Spaces:
Runtime error
Runtime error
Commit
·
3e0bc32
1
Parent(s):
6a3902b
lol
Browse files- app.py +81 -0
- packages.txt +2 -0
- requirements.txt +5 -0
app.py
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
|
| 3 |
+
import requests, re, base64, string, random
|
| 4 |
+
from PIL import Image, ImageEnhance
|
| 5 |
+
from io import BytesIO
|
| 6 |
+
import os
|
| 7 |
+
processor = TrOCRProcessor.from_pretrained("microsoft/trocr-small-printed")
|
| 8 |
+
model = VisionEncoderDecoderModel.from_pretrained("jonahgoldberg/bk_wht_8kun")
|
| 9 |
+
|
| 10 |
+
def random_string(string_length):
|
| 11 |
+
input = string.ascii_lowercase + string.digits
|
| 12 |
+
return ''.join(random.choice(input) for i in range(string_length))
|
| 13 |
+
|
| 14 |
+
# # load image examples
|
| 15 |
+
# urls = [
|
| 16 |
+
# 'https://storage.googleapis.com/trocr-captcha.appspot.com/captcha_images_v2/nfcb5.png',
|
| 17 |
+
# 'https://storage.googleapis.com/trocr-captcha.appspot.com/captcha_images_v2/p57fn.png',
|
| 18 |
+
# 'https://storage.googleapis.com/trocr-captcha.appspot.com/captcha_images_v2/w2yp7.png',
|
| 19 |
+
# 'https://storage.googleapis.com/trocr-captcha.appspot.com/captcha_images_v2/pme86.png',
|
| 20 |
+
# 'https://storage.googleapis.com/trocr-captcha.appspot.com/captcha_images_v2/w4nfx.png',
|
| 21 |
+
# 'https://storage.googleapis.com/trocr-captcha.appspot.com/captcha_images_v2/nf8b8.png'
|
| 22 |
+
# ]
|
| 23 |
+
# for idx, url in enumerate(urls):
|
| 24 |
+
# image = Image.open(requests.get(url, stream=True).raw)
|
| 25 |
+
# image.save(f"image_{idx}.png")
|
| 26 |
+
def execit(command):
|
| 27 |
+
return os.system(command)
|
| 28 |
+
###git add *.txt && git add *.py && git commit -m "lol" && git push
|
| 29 |
+
def process_image(image):
|
| 30 |
+
# prepare image
|
| 31 |
+
image_data = re.sub('^data:image/.+;base64,', '', image)
|
| 32 |
+
im = Image.open(BytesIO(base64.b64decode(image_data))).convert("RGB")
|
| 33 |
+
|
| 34 |
+
filter = ImageEnhance.Color(im)
|
| 35 |
+
im = filter.enhance(0)
|
| 36 |
+
|
| 37 |
+
# input_location = f"{random_string(9)}.png"
|
| 38 |
+
# outputfile_tmp = f"{random_string(9)}.png"
|
| 39 |
+
# outputfile_usable = f"{random_string(9)}.png"
|
| 40 |
+
# execit("input_location="+input_location)
|
| 41 |
+
# execit("outputfile_tmp="+outputfile_tmp)
|
| 42 |
+
# execit("outputfile_usable="+outputfile_usable)
|
| 43 |
+
# im.save(input_location, "png")
|
| 44 |
+
# execit('''gegl -x "<?xml version='1.0' encoding='UTF-8'?> <gegl> <node operation='gegl:ripple'> <params> <param name='amplitude'>9.9</param> <param name='period'>125.0</param> <param name='sampler-type'>nearest</param> <param name='abyss-policy'>none</param> </params> </node> <node operation='gegl:brightness-contrast'> <params> <param name='contrast'>5</param> <param name='brightness'>-1.0</param> </params> </node> <node operation='gegl:c2g'/> <node operation='gegl:load'> <params> <param name='path'>"$input_location"</param> </params> </node> </gegl>" -o $outputfile_tmp''')
|
| 45 |
+
# execit('convert $outputfile_tmp -background white -alpha remove -alpha off $outputfile_usable')
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
#Take's the picture
|
| 50 |
+
pixel_values = processor(im, return_tensors="pt").pixel_values
|
| 51 |
+
|
| 52 |
+
# generate (no beam search)
|
| 53 |
+
generated_ids = model.generate(pixel_values)
|
| 54 |
+
|
| 55 |
+
# os.remove(input_location)
|
| 56 |
+
# os.remove(outputfile_tmp)
|
| 57 |
+
# os.remove(outputfile_usable)
|
| 58 |
+
|
| 59 |
+
# decode
|
| 60 |
+
generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
| 61 |
+
|
| 62 |
+
return generated_text
|
| 63 |
+
|
| 64 |
+
title = "8kun captcha solver 1 in 8"
|
| 65 |
+
description = "Due to events. in 8chan staff moderation. I am attacking it. The gamergate shitposting days are over. and so is 8chan."
|
| 66 |
+
# article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2109.10282'>TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models</a> | <a href='https://github.com/microsoft/unilm/tree/master/trocr'>Github Repo</a></p>"
|
| 67 |
+
# examples =[["image_0.png"], ["image_1.png"], ["image_2.png"], ["image_3.png"], ["image_4.png"], ["image_5.png"]]
|
| 68 |
+
|
| 69 |
+
#css = """.output_image, .input_image {height: 600px !important}"""
|
| 70 |
+
|
| 71 |
+
iface = gr.Interface(fn=process_image,
|
| 72 |
+
# inputs=gr.inputs.Image(type="pil"),
|
| 73 |
+
inputs=gr.Textbox(placeholder="base64 string (right-click => copy-link) ..."),
|
| 74 |
+
outputs=gr.outputs.Textbox(),
|
| 75 |
+
|
| 76 |
+
title=title,
|
| 77 |
+
description=description,
|
| 78 |
+
# article=article,
|
| 79 |
+
# examples=examples
|
| 80 |
+
)
|
| 81 |
+
iface.launch(debug=True)
|
packages.txt
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
gegl
|
| 2 |
+
imagemagick
|
requirements.txt
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
gradio
|
| 2 |
+
torch
|
| 3 |
+
Pillow
|
| 4 |
+
sentencepiece
|
| 5 |
+
git+https://github.com/huggingface/transformers.git
|