import os
import time
from typing import List, Tuple, Optional
from pytube import YouTube
from moviepy.editor import *
import speech_recognition as sr
# import stanza
import pandas as pd
import numpy as np
import google.generativeai as genai
from tqdm.auto import tqdm
import time
import google.generativeai as genai
import gradio as gr
from PIL import Image
print("google-generativeai:", genai.__version__)
GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
TITLE = """Gemini Playground 😎
"""
SUBTITLE = """Play with Gemini Pro and Gemini Pro Vision API 🖇️
"""
DUPLICATE = """
"""
IMAGE_WIDTH = 512
safety_settings = [
{
"category": "HARM_CATEGORY_DANGEROUS",
"threshold": "BLOCK_NONE",
},
{
"category": "HARM_CATEGORY_HARASSMENT",
"threshold": "BLOCK_NONE",
},
{
"category": "HARM_CATEGORY_HATE_SPEECH",
"threshold": "BLOCK_NONE",
},
{
"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
"threshold": "BLOCK_NONE",
},
{
"category": "HARM_CATEGORY_DANGEROUS_CONTENT",
"threshold": "BLOCK_NONE",
},]
def extract_text_from(vid_link):
video_url = vid_link
yt = YouTube(video_url)
text = ""
audio_stream = yt.streams.get_audio_only()
audio_stream.download(filename='tmp.mp4')
audio_clip = AudioFileClip('tmp.mp4')
audio_clip.write_audiofile('tmp.wav')
r = sr.Recognizer()
with sr.AudioFile('tmp.wav') as source:
audio_data = r.record(source)
try:
text = r.recognize_google(audio_data, language='ar')
except sr.UnknownValueError:
print("Google Speech Recognition could not understand audio")
except sr.RequestError as e:
print("Could not request results from Google Speech Recognition service; {0}".format(e))
return text
def preprocess_stop_sequences(stop_sequences: str) -> Optional[List[str]]:
if not stop_sequences:
return None
return [sequence.strip() for sequence in stop_sequences.split(",")]
def preprocess_image(image: Image.Image) -> Optional[Image.Image]:
image_height = int(image.height * IMAGE_WIDTH / image.width)
return image.resize((IMAGE_WIDTH, image_height))
def user(text_prompt: str, chatbot: List[Tuple[str, str]]):
return "", chatbot + [[text_prompt, None]]
#https://www.youtube.com/watch?v=5Abk7EU5EJI
def bot(
google_key: str,
image_prompt: Optional[Image.Image],
temperature: float,
max_output_tokens: int,
stop_sequences: str,
top_k: int,
top_p: float,
chatbot: List[Tuple[str, str]]
):
google_key = google_key if google_key else GOOGLE_API_KEY
if not google_key:
raise ValueError(
"GOOGLE_API_KEY is not set. "
"Please follow the instructions in the README to set it up.")
#text_prompt = chatbot[-1][0]
txt_in = chatbot[-1][0]
if "youtube" in txt_in:
text_prompt = extract_text_from(txt_in)
else:
text_prompt = txt_in
genai.configure(api_key=google_key)
generation_config = genai.types.GenerationConfig(
temperature=0.7,
max_output_tokens=2048,
stop_sequences=preprocess_stop_sequences(stop_sequences=stop_sequences),
top_k=40,
top_p=0.95)
if image_prompt is None:
prompt= "استخرج كلمات مفتاحية من النص التالي: "+text_prompt
model = genai.GenerativeModel('gemini-pro')
response = model.generate_content(
prompt,
stream=True,
generation_config=generation_config,safety_settings=safety_settings)
response.resolve()
out1 = response.text
prompt = "أذكر لي آية من القران الكريم تتحدث عن أحد هذه المواضيع او اكثر: "+ out1 + " واشرح الآيه وفسرها باللغة العربية."
response = model.generate_content(
prompt,
stream=True,
generation_config=generation_config, safety_settings=safety_settings)
response.resolve()
else:
image_prompt = preprocess_image(image_prompt)
model = genai.GenerativeModel('gemini-pro-vision')
response = model.generate_content(
contents=[text_prompt, image_prompt],
stream=True,
generation_config=generation_config, safety_settings=safety_settings)
response.resolve()
# streaming effect
chatbot[-1][1] = ""
for chunk in response:
for i in range(0, len(chunk.text), 10):
section = chunk.text[i:i + 10]
chatbot[-1][1] += section
time.sleep(0.01)
yield chatbot
google_key_component = gr.Textbox(
label="GOOGLE API KEY",
value="",
type="password",
placeholder="...",
info="You have to provide your own GOOGLE_API_KEY for this app to function properly",
visible=GOOGLE_API_KEY is None
)
image_prompt_component = gr.Image(type="pil", label="Image", scale=1)
chatbot_component = gr.Chatbot(
label='Gemini',
bubble_full_width=False,
scale=2
)
text_prompt_component = gr.Textbox(
placeholder="Hi there!",
label="Ask me anything and press Enter"
)
run_button_component = gr.Button()
temperature_component = gr.Slider(
minimum=0,
maximum=1.0,
value=0.4,
step=0.05,
label="Temperature",
info=(
"Temperature controls the degree of randomness in token selection. Lower "
"temperatures are good for prompts that expect a true or correct response, "
"while higher temperatures can lead to more diverse or unexpected results. "
))
max_output_tokens_component = gr.Slider(
minimum=1,
maximum=2048,
value=1024,
step=1,
label="Token limit",
info=(
"Token limit determines the maximum amount of text output from one prompt. A "
"token is approximately four characters. The default value is 2048."
))
stop_sequences_component = gr.Textbox(
label="Add stop sequence",
value="",
type="text",
placeholder="STOP, END",
info=(
"A stop sequence is a series of characters (including spaces) that stops "
"response generation if the model encounters it. The sequence is not included "
"as part of the response. You can add up to five stop sequences."
))
top_k_component = gr.Slider(
minimum=1,
maximum=40,
value=32,
step=1,
label="Top-K",
info=(
"Top-k changes how the model selects tokens for output. A top-k of 1 means the "
"selected token is the most probable among all tokens in the model’s "
"vocabulary (also called greedy decoding), while a top-k of 3 means that the "
"next token is selected from among the 3 most probable tokens (using "
"temperature)."
))
top_p_component = gr.Slider(
minimum=0,
maximum=1,
value=1,
step=0.01,
label="Top-P",
info=(
"Top-p changes how the model selects tokens for output. Tokens are selected "
"from most probable to least until the sum of their probabilities equals the "
"top-p value. For example, if tokens A, B, and C have a probability of .3, .2, "
"and .1 and the top-p value is .5, then the model will select either A or B as "
"the next token (using temperature). "
))
user_inputs = [
text_prompt_component,
chatbot_component
]
bot_inputs = [
google_key_component,
image_prompt_component,
temperature_component,
max_output_tokens_component,
stop_sequences_component,
top_k_component,
top_p_component,
chatbot_component
]
with gr.Blocks() as demo:
gr.HTML(TITLE)
gr.HTML(SUBTITLE)
gr.HTML(DUPLICATE)
with gr.Column():
google_key_component.render()
with gr.Row():
image_prompt_component.render()
chatbot_component.render()
text_prompt_component.render()
run_button_component.render()
with gr.Accordion("Parameters", open=False):
temperature_component.render()
max_output_tokens_component.render()
stop_sequences_component.render()
with gr.Accordion("Advanced", open=False):
top_k_component.render()
top_p_component.render()
run_button_component.click(
fn=user,
inputs=user_inputs,
outputs=[text_prompt_component, chatbot_component],
queue=False
).then(
fn=bot, inputs=bot_inputs, outputs=[chatbot_component],
)
text_prompt_component.submit(
fn=user,
inputs=user_inputs,
outputs=[text_prompt_component, chatbot_component],
queue=False
).then(
fn=bot, inputs=bot_inputs, outputs=[chatbot_component],
)
demo.queue(max_size=99).launch(debug=False, show_error=True)