doc: add more doc on env.example and rename stt, remove instrumentation if not locally run

app.py

@@ -11,8 +11,11 @@ from multiagents import MultiAgent
 from phoenix.otel import register
 from openinference.instrumentation.smolagents import SmolagentsInstrumentor
 
-register()
-SmolagentsInstrumentor().instrument()
+# use space_host var to determine if running in HF space or locally, if so register local instrumentation
+space_host_startup = os.getenv("SPACE_HOST")
+if not space_host_startup:
+    register()
+    SmolagentsInstrumentor().instrument()
 
 # (Keep Constants as is)
 # --- Constants ---
@@ -21,14 +24,13 @@ load_dotenv()
 
 max_questions = 20
 
-
 def run_and_submit_all(nb_questions: int, profile: gr.OAuthProfile | None): 
     """
     Fetches all questions, runs my Agent on them, submits all answers,
     and displays the results.
     """
     # --- Determine HF Space Runtime URL and Repo URL ---
-    space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
+    space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code   
 
     if profile:
         username= f"{profile.username}"
@@ -96,8 +98,9 @@ def run_and_submit_all(nb_questions: int, profile: gr.OAuthProfile | None):
             agent_question = question_text
             if file_question_url:
                 agent_question += f"\n\nFile URL: {file_question_url}"
-            
+
             submitted_answer = agent(agent_question)
+            
             answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
         except Exception as e:

env.example

@@ -1 +1,3 @@
-HF_TOKEN=<<your hf token>>
+HF_TOKEN=<<your hf token>>
+SPACE_ID=<<your hf sapce id>>
+OPENAI_API_KEY=<<your open ai key>>

multiagents.py

@@ -7,7 +7,7 @@ from smolagents import OpenAIServerModel
 from tools.fetch import fetch_webpage, search_web
 from smolagents import PythonInterpreterTool
 from tools.yttranscript import get_youtube_transcript, get_youtube_title_description
-from tools.stt import stt
+from tools.stt import get_text_transcript_from_audio_file
 from tools.image import analyze_image
 from tools.mylogger import save_file_with_timestamp, mylog
 import myprompts
@@ -70,7 +70,7 @@ audiovideo_agent = CodeAgent(
     tools=[
         get_youtube_transcript,
         get_youtube_title_description,
-        stt,
+        get_text_transcript_from_audio_file,
         analyze_image
     ],
     name="audiovideo_agent",

tools/stt.py

@@ -5,9 +5,11 @@ import subprocess
 import requests
 import uuid
 from smolagents import tool
+import dotenv
+dotenv.load_dotenv()
 
 @tool
-def stt(file_url: str, language: str = "en-US") -> str:
+def get_text_transcript_from_audio_file(file_url: str, language: str = "en-US") -> str:
     """
     Convert speech to text using local whisper model.
     This function downloads an audio file from a given URL, converts it to WAV format if necessary,
@@ -23,7 +25,9 @@ def stt(file_url: str, language: str = "en-US") -> str:
 
     file_name = uuid.uuid4().hex +".mp3"
 
-    dest_folder = "c:\\Users\\mathi\\dev\\stt"
+    dest_folder = os.getenv("STT_FOLDER")
+    if not dest_folder:
+        dest_folder = '.'
     file_path = os.path.join(dest_folder + "\\tmp", file_name)
     # 1. download the file from url (in pure python without wget or curl)
     if not os.path.exists(file_name):            
@@ -64,5 +68,5 @@ def stt(file_url: str, language: str = "en-US") -> str:
     return text
 
 if __name__ == "__main__":
-    transcript = stt("https://agents-course-unit4-scoring.hf.space/files/99c9cc74-fdc8-46c6-8f8d-3ce2d3bfeea3", )
+    transcript = get_text_transcript_from_audio_file("https://agents-course-unit4-scoring.hf.space/files/99c9cc74-fdc8-46c6-8f8d-3ce2d3bfeea3", )
     print(transcript)