File size: 9,048 Bytes
75c2120 008fdc7 75c2120 1e02081 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 |
import glob
import os
import random
import re
import string
import gradio as gr
import openai
from icrawler import ImageDownloader
from icrawler.builtin import GoogleImageCrawler, BingImageCrawler
from uuid import uuid4
from pptx import Presentation
bad_coding_practice = ''.join(random.choice(string.ascii_uppercase + string.ascii_lowercase + string.digits) for _ in
range(16))
def refresh_bad_coding_practice():
global bad_coding_practice
bad_coding_practice = ''.join(random.choice(string.ascii_uppercase + string.ascii_lowercase + string.digits)
for _ in range(16))
return
class PrefixNameDownloader(ImageDownloader):
def get_filename(self, task, default_ext):
filename = super(PrefixNameDownloader, self).get_filename(
task, default_ext)
print(bad_coding_practice)
return 'prefix_' + bad_coding_practice + filename
def generate_ppt(file, topic, slide_length, api_key):
print(file.name)
root = Presentation(file.name)
openai.api_key = api_key
message = f"""
Create content for a slideshow presentation.
The content's topic is {topic}.
The slideshow is {slide_length} slides long.
The content is written in the language of the content I give you above.
You are allowed to use the following slide types:
Slide types:
Title Slide - (Title, Subtitle)
Content Slide - (Title, Content)
Image Slide - (Title, Content, Image)
Thanks Slide - (Title)
Put this tag before the Title Slide: [L_TS]
Put this tag before the Content Slide: [L_CS]
Put this tag before the Image Slide: [L_IS]
Put this tag before the Thanks Slide: [L_THS]
Put "[SLIDEBREAK]" after each slide
For example:
[L_TS]
[TITLE]Mental Health[/TITLE]
[SLIDEBREAK]
[L_CS]
[TITLE]Mental Health Definition[/TITLE]
[CONTENT]
1. Definition: A person’s condition with regard to their psychological and emotional well-being
2. Can impact one's physical health
3. Stigmatized too often.
[/CONTENT]
[SLIDEBREAK]
Put this tag before the Title: [TITLE]
Put this tag after the Title: [/TITLE]
Put this tag before the Subitle: [SUBTITLE]
Put this tag after the Subtitle: [/SUBTITLE]
Put this tag before the Content: [CONTENT]
Put this tag after the Content: [/CONTENT]
Put this tag before the Image: [IMAGE]
Put this tag after the Image: [/IMAGE]
Elaborate on the Content, provide as much information as possible.
You put a [/CONTENT] at the end of the Content.
Do not reply as if you are talking about the slideshow itself. (ex. "Include pictures here about...")
Do not include any special characters (?, !, ., :, ) in the Title.
Do not include any additional information in your response and stick to the format."""
response = openai.ChatCompletion.create(
model="gpt-3.5-turbo",
messages=[
{"role": "user", "content": message}
]
)
# """ Ref for slide types:
# 0 -> title and subtitle
# 1 -> title and content
# 2 -> section header
# 3 -> two content
# 4 -> Comparison
# 5 -> Title only
# 6 -> Blank
# 7 -> Content with caption
# 8 -> Pic with caption
# """
def delete_all_slides():
for i in range(len(root.slides) - 1, -1, -1):
r_id = root.slides._sldIdLst[i].rId
root.part.drop_rel(r_id)
del root.slides._sldIdLst[i]
def create_title_slide(title, subtitle):
layout = root.slide_layouts[0]
slide = root.slides.add_slide(layout)
slide.shapes.title.text = title
slide.placeholders[1].text = subtitle
def create_section_header_slide(title):
layout = root.slide_layouts[2]
slide = root.slides.add_slide(layout)
slide.shapes.title.text = title
def create_title_and_content_slide(title, content):
layout = root.slide_layouts[1]
slide = root.slides.add_slide(layout)
slide.shapes.title.text = title
slide.placeholders[1].text = content
def create_title_and_content_and_image_slide(title, content, image_query):
layout = root.slide_layouts[8]
slide = root.slides.add_slide(layout)
slide.shapes.title.text = title
slide.placeholders[2].text = content
refresh_bad_coding_practice()
bing_crawler = GoogleImageCrawler(downloader_cls=PrefixNameDownloader, storage={'root_dir': os.getcwd()})
bing_crawler.crawl(keyword=image_query, max_num=1)
dir_path = os.path.dirname(os.path.realpath(__file__))
file_name = glob.glob(f"prefix_{bad_coding_practice}*")
print(file_name)
img_path = os.path.join(dir_path, file_name[0])
slide.shapes.add_picture(img_path, slide.placeholders[1].left, slide.placeholders[1].top,
slide.placeholders[1].width, slide.placeholders[1].height)
def find_text_in_between_tags(text, start_tag, end_tag):
start_pos = text.find(start_tag)
end_pos = text.find(end_tag)
result = []
while start_pos > -1 and end_pos > -1:
text_between_tags = text[start_pos + len(start_tag):end_pos]
result.append(text_between_tags)
start_pos = text.find(start_tag, end_pos + len(end_tag))
end_pos = text.find(end_tag, start_pos)
res1 = "".join(result)
res2 = re.sub(r"\[IMAGE\].*?\[/IMAGE\]", '', res1)
if len(result) > 0:
return res2
else:
return ""
def search_for_slide_type(text):
tags = ["[L_TS]", "[L_CS]", "[L_IS]", "[L_THS]"]
found_text = next((s for s in tags if s in text), None)
return found_text
def parse_response(reply):
list_of_slides = reply.split("[SLIDEBREAK]")
for slide in list_of_slides:
slide_type = search_for_slide_type(slide)
if slide_type == "[L_TS]":
create_title_slide(find_text_in_between_tags(str(slide), "[TITLE]", "[/TITLE]"),
find_text_in_between_tags(str(slide), "[SUBTITLE]", "[/SUBTITLE]"))
elif slide_type == "[L_CS]":
create_title_and_content_slide("".join(find_text_in_between_tags(str(slide), "[TITLE]", "[/TITLE]")),
"".join(find_text_in_between_tags(str(slide), "[CONTENT]",
"[/CONTENT]")))
elif slide_type == "[L_IS]":
create_title_and_content_and_image_slide("".join(find_text_in_between_tags(str(slide), "[TITLE]",
"[/TITLE]")),
"".join(find_text_in_between_tags(str(slide), "[CONTENT]",
"[/CONTENT]")),
"".join(find_text_in_between_tags(str(slide), "[IMAGE]",
"[/IMAGE]")))
elif slide_type == "[L_THS]":
create_section_header_slide("".join(find_text_in_between_tags(str(slide), "[TITLE]", "[/TITLE]")))
def find_title():
return root.slides[0].shapes.title.text
delete_all_slides()
print(response)
parse_response(response['choices'][0]['message']['content'])
name_ = str(uuid4()).replace('-', '')
root.save(f"./{name_}.pptx")
print("done")
dir_path = "./"
prefix = "prefix_"
for file_name in os.listdir(dir_path):
if file_name.startswith(prefix):
file_path = os.path.join(dir_path, file_name)
if os.path.isfile(file_path):
os.remove(file_path)
return f"./{name_}.pptx"
with gr.Blocks(title="ChatGPT PPT框架生成") as demo:
gr.Markdown("""<h1><center>ChatGPT PPT框架生成</center></h1>""")
with gr.Row():
with gr.Column():
openai_token = gr.Textbox(label="OpenAI API Key")
topic = gr.Textbox(label="PPT的主题或内容")
length = gr.Slider(minimum=1, maximum=20, value=6, label="生成的PPT页数", step=1)
theme = gr.File(value="./theme.pptx", file_types=['pptx', 'ppt'], label="PPT模版")
output_file = gr.File(interactive=False)
topic.submit(
fn=generate_ppt,
inputs=[theme, topic, length, openai_token],
outputs=[output_file]
)
submit = gr.Button("生成")
submit.click(
fn=generate_ppt,
inputs=[theme, topic, length, openai_token],
outputs=[output_file]
)
if __name__ == "__main__":
demo.launch()
|