ChatGPT Tutorial#
Run on Google Colab | View source on GitHub | Download notebook |
Start EVA server#
We are reusing the start server notebook for launching the EVA server
!wget -nc "https://raw.githubusercontent.com/georgia-tech-db/eva/master/tutorials/00-start-eva-server.ipynb"
%run 00-start-eva-server.ipynb
cursor = connect_to_server()
File ‘00-start-eva-server.ipynb’ already there; not retrieving.
[notice] A new release of pip is available: 23.0.1 -> 23.1.2
[notice] To update, run: pip install --upgrade pip
Note: you may need to restart the kernel to use updated packages.
nohup eva_server > eva.log 2>&1 &
[notice] A new release of pip is available: 23.0.1 -> 23.1.2
[notice] To update, run: pip install --upgrade pip
Note: you may need to restart the kernel to use updated packages.
Download News Video and ChatGPT UDF#
# Download News Video
!wget -nc "https://www.dropbox.com/s/rfm1kds2mv77pca/russia_ukraine.mp4?dl=0" -O russia_ukraine.mp4
# Download ChatGPT UDF if needed
!wget -nc https://raw.githubusercontent.com/georgia-tech-db/eva/master/eva/udfs/chatgpt.py -O chatgpt.py
File ‘russia_ukraine.mp4’ already there; not retrieving.
File ‘chatgpt.py’ already there; not retrieving.
Visualize Video#
from IPython.display import Video
Video("russia_ukraine.mp4", height=450, width=800, embed=True)
Set your OpenAI API key here#
from eva.configuration.configuration_manager import ConfigurationManager
import os
# Assuming that the key is stored as an environment variable
open_ai_key = os.environ.get('OPENAI_KEY')
ConfigurationManager().update_value("third_party", "openai_api_key", open_ai_key)
# Drop the UDF if it already exists
drop_udf_query = f"DROP UDF IF EXISTS ChatGPT;"
cursor.execute(drop_udf_query)
response = cursor.fetch_all()
response.as_df()
# Register the ChatGPT UDF in EVA
create_udf_query = f"""CREATE UDF ChatGPT
IMPL 'chatgpt.py' """
cursor.execute(create_udf_query)
response = cursor.fetch_all()
response.as_df()
0 | |
---|---|
0 | UDF ChatGPT successfully added to the database. |
Run the ChatGPT UDF#
#load the video
cursor.execute("LOAD VIDEO 'russia_ukraine.mp4' INTO VIDEOS;")
response = cursor.fetch_all()
response.as_df()
0 | |
---|---|
0 | Number of loaded VIDEO: 1 |
# Drop the Text Summarization UDF if needed
cursor.execute("DROP UDF IF EXISTS SpeechRecognizer;")
response = cursor.fetch_all()
response.as_df()
# Create a Text Summarization UDF using Hugging Face
text_summarizer_udf_creation = """
CREATE UDF SpeechRecognizer
TYPE HuggingFace
'task' 'automatic-speech-recognition'
'model' 'openai/whisper-base';
"""
cursor.execute(text_summarizer_udf_creation)
response = cursor.fetch_all()
response.as_df()
0 | |
---|---|
0 | UDF SpeechRecognizer successfully added to the... |
# Drop the table if needed
cursor.execute("DROP TABLE IF EXISTS TEXT_SUMMARY;")
response = cursor.fetch_all()
response.as_df()
# Create a materialized view of the text summarization output
text_summarization_query = """
CREATE MATERIALIZED VIEW
TEXT_SUMMARY(text) AS
SELECT SpeechRecognizer(audio) FROM VIDEOS;
"""
cursor.execute(text_summarization_query)
response = cursor.fetch_all()
response.as_df()
# Run ChatGPT over the Text Summary extracted by Whisper
chatgpt_udf = """
SELECT ChatGPT('Is this video summary related to Ukraine russia war',text)
FROM TEXT_SUMMARY;
"""
cursor.execute(chatgpt_udf)
response = cursor.fetch_all()
response.as_df()
chatgpt.response | |
---|---|
0 | No, this video summary is not related to the U... |
1 | Yes, the video summary is related to the Ukrai... |
Check if it works on an SNL Video#
# Download Entertainment Video
!wget -nc "https://www.dropbox.com/s/u66im8jw2s1dmuw/snl.mp4?dl=0" -O snl.mp4
cursor.execute("DROP TABLE IF EXISTS SNL_VIDEO;")
response = cursor.fetch_all()
response.as_df()
cursor.execute("LOAD VIDEO 'snl.mp4' INTO SNL_VIDEO;")
response = cursor.fetch_all()
response.as_df()
File ‘snl.mp4’ already there; not retrieving.
0 | |
---|---|
0 | Number of loaded VIDEO: 1 |
from IPython.display import Video
Video("snl.mp4", height=450, width=800, embed=True)