class BaseApp(BaseModel):

src/inferencesh/models/output_meta.py

@@ -0,0 +1,112 @@
+"""Output metadata types for pricing and usage tracking."""
+
+from typing import Any, List, Optional
+from enum import Enum
+from pydantic import BaseModel, Field
+
+
+class MetaItemType(str, Enum):
+    """Type discriminator for MetaItem."""
+    TEXT = "text"
+    IMAGE = "image"
+    VIDEO = "video"
+    AUDIO = "audio"
+
+
+class VideoResolution(str, Enum):
+    """Standard video resolution presets."""
+    RES_480P = "480p"
+    RES_720P = "720p"
+    RES_1080P = "1080p"
+    RES_1440P = "1440p"
+    RES_4K = "4k"
+
+
+class MetaItem(BaseModel):
+    """Base class for input/output metadata items."""
+    type: str  # "text", "image", "video", "audio"
+    extra: Optional[dict[str, Any]] = Field(
+        default=None,
+        description="App-specific key-value pairs for custom pricing factors"
+    )
+
+
+class TextMeta(MetaItem):
+    """Metadata for text inputs/outputs (e.g., LLM tokens)."""
+    type: str = MetaItemType.TEXT.value
+    tokens: int = Field(
+        default=0,
+        description="Token count - in inputs[] = input tokens, in outputs[] = output tokens"
+    )
+
+
+class ImageMeta(MetaItem):
+    """Metadata for image inputs/outputs."""
+    type: str = MetaItemType.IMAGE.value
+    width: int = Field(default=0, description="Image width in pixels")
+    height: int = Field(default=0, description="Image height in pixels")
+    resolution_mp: float = Field(
+        default=0,
+        description="Resolution in megapixels (width * height / 1_000_000)"
+    )
+    steps: int = Field(default=0, description="Number of diffusion steps")
+    count: int = Field(default=1, description="Number of images")
+
+
+class VideoMeta(MetaItem):
+    """Metadata for video inputs/outputs."""
+    type: str = MetaItemType.VIDEO.value
+    width: int = Field(default=0, description="Video width in pixels")
+    height: int = Field(default=0, description="Video height in pixels")
+    resolution_mp: float = Field(
+        default=0,
+        description="Resolution in megapixels per frame"
+    )
+    resolution: Optional[VideoResolution] = Field(
+        default=None,
+        description="Standard resolution preset (480p, 720p, 1080p, 1440p, 4k)"
+    )
+    seconds: float = Field(default=0, description="Duration in seconds")
+    fps: int = Field(default=0, description="Frames per second")
+
+
+class AudioMeta(MetaItem):
+    """Metadata for audio inputs/outputs."""
+    type: str = MetaItemType.AUDIO.value
+    seconds: float = Field(default=0, description="Duration in seconds")
+    sample_rate: int = Field(default=0, description="Sample rate in Hz")
+
+
+class OutputMeta(BaseModel):
+    """
+    Structured metadata about task inputs and outputs for pricing calculation.
+    
+    Apps include this in their output to report what was consumed (inputs)
+    and what was produced (outputs). The backend uses this with CEL expressions
+    to calculate app-level pricing.
+    
+    Example usage in an LLM app:
+        output_meta = OutputMeta(
+            inputs=[TextMeta(tokens=150)],
+            outputs=[TextMeta(tokens=500)]
+        )
+    
+    Example usage in a video generation app:
+        output_meta = OutputMeta(
+            outputs=[VideoMeta(
+                resolution=VideoResolution.RES_1080P,
+                resolution_mp=2.07,
+                seconds=10.5,
+                fps=30
+            )]
+        )
+    """
+    inputs: List[MetaItem] = Field(
+        default_factory=list,
+        description="Metadata about consumed inputs"
+    )
+    outputs: List[MetaItem] = Field(
+        default_factory=list,
+        description="Metadata about produced outputs"
+    )
+