frames: added vision response and text frames

changelog/3252.added.md

@@ -0,0 +1,2 @@
+- Added `VisionFullResponseStartFrame`, `VisionFullResponseEndFrame` and
+  `VisionTextFrame`. This are used by vision services similar to LLM services.

src/pipecat/frames/frames.py

@@ -400,6 +400,13 @@ class AggregatedTextFrame(TextFrame):
     aggregated_by: AggregationType | str
 
 
+@dataclass
+class VisionTextFrame(LLMTextFrame):
+    """Text frame generated by vision services."""
+
+    pass
+
+
 @dataclass
 class TTSTextFrame(AggregatedTextFrame):
     """Text frame generated by Text-to-Speech services."""
@@ -1766,6 +1773,25 @@ class FunctionCallInProgressFrame(ControlFrame, UninterruptibleFrame):
     cancel_on_interruption: bool = False
 
 
+@dataclass
+class VisionFullResponseStartFrame(LLMFullResponseStartFrame):
+    """Frame indicating the beginning of a vision model response.
+
+    Used to indicate the beginning of a vision model response. Followed by one
+    or more VisionTextFrames and a final VisionFullResponseEndFrame.
+
+    """
+
+    pass
+
+
+@dataclass
+class VisionFullResponseEndFrame(LLMFullResponseEndFrame):
+    """Frame indicating the end of a Vision model response."""
+
+    pass
+
+
 @dataclass
 class TTSStartedFrame(ControlFrame):
     """Frame indicating the beginning of a TTS response.