refactor: update examples to use .run() method consistently

Author: Pierre

examples/03_caching.py

@@ -17,7 +17,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 # Import CacheUsage type
 CacheUsage = Literal["auto", "always", "never"]
@@ -54,7 +54,7 @@ class SOAPNote(BaseModel):
     id="soap-extractor",
     model=Model.LLAMA_3_3_70B,
 )
-async def extract_soap_notes(soap_input: SOAPInput) -> Run[SOAPNote]:
+async def extract_soap_notes(soap_input: SOAPInput) -> SOAPNote:
     """
     Extract SOAP notes from a medical consultation transcript.
 
@@ -91,7 +91,7 @@ async def demonstrate_caching(transcript: str):
     for cache_option in cache_options:
         start_time = time.time()
 
-        run = await extract_soap_notes(
+        run = await extract_soap_notes.run(
             SOAPInput(transcript=transcript),
             use_cache=cache_option,
         )

examples/04_audio_classifier_agent.py

@@ -13,7 +13,7 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 from workflowai.fields import Audio
 
 
@@ -67,7 +67,7 @@ class AudioClassification(BaseModel):
     id="audio-spam-detector",
     model=Model.GEMINI_1_5_FLASH_LATEST,
 )
-async def classify_audio(audio_input: AudioInput) -> Run[AudioClassification]:
+async def classify_audio(audio_input: AudioInput) -> AudioClassification:
     """
     Analyze the audio recording to determine if it's a spam/robocall.
 
@@ -119,7 +119,7 @@ async def main():
     # )
 
     # Classify the audio
-    run = await classify_audio(AudioInput(audio=audio))
+    run = await classify_audio.run(AudioInput(audio=audio))
 
     # Print results including cost and latency information
     print(run)

examples/05_browser_text_uptime_agent.py

@@ -11,7 +11,7 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class UptimeInput(BaseModel):
@@ -40,7 +40,7 @@ class UptimeOutput(BaseModel):
     id="uptime-checker",
     model=Model.GPT_4O_MINI_LATEST,
 )
-async def check_uptime(uptime_input: UptimeInput, use_cache: str = "never") -> Run[UptimeOutput]:
+async def check_uptime(uptime_input: UptimeInput, use_cache: str = "never") -> UptimeOutput:
     """
     Fetch and analyze uptime data from an API status page.
     Use @browser-text to get the page content.
@@ -68,7 +68,7 @@ async def main():
     print("-" * 50)
 
     # Get uptime data with caching disabled
-    run = await check_uptime(uptime_input, use_cache="never")
+    run = await check_uptime.run(uptime_input, use_cache="never")
 
     # Print the run
     print(run)

examples/06_streaming_summary.py

@@ -6,12 +6,11 @@
 """
 
 import asyncio
-from collections.abc import AsyncIterator
 
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class TranslationInput(BaseModel):
@@ -30,7 +29,7 @@ class TranslationOutput(BaseModel):
 
 
 @workflowai.agent(id="french-translator", model=Model.CLAUDE_3_5_SONNET_LATEST)
-def translate_to_english(_: TranslationInput) -> AsyncIterator[Run[TranslationOutput]]:
+def translate_to_english(_: TranslationInput) -> TranslationOutput:
     """
     Translate French text into natural, fluent English.
 
@@ -74,7 +73,7 @@ async def main():
     # This ensures we can see the streaming effect in the example
     # Otherwise, subsequent runs would return the cached result instantly,
     # making it hard to observe the incremental streaming behavior
-    async for chunk in translate_to_english(TranslationInput(text=french_text), use_cache="never"):
+    async for chunk in translate_to_english.stream(TranslationInput(text=french_text), use_cache="never"):
         print(f"--- Translation Progress (Chunk {chunk_num}) ---")
         print(chunk.output.translation)
         print("-" * 50)

examples/08_pdf_agent.py

@@ -5,7 +5,7 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Run, WorkflowAIError
+from workflowai import WorkflowAIError
 from workflowai.core.domain.model import Model
 from workflowai.fields import PDF
 
@@ -21,7 +21,7 @@ class PDFAnswerOutput(BaseModel):
 
 
 @workflowai.agent(id="pdf-answer", model=Model.CLAUDE_3_5_SONNET_LATEST)
-async def answer_pdf_question(_: PDFQuestionInput) -> Run[PDFAnswerOutput]:
+async def answer_pdf_question(_: PDFQuestionInput) -> PDFAnswerOutput:
     """
     Analyze the provided PDF document and answer the given question.
     Provide a clear and concise answer based on the content found in the PDF.
@@ -55,17 +55,15 @@ async def run_pdf_answer():
     question = "How many stocks were sold? What is the total amount in USD?"
 
     try:
-        agent_run = await answer_pdf_question(
+        run = await answer_pdf_question.run(
             PDFQuestionInput(pdf=pdf, question=question),
             use_cache="auto",
         )
     except WorkflowAIError as e:
         print(f"Failed to run task. Code: {e.error.code}. Message: {e.error.message}")
         return
 
-    print("\n--------\nAgent output:\n", agent_run.output, "\n--------\n")
-    print(f"Cost: ${agent_run.cost_usd:.10f}")
-    print(f"Latency: {agent_run.duration_seconds:.2f}s")
+    print(run)
 
 
 if __name__ == "__main__":

examples/09_reply.py

@@ -20,7 +20,7 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class NameExtractionInput(BaseModel):
@@ -43,7 +43,7 @@ class NameExtractionOutput(BaseModel):
 
 
 @workflowai.agent(id="name-extractor", model=Model.GPT_4O_MINI_LATEST)
-async def extract_name(_: NameExtractionInput) -> Run[NameExtractionOutput]:
+async def extract_name(_: NameExtractionInput) -> NameExtractionOutput:
     """
     Extract a person's first and last name from a sentence.
     Be precise and consider cultural variations in name formats.
@@ -64,7 +64,7 @@ async def main():
         print(f"\nProcessing: {sentence}")
 
         # Initial extraction
-        run = await extract_name(NameExtractionInput(sentence=sentence))
+        run = await extract_name.run(NameExtractionInput(sentence=sentence))
 
         print(f"Extracted: {run.output.first_name} {run.output.last_name}")
 

examples/10_calendar_event_extraction.py

@@ -17,7 +17,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 from workflowai.fields import File
 
 
@@ -79,7 +79,7 @@ class CalendarEventOutput(BaseModel):
     id="calendar-event-extractor",
     model=Model.GPT_4O_MINI_LATEST,
 )
-async def extract_calendar_event_from_email(email_input: EmailInput) -> Run[CalendarEventOutput]:
+async def extract_calendar_event_from_email(email_input: EmailInput) -> CalendarEventOutput:
     """
     Extract calendar event details from email content.
 
@@ -112,7 +112,7 @@ async def extract_calendar_event_from_email(email_input: EmailInput) -> Run[Cale
     id="calendar-event-extractor",
     model=Model.GPT_4O_MINI_LATEST,
 )
-async def extract_calendar_event_from_image(image_input: ImageInput) -> Run[CalendarEventOutput]:
+async def extract_calendar_event_from_image(image_input: ImageInput) -> CalendarEventOutput:
     """
     Extract calendar event details from an event poster or flyer image.
 
@@ -160,7 +160,7 @@ async def main():
         """,
     )
 
-    run = await extract_calendar_event_from_email(email1)
+    run = await extract_calendar_event_from_email.run(email1)
     print(run)
 
     # Example 2: Virtual meeting with more details
@@ -238,7 +238,7 @@ async def main():
     )
 
     try:
-        run = await extract_calendar_event_from_email(email4)
+        run = await extract_calendar_event_from_email.run(email4)
         print(run)
     except workflowai.WorkflowAIError as e:
         print(f"As expected, no calendar event found: {e!s}")

examples/12_contextual_retrieval.py

@@ -9,7 +9,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class ContextGeneratorInput(BaseModel):
@@ -37,7 +37,7 @@ class ContextGeneratorOutput(BaseModel):
     id="context-generator",
     model=Model.CLAUDE_3_5_SONNET_LATEST,
 )
-async def generate_chunk_context(context_input: ContextGeneratorInput) -> Run[ContextGeneratorOutput]:
+async def generate_chunk_context(context_input: ContextGeneratorInput) -> ContextGeneratorOutput:
     """
     Here is the chunk we want to situate within the whole document.
     Please give a short succinct context to situate this chunk within the overall document
@@ -80,9 +80,9 @@ async def main():
         chunk_content=chunk_content,
     )
 
-    run = await generate_chunk_context(context_input)
+    run = await generate_chunk_context.run(context_input)
     print("\nGenerated Context:")
-    print(run.output.context)
+    print(run)
 
 
 if __name__ == "__main__":

examples/14_templated_instructions.py

@@ -23,7 +23,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class CodeReviewInput(BaseModel):
@@ -77,7 +77,7 @@ class CodeReviewOutput(BaseModel):
     id="templated-code-reviewer",
     model=Model.CLAUDE_3_5_SONNET_LATEST,
 )
-async def review_code(review_input: CodeReviewInput) -> Run[CodeReviewOutput]:
+async def review_code(review_input: CodeReviewInput) -> CodeReviewOutput:
     """
     Review code based on specified parameters and guidelines.
 
@@ -142,7 +142,7 @@ def calculate_sum(numbers):
     return result
     """
 
-    run = await review_code(
+    run = await review_code.run(
         CodeReviewInput(
             language="python",
             code=python_code,

examples/15_text_to_sql.py

@@ -17,7 +17,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class SQLGenerationInput(BaseModel):
@@ -49,7 +49,7 @@ class SQLGenerationOutput(BaseModel):
     id="text-to-sql",
     model=Model.CLAUDE_3_5_SONNET_LATEST,
 )
-async def generate_sql(review_input: SQLGenerationInput) -> Run[SQLGenerationOutput]:
+async def generate_sql(review_input: SQLGenerationInput) -> SQLGenerationOutput:
     """
     Convert natural language questions to SQL queries based on the provided schema.
 
@@ -121,7 +121,7 @@ async def main():
     # Example 1: Simple SELECT with conditions
     print("\nExample 1: Find expensive products")
     print("-" * 50)
-    run = await generate_sql(
+    run = await generate_sql.run(
         SQLGenerationInput(
             db_schema=schema,
             question="Show me all products that cost more than $100, ordered by price descending",
@@ -132,7 +132,7 @@ async def main():
     # Example 2: JOIN with aggregation
     print("\nExample 2: Customer order summary")
     print("-" * 50)
-    run = await generate_sql(
+    run = await generate_sql.run(
         SQLGenerationInput(
             db_schema=schema,
             question=(
@@ -146,7 +146,7 @@ async def main():
     # Example 3: Complex query
     print("\nExample 3: Product category analysis")
     print("-" * 50)
-    run = await generate_sql(
+    run = await generate_sql.run(
         SQLGenerationInput(
             db_schema=schema,
             question=(