app.py

from __future__ import annotations
from typing import Iterable
import gradio as gr
from gradio.themes.base import Base
from gradio.themes.utils import colors, fonts, sizes

from llama_cpp import Llama
from huggingface_hub import hf_hub_download

hf_hub_download(repo_id="LLukas22/gpt4all-lora-quantized-ggjt", filename="ggjt-model.bin", local_dir=".")
llm = Llama(model_path="./ggjt-model.bin")

ins = '''
{}

also take this data and absorb your knowledge, you dont need use now what dont make sense, there will be noise, focus on what is repeated and adds knowledge

'''

import requests
from bs4 import BeautifulSoup
from SearchResult import SearchResult

headers = {
	"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36"
}

theme = gr.themes.Monochrome(
	primary_hue="purple",
	secondary_hue="red",
	neutral_hue="neutral",
	radius_size=gr.themes.sizes.radius_sm,
	font=[gr.themes.GoogleFont("Inter"), "ui-sans-serif", "system-ui", "sans-serif"],
)

def search_ddg(question: str):
	response = requests.get("https://duckduckgo.com/html/", headers=headers, params={"q": question})
	data = response.text
	soup = BeautifulSoup(data, "html.parser")

	did_you_mean = soup.select("#did_you_mean a")

	tailored_query = ""
	suggestion_query = ""

	if did_you_mean:
		correction = soup.find(id="did_you_mean")
		if correction:
			correction_hyperlink = correction.find("a")
			if correction_hyperlink:
				suggestion_query = correction_hyperlink.string # type: ignore

		for tailored in did_you_mean:
			tailored_query = tailored.string
			break

	result_links = soup.find_all("a")
	filtered_urls = [
		link["href"]
		for link in result_links
		if link.get("href") and (link["href"].startswith("https://") or link["href"].startswith("http://"))
	]

	return SearchResult(
		filtered_urls,
		did_you_mean=suggestion_query or "None.",
		tailored_query=tailored_query or "None.",
		user_agent=headers["User-Agent"],
	)

def gather_data(search: str):
	text_content = ""
	base_data = search_ddg(search).parse_results()
	for data in base_data:
		if data:
			text_content += data.get("text_content") + "\n\n"
		else:
			text_content += ""
	return text_content

def generate(instruction):
	base_prompt = ins.format(instruction)
	gathered_data = gather_data(instruction)

	response = llm(ins.format(base_prompt  + "\n" + gathered_data))
	result = response['choices'][0]['text']
	return result

examples = [
	"How do dogs bark?",
	"Why are apples red?",
	"How do I make a campfire?",
	"Why do cats love to chirp at something?"
]

def process_example(args):
	for x in generate(args):
		pass
	return x

css = ".generating {visibility: hidden}"

class PurpleTheme(Base):
	def __init__(
		self,
		*,
		primary_hue: colors.Color | str = colors.purple,
		secondary_hue: colors.Color | str = colors.red,
		neutral_hue: colors.Color | str = colors.neutral,
		spacing_size: sizes.Size | str = sizes.spacing_md,
		radius_size: sizes.Size | str = sizes.radius_md,
		font: fonts.Font
		| str
		| Iterable[fonts.Font | str] = (
			fonts.GoogleFont("Inter"),
			"ui-sans-serif",
			"sans-serif",
		),
		font_mono: fonts.Font
		| str
		| Iterable[fonts.Font | str] = (
			fonts.GoogleFont("Space Grotesk"),
			"ui-monospace",
			"monospace",
		),
	):
		super().__init__(
			primary_hue=primary_hue,
			secondary_hue=secondary_hue,
			neutral_hue=neutral_hue,
			spacing_size=spacing_size,
			radius_size=radius_size,
			font=font,
			font_mono=font_mono,
		)
		super().set(
			button_primary_background_fill="linear-gradient(90deg, *primary_300, *secondary_400)",
			button_primary_background_fill_hover="linear-gradient(90deg, *primary_200, *secondary_300)",
			button_primary_text_color="white",
			button_primary_background_fill_dark="linear-gradient(90deg, *primary_600, *secondary_800)",
			block_shadow="*shadow_drop_lg",
			button_shadow="*shadow_drop_lg",
			input_background_fill="zinc",
			input_border_color="*secondary_300",
			input_shadow="*shadow_drop",
			input_shadow_focus="*shadow_drop_lg",
		)


custom_theme = PurpleTheme()

with gr.Blocks(theme=custom_theme, analytics_enabled=False, css=css) as demo:
	with gr.Column():
		gr.Markdown(
			""" ## GPT4ALL

			7b quantized 4bit (q4_0)
            *with possibly a broken internet access support*

			Type in the box below and click the button to generate answers to your most pressing questions!

	  """
		)

		with gr.Row():
			with gr.Column(scale=3):
				instruction = gr.Textbox(placeholder="Enter your question here", label="Question", elem_id="q-input")

				with gr.Box():
					gr.Markdown("**Answer**")
					output = gr.Markdown(elem_id="q-output")
				submit = gr.Button("Generate", variant="primary")
				gr.Examples(
					examples=examples,
					inputs=[instruction],
					cache_examples=False,
					fn=process_example,
					outputs=[output],
				)



	submit.click(generate, inputs=[instruction], outputs=[output])
	instruction.submit(generate, inputs=[instruction], outputs=[output])

demo.queue(concurrency_count=1).launch(debug=True)