Spaces:
Sleeping
Sleeping
File size: 5,056 Bytes
d2cb17f 409e708 d2cb17f 2591f57 d2cb17f 409e708 d2cb17f 409e708 d2cb17f 2591f57 d2cb17f 04e96c0 d2cb17f 07216f8 d2cb17f |
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 |
import pymupdf
from io import BytesIO
from PIL import Image
import pdfplumber
import ast
import google.generativeai as genai
from PIL import Image, ImageDraw
import openai
import requests
import os
# from constants import GEMINI_API_KEY, OPENAI_API_KEY
from utils import (
draw_boxes,
pdf_to_images,
parse_bboxs_gemini_flash,
convert_pdf_to_images,
encode_image_to_base64,
)
def extract_images_pymupdf(pdf_file):
pdf_path = "extract_images/input_docs/uploaded_pdf.pdf"
with open(pdf_path, "wb") as f:
f.write(pdf_file)
doc = pymupdf.open(pdf_path)
images = []
for page_idx, page in enumerate(doc):
for img_index, img in enumerate(doc.get_page_images(page_idx)):
xref = img[0]
base_image = doc.extract_image(xref)
image_bytes = base_image["image"]
image = Image.open(BytesIO(image_bytes))
images.append(image)
return images if images != [] else None
def extract_images_pdfplumber(pdf_file):
pdf_path = "extract_images/input_docs/uploaded_pdf.pdf"
with open(pdf_path, "wb") as f:
f.write(pdf_file)
images = []
output_dir = "extract_tables/table_outputs"
pdf_obj = pdfplumber.open(pdf_path)
for page_idx, page in enumerate(pdf_obj.pages):
page_bbox = []
for image_idx, image in enumerate(page.images):
page_height = page.height
image_bbox = (
image["x0"],
page_height - image["y1"],
image["x1"],
page_height - image["y0"],
)
page_bbox.append(image_bbox)
cropped_page = page.crop(image_bbox)
image_obj = cropped_page.to_image(resolution=400)
image_path = os.path.join(
output_dir, f"image-{page_idx + 1}-{image_idx}.png"
)
image_obj.save(image_path)
image = Image.open(image_path)
images.append(image)
return images if images != [] else None
def extract_images_gemini(model, pdf_file):
gemini_api_key = os.getenv("GEMINI_API_KEY")
genai.configure(api_key=gemini_api_key)
gemini_model = genai.GenerativeModel(model)
prompt = f"Extract the bounding boxes of all the images present in this page. Return the bounding boxes as list of lists. Do not include anyother text or symbols in the output"
pdf_path = "extract_images/input_docs/uploaded_pdf.pdf"
with open(pdf_path, "wb") as f:
f.write(pdf_file)
images = []
pdf_images = pdf_to_images(pdf_path)
for page in pdf_images:
img = Image.open(page).convert("RGB")
response = gemini_model.generate_content([img, prompt], stream=False)
response.resolve()
print(response.text)
if model == "gemini-pro-vision":
page_bbox = ast.literal_eval(response.text)
elif model == "gemini-1.5-flash-latest":
page_bbox = parse_bboxs_gemini_flash(response.text)
image = draw_boxes(page, page_bbox)
images.append(image)
return images
def extract_images_gpt(model, pdf_file):
openai_api_key = os.getenv("OPENAI_API_KEY")
openai.api_key = openai_api_key
image_media_type = "image/png"
pdf_path = "extract_images/input_docs/uploaded_pdf.pdf"
with open(pdf_path, "wb") as f:
f.write(pdf_file)
images = convert_pdf_to_images(pdf_path)
image_paths = pdf_to_images(pdf_path)
headers = {
"Content-Type": "application/json",
"Authorization": f"Bearer {openai.api_key}",
}
extracted_images = []
for page_idx, image in enumerate(images):
base64_string = encode_image_to_base64(image)
payload = {
"model": model,
"messages": [
{
"role": "user",
"content": [
{
"type": "text",
"text": "Extract bounding boxes of all the images present in this page. Return bounding boxes as liat of lists and don't provide any other text in the response.",
},
{
"type": "image_url",
"image_url": {
"url": f"data:image/jpeg;base64,{base64_string}"
},
},
],
}
],
}
response = requests.post(
"https://api.openai.com/v1/chat/completions", headers=headers, json=payload
)
response_json = response.json()
print(response_json["choices"][0]["message"]["content"])
if "choices" in response_json and len(response_json["choices"]) > 0:
extracted_images.append(
draw_boxes(
image_paths[page_idx],
ast.literal_eval(response_json["choices"][0]["message"]["content"]),
)
)
return extracted_images
|