Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -11,7 +11,6 @@ import fitz
|
|
11 |
import gradio as gr
|
12 |
import requests
|
13 |
import torch
|
14 |
-
from huggingface_hub import snapshot_download
|
15 |
from PIL import Image, ImageDraw, ImageFont
|
16 |
from transformers import (
|
17 |
Qwen2_5_VLForConditionalGeneration,
|
@@ -46,7 +45,7 @@ prompt = """Please output the layout information from the PDF image, including e
|
|
46 |
5. Final Output: The entire output must be a single JSON object.
|
47 |
"""
|
48 |
|
49 |
-
# Load
|
50 |
MODEL_ID_M = "prithivMLmods/Camel-Doc-OCR-062825"
|
51 |
processor_m = AutoProcessor.from_pretrained(MODEL_ID_M, trust_remote_code=True)
|
52 |
model_m = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
@@ -55,7 +54,6 @@ model_m = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
|
55 |
torch_dtype=torch.float16
|
56 |
).to(device).eval()
|
57 |
|
58 |
-
# Load Megalodon-OCR-Sync-0713
|
59 |
MODEL_ID_T = "prithivMLmods/Megalodon-OCR-Sync-0713"
|
60 |
processor_t = AutoProcessor.from_pretrained(MODEL_ID_T, trust_remote_code=True)
|
61 |
model_t = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
@@ -64,7 +62,6 @@ model_t = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
|
64 |
torch_dtype=torch.float16
|
65 |
).to(device).eval()
|
66 |
|
67 |
-
# Load Nanonets-OCR-s
|
68 |
MODEL_ID_C = "nanonets/Nanonets-OCR-s"
|
69 |
processor_c = AutoProcessor.from_pretrained(MODEL_ID_C, trust_remote_code=True)
|
70 |
model_c = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
@@ -73,7 +70,6 @@ model_c = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
|
73 |
torch_dtype=torch.float16
|
74 |
).to(device).eval()
|
75 |
|
76 |
-
# Load MonkeyOCR
|
77 |
MODEL_ID_G = "echo840/MonkeyOCR"
|
78 |
SUBFOLDER = "Recognition"
|
79 |
processor_g = AutoProcessor.from_pretrained(
|
@@ -88,13 +84,10 @@ model_g = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
|
88 |
torch_dtype=torch.float16
|
89 |
).to(device).eval()
|
90 |
|
91 |
-
|
92 |
# Utility functions
|
93 |
def round_by_factor(number: int, factor: int) -> int:
|
94 |
-
"""Returns the closest integer to 'number' that is divisible by 'factor'."""
|
95 |
return round(number / factor) * factor
|
96 |
|
97 |
-
|
98 |
def smart_resize(
|
99 |
height: int,
|
100 |
width: int,
|
@@ -102,18 +95,10 @@ def smart_resize(
|
|
102 |
min_pixels: int = 3136,
|
103 |
max_pixels: int = 11289600,
|
104 |
):
|
105 |
-
"""Rescales the image so that the following conditions are met:
|
106 |
-
1. Both dimensions (height and width) are divisible by 'factor'.
|
107 |
-
2. The total number of pixels is within the range ['min_pixels', 'max_pixels'].
|
108 |
-
3. The aspect ratio of the image is maintained as closely as possible.
|
109 |
-
"""
|
110 |
if max(height, width) / min(height, width) > 200:
|
111 |
-
raise ValueError(
|
112 |
-
f"absolute aspect ratio must be smaller than 200, got {max(height, width) / min(height, width)}"
|
113 |
-
)
|
114 |
h_bar = max(factor, round_by_factor(height, factor))
|
115 |
w_bar = max(factor, round_by_factor(width, factor))
|
116 |
-
|
117 |
if h_bar * w_bar > max_pixels:
|
118 |
beta = math.sqrt((height * width) / max_pixels)
|
119 |
h_bar = round_by_factor(height / beta, factor)
|
@@ -124,9 +109,7 @@ def smart_resize(
|
|
124 |
w_bar = round_by_factor(width * beta, factor)
|
125 |
return h_bar, w_bar
|
126 |
|
127 |
-
|
128 |
def fetch_image(image_input, min_pixels: int = None, max_pixels: int = None):
|
129 |
-
"""Fetch and process an image"""
|
130 |
if isinstance(image_input, str):
|
131 |
if image_input.startswith(("http://", "https://")):
|
132 |
response = requests.get(image_input)
|
@@ -137,31 +120,20 @@ def fetch_image(image_input, min_pixels: int = None, max_pixels: int = None):
|
|
137 |
image = image_input.convert('RGB')
|
138 |
else:
|
139 |
raise ValueError(f"Invalid image input type: {type(image_input)}")
|
140 |
-
|
141 |
-
if min_pixels is not None or max_pixels is not None:
|
142 |
min_pixels = min_pixels or MIN_PIXELS
|
143 |
max_pixels = max_pixels or MAX_PIXELS
|
144 |
-
height, width = smart_resize(
|
145 |
-
image.height,
|
146 |
-
image.width,
|
147 |
-
factor=IMAGE_FACTOR,
|
148 |
-
min_pixels=min_pixels,
|
149 |
-
max_pixels=max_pixels
|
150 |
-
)
|
151 |
image = image.resize((width, height), Image.LANCZOS)
|
152 |
-
|
153 |
return image
|
154 |
|
155 |
-
|
156 |
def load_images_from_pdf(pdf_path: str) -> List[Image.Image]:
|
157 |
-
"""Load images from PDF file"""
|
158 |
images = []
|
159 |
try:
|
160 |
pdf_document = fitz.open(pdf_path)
|
161 |
for page_num in range(len(pdf_document)):
|
162 |
page = pdf_document.load_page(page_num)
|
163 |
-
|
164 |
-
mat = fitz.Matrix(2.0, 2.0) # Increase resolution
|
165 |
pix = page.get_pixmap(matrix=mat)
|
166 |
img_data = pix.tobytes("ppm")
|
167 |
image = Image.open(BytesIO(img_data)).convert('RGB')
|
@@ -169,157 +141,86 @@ def load_images_from_pdf(pdf_path: str) -> List[Image.Image]:
|
|
169 |
pdf_document.close()
|
170 |
except Exception as e:
|
171 |
print(f"Error loading PDF: {e}")
|
172 |
-
return []
|
173 |
return images
|
174 |
|
175 |
-
|
176 |
def draw_layout_on_image(image: Image.Image, layout_data: List[Dict]) -> Image.Image:
|
177 |
-
"""Draw layout bounding boxes on image"""
|
178 |
img_copy = image.copy()
|
179 |
draw = ImageDraw.Draw(img_copy)
|
180 |
-
|
181 |
-
# Colors for different categories
|
182 |
colors = {
|
183 |
-
'Caption': '#FF6B6B',
|
184 |
-
'
|
185 |
-
'
|
186 |
-
'
|
187 |
-
'Page-footer': '#FFEAA7',
|
188 |
-
'Page-header': '#DDA0DD',
|
189 |
-
'Picture': '#FFD93D',
|
190 |
-
'Section-header': '#6C5CE7',
|
191 |
-
'Table': '#FD79A8',
|
192 |
-
'Text': '#74B9FF',
|
193 |
-
'Title': '#E17055'
|
194 |
}
|
195 |
-
|
196 |
try:
|
197 |
-
|
198 |
-
|
199 |
-
|
200 |
-
|
201 |
-
|
202 |
-
|
203 |
-
|
204 |
-
|
205 |
-
|
206 |
-
|
207 |
-
|
208 |
-
|
209 |
-
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
|
214 |
-
label_bbox = draw.textbbox((0, 0), label, font=font)
|
215 |
-
label_width = label_bbox[2] - label_bbox[0]
|
216 |
-
label_height = label_bbox[3] - label_bbox[1]
|
217 |
-
|
218 |
-
# Position label above the box
|
219 |
-
label_x = bbox[0]
|
220 |
-
label_y = max(0, bbox[1] - label_height - 2)
|
221 |
-
|
222 |
-
# Draw background for label
|
223 |
-
draw.rectangle(
|
224 |
-
[label_x, label_y, label_x + label_width + 4, label_y + label_height + 2],
|
225 |
-
fill=color
|
226 |
-
)
|
227 |
-
|
228 |
-
# Draw text
|
229 |
-
draw.text((label_x + 2, label_y + 1), label, fill='white', font=font)
|
230 |
-
|
231 |
-
except Exception as e:
|
232 |
-
print(f"Error drawing layout: {e}")
|
233 |
-
|
234 |
return img_copy
|
235 |
|
236 |
-
|
237 |
def is_arabic_text(text: str) -> bool:
|
238 |
-
"""Check if text in headers and paragraphs contains mostly Arabic characters"""
|
239 |
if not text:
|
240 |
return False
|
241 |
-
|
242 |
-
# Extract text from headers and paragraphs only
|
243 |
-
# Match markdown headers (# ## ###) and regular paragraph text
|
244 |
header_pattern = r'^#{1,6}\s+(.+)$'
|
245 |
paragraph_pattern = r'^(?!#{1,6}\s|!\[|```|\||\s*[-*+]\s|\s*\d+\.\s)(.+)$'
|
246 |
-
|
247 |
content_text = []
|
248 |
-
|
249 |
for line in text.split('\n'):
|
250 |
line = line.strip()
|
251 |
if not line:
|
252 |
continue
|
253 |
-
|
254 |
-
# Check for headers
|
255 |
header_match = re.match(header_pattern, line, re.MULTILINE)
|
256 |
if header_match:
|
257 |
content_text.append(header_match.group(1))
|
258 |
continue
|
259 |
-
|
260 |
-
# Check for paragraph text (exclude lists, tables, code blocks, images)
|
261 |
if re.match(paragraph_pattern, line, re.MULTILINE):
|
262 |
content_text.append(line)
|
263 |
-
|
264 |
if not content_text:
|
265 |
return False
|
266 |
-
|
267 |
-
# Join all content text and check for Arabic characters
|
268 |
combined_text = ' '.join(content_text)
|
269 |
-
|
270 |
-
# Arabic Unicode ranges
|
271 |
arabic_chars = 0
|
272 |
total_chars = 0
|
273 |
-
|
274 |
for char in combined_text:
|
275 |
if char.isalpha():
|
276 |
total_chars += 1
|
277 |
-
# Arabic script ranges
|
278 |
if ('\u0600' <= char <= '\u06FF') or ('\u0750' <= char <= '\u077F') or ('\u08A0' <= char <= '\u08FF'):
|
279 |
arabic_chars += 1
|
280 |
-
|
281 |
-
if total_chars == 0:
|
282 |
-
return False
|
283 |
-
|
284 |
-
# Consider text as Arabic if more than 50% of alphabetic characters are Arabic
|
285 |
-
return (arabic_chars / total_chars) > 0.5
|
286 |
-
|
287 |
|
288 |
def layoutjson2md(image: Image.Image, layout_data: List[Dict], text_key: str = 'text') -> str:
|
289 |
-
"""Convert layout JSON to markdown format"""
|
290 |
import base64
|
291 |
from io import BytesIO
|
292 |
-
|
293 |
markdown_lines = []
|
294 |
-
|
295 |
try:
|
296 |
-
|
297 |
-
sorted_items = sorted(layout_data, key=lambda x: (x.get('bbox',), x.get('bbox',)))
|
298 |
-
|
299 |
for item in sorted_items:
|
300 |
category = item.get('category', '')
|
301 |
text = item.get(text_key, '')
|
302 |
bbox = item.get('bbox', [])
|
303 |
-
|
304 |
if category == 'Picture':
|
305 |
-
# Extract image region and embed it
|
306 |
if bbox and len(bbox) == 4:
|
307 |
try:
|
308 |
-
# Extract the image region
|
309 |
x1, y1, x2, y2 = bbox
|
310 |
-
# Ensure coordinates are within image bounds
|
311 |
x1, y1 = max(0, int(x1)), max(0, int(y1))
|
312 |
x2, y2 = min(image.width, int(x2)), min(image.height, int(y2))
|
313 |
-
|
314 |
if x2 > x1 and y2 > y1:
|
315 |
cropped_img = image.crop((x1, y1, x2, y2))
|
316 |
-
|
317 |
-
# Convert to base64 for embedding
|
318 |
buffer = BytesIO()
|
319 |
cropped_img.save(buffer, format='PNG')
|
320 |
img_data = base64.b64encode(buffer.getvalue()).decode()
|
321 |
-
|
322 |
-
# Add as markdown image
|
323 |
markdown_lines.append(f"\n")
|
324 |
else:
|
325 |
markdown_lines.append("\n")
|
@@ -339,13 +240,11 @@ def layoutjson2md(image: Image.Image, layout_data: List[Dict], text_key: str = '
|
|
339 |
elif category == 'List-item':
|
340 |
markdown_lines.append(f"- {text}\n")
|
341 |
elif category == 'Table':
|
342 |
-
# If text is already HTML, keep it as is
|
343 |
if text.strip().startswith('<'):
|
344 |
markdown_lines.append(f"{text}\n")
|
345 |
else:
|
346 |
markdown_lines.append(f"**Table:** {text}\n")
|
347 |
elif category == 'Formula':
|
348 |
-
# If text is LaTeX, format it properly
|
349 |
if text.strip().startswith('$') or '\\' in text:
|
350 |
markdown_lines.append(f"$$\n{text}\n$$\n")
|
351 |
else:
|
@@ -355,20 +254,15 @@ def layoutjson2md(image: Image.Image, layout_data: List[Dict], text_key: str = '
|
|
355 |
elif category == 'Footnote':
|
356 |
markdown_lines.append(f"^{text}^\n")
|
357 |
elif category in ['Page-header', 'Page-footer']:
|
358 |
-
# Skip headers and footers in main content
|
359 |
continue
|
360 |
else:
|
361 |
markdown_lines.append(f"{text}\n")
|
362 |
-
|
363 |
-
markdown_lines.append("") # Add spacing
|
364 |
-
|
365 |
except Exception as e:
|
366 |
print(f"Error converting to markdown: {e}")
|
367 |
return str(layout_data)
|
368 |
-
|
369 |
return "\n".join(markdown_lines)
|
370 |
|
371 |
-
|
372 |
# PDF handling state
|
373 |
pdf_cache = {
|
374 |
"images": [],
|
@@ -378,9 +272,9 @@ pdf_cache = {
|
|
378 |
"is_parsed": False,
|
379 |
"results": []
|
380 |
}
|
|
|
381 |
@spaces.GPU
|
382 |
-
def inference(model_name: str, image: Image.Image, prompt: str, max_new_tokens: int =
|
383 |
-
"""Run inference on an image with the given prompt using the selected model."""
|
384 |
try:
|
385 |
if model_name == "Camel-Doc-OCR-062825":
|
386 |
processor = processor_m
|
@@ -401,23 +295,14 @@ def inference(model_name: str, image: Image.Image, prompt: str, max_new_tokens:
|
|
401 |
{
|
402 |
"role": "user",
|
403 |
"content": [
|
404 |
-
{"type": "
|
405 |
-
{"type": "
|
406 |
]
|
407 |
}
|
408 |
]
|
409 |
-
|
410 |
-
text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
411 |
-
image_inputs, video_inputs = process_vision_info(messages)
|
412 |
-
|
413 |
-
inputs = processor(
|
414 |
-
text=[text],
|
415 |
-
images=[image],
|
416 |
-
videos=video_inputs,
|
417 |
-
padding=True,
|
418 |
-
return_tensors="pt"
|
419 |
-
).to(device)
|
420 |
|
|
|
|
|
421 |
|
422 |
with torch.no_grad():
|
423 |
generated_ids = model.generate(
|
@@ -427,34 +312,27 @@ def inference(model_name: str, image: Image.Image, prompt: str, max_new_tokens:
|
|
427 |
temperature=0.1
|
428 |
)
|
429 |
|
430 |
-
|
431 |
out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
|
432 |
]
|
433 |
-
output_text = processor.batch_decode(
|
434 |
-
return output_text
|
435 |
|
436 |
except Exception as e:
|
437 |
print(f"Error during inference: {e}")
|
438 |
traceback.print_exc()
|
439 |
return f"Error during inference: {str(e)}"
|
440 |
|
441 |
-
|
442 |
def process_image(
|
443 |
model_name: str,
|
444 |
image: Image.Image,
|
445 |
min_pixels: Optional[int] = None,
|
446 |
max_pixels: Optional[int] = None
|
447 |
) -> Dict[str, Any]:
|
448 |
-
"""Process a single image with the specified prompt mode"""
|
449 |
try:
|
450 |
-
|
451 |
-
if min_pixels is not None or max_pixels is not None:
|
452 |
image = fetch_image(image, min_pixels=min_pixels, max_pixels=max_pixels)
|
453 |
-
|
454 |
-
# Run inference with the default prompt
|
455 |
raw_output = inference(model_name, image, prompt)
|
456 |
-
|
457 |
-
# Process results based on prompt mode
|
458 |
result = {
|
459 |
'original_image': image,
|
460 |
'raw_output': raw_output,
|
@@ -462,42 +340,26 @@ def process_image(
|
|
462 |
'layout_result': None,
|
463 |
'markdown_content': None
|
464 |
}
|
465 |
-
|
466 |
-
# Try to parse JSON and create visualizations (since we're doing layout analysis)
|
467 |
try:
|
468 |
-
# Clean the output to be valid JSON
|
469 |
-
# Models sometimes add ```json ... ``` markers
|
470 |
json_match = re.search(r'```json\s*([\s\S]+?)\s*```', raw_output)
|
471 |
-
if json_match
|
472 |
-
json_str = json_match.group(1)
|
473 |
-
else:
|
474 |
-
json_str = raw_output
|
475 |
-
|
476 |
layout_data = json.loads(json_str)
|
477 |
result['layout_result'] = layout_data
|
478 |
-
|
479 |
-
# Create visualization with bounding boxes
|
480 |
try:
|
481 |
processed_image = draw_layout_on_image(image, layout_data)
|
482 |
result['processed_image'] = processed_image
|
483 |
except Exception as e:
|
484 |
print(f"Error drawing layout: {e}")
|
485 |
-
result['processed_image'] = image
|
486 |
-
|
487 |
-
# Generate markdown from layout data
|
488 |
try:
|
489 |
markdown_content = layoutjson2md(image, layout_data, text_key='text')
|
490 |
result['markdown_content'] = markdown_content
|
491 |
except Exception as e:
|
492 |
print(f"Error generating markdown: {e}")
|
493 |
result['markdown_content'] = raw_output
|
494 |
-
|
495 |
except json.JSONDecodeError:
|
496 |
print("Failed to parse JSON output, using raw output")
|
497 |
result['markdown_content'] = raw_output
|
498 |
-
|
499 |
return result
|
500 |
-
|
501 |
except Exception as e:
|
502 |
print(f"Error processing image: {e}")
|
503 |
traceback.print_exc()
|
@@ -509,24 +371,16 @@ def process_image(
|
|
509 |
'markdown_content': f"Error processing image: {str(e)}"
|
510 |
}
|
511 |
|
512 |
-
|
513 |
def load_file_for_preview(file_path: str) -> Tuple[Optional[Image.Image], str]:
|
514 |
-
"""Load file for preview (supports PDF and images)"""
|
515 |
global pdf_cache
|
516 |
-
|
517 |
if not file_path or not os.path.exists(file_path):
|
518 |
return None, "No file selected"
|
519 |
-
|
520 |
-
# FIX 1: Access the second element of the tuple returned by os.path.splitext
|
521 |
-
file_ext = os.path.splitext(file_path).lower()
|
522 |
-
|
523 |
try:
|
524 |
if file_ext == '.pdf':
|
525 |
-
# Load PDF pages
|
526 |
images = load_images_from_pdf(file_path)
|
527 |
if not images:
|
528 |
return None, "Failed to load PDF"
|
529 |
-
|
530 |
pdf_cache.update({
|
531 |
"images": images,
|
532 |
"current_page": 0,
|
@@ -535,14 +389,9 @@ def load_file_for_preview(file_path: str) -> Tuple[Optional[Image.Image], str]:
|
|
535 |
"is_parsed": False,
|
536 |
"results": []
|
537 |
})
|
538 |
-
|
539 |
-
# FIX 2: Return only the first image for the preview component
|
540 |
-
return images, f"Page 1 / {len(images)}"
|
541 |
-
|
542 |
elif file_ext in ['.jpg', '.jpeg', '.png', '.bmp', '.tiff']:
|
543 |
-
# Load single image
|
544 |
image = Image.open(file_path).convert('RGB')
|
545 |
-
|
546 |
pdf_cache.update({
|
547 |
"images": [image],
|
548 |
"current_page": 0,
|
@@ -551,73 +400,50 @@ def load_file_for_preview(file_path: str) -> Tuple[Optional[Image.Image], str]:
|
|
551 |
"is_parsed": False,
|
552 |
"results": []
|
553 |
})
|
554 |
-
|
555 |
return image, "Page 1 / 1"
|
556 |
else:
|
557 |
return None, f"Unsupported file format: {file_ext}"
|
558 |
-
|
559 |
except Exception as e:
|
560 |
print(f"Error loading file: {e}")
|
561 |
return None, f"Error loading file: {str(e)}"
|
562 |
|
563 |
-
|
564 |
def turn_page(direction: str) -> Tuple[Optional[Image.Image], str, Any, Optional[Image.Image], Optional[Dict]]:
|
565 |
-
"""Navigate through PDF pages and update all relevant outputs."""
|
566 |
global pdf_cache
|
567 |
-
|
568 |
if not pdf_cache["images"]:
|
569 |
return None, '<div class="page-info">No file loaded</div>', "No results yet", None, None
|
570 |
-
|
571 |
if direction == "prev":
|
572 |
pdf_cache["current_page"] = max(0, pdf_cache["current_page"] - 1)
|
573 |
elif direction == "next":
|
574 |
-
pdf_cache["current_page"] = min(
|
575 |
-
pdf_cache["total_pages"] - 1,
|
576 |
-
pdf_cache["current_page"] + 1
|
577 |
-
)
|
578 |
-
|
579 |
index = pdf_cache["current_page"]
|
580 |
current_image_preview = pdf_cache["images"][index]
|
581 |
page_info_html = f'<div class="page-info">Page {index + 1} / {pdf_cache["total_pages"]}</div>'
|
582 |
-
|
583 |
-
# Initialize default result values
|
584 |
markdown_content = "Page not processed yet"
|
585 |
processed_img = None
|
586 |
layout_json = None
|
587 |
-
|
588 |
-
# Get results for current page if available
|
589 |
-
if (pdf_cache["is_parsed"] and
|
590 |
-
index < len(pdf_cache["results"]) and
|
591 |
-
pdf_cache["results"][index]):
|
592 |
-
|
593 |
result = pdf_cache["results"][index]
|
594 |
markdown_content = result.get('markdown_content') or result.get('raw_output', 'No content available')
|
595 |
-
processed_img = result.get('processed_image', None)
|
596 |
-
layout_json = result.get('layout_result', None)
|
597 |
-
|
598 |
-
# Check for Arabic text to set RTL property
|
599 |
if is_arabic_text(markdown_content):
|
600 |
markdown_update = gr.update(value=markdown_content, rtl=True)
|
601 |
else:
|
602 |
markdown_update = markdown_content
|
603 |
-
|
604 |
return current_image_preview, page_info_html, markdown_update, processed_img, layout_json
|
605 |
|
606 |
-
|
607 |
def create_gradio_interface():
|
608 |
-
"""Create the Gradio interface"""
|
609 |
-
|
610 |
css = """
|
611 |
.main-container { max-width: 1400px; margin: 0 auto; }
|
612 |
.header-text { text-align: center; color: #2c3e50; margin-bottom: 20px; }
|
613 |
-
.process-button {
|
614 |
-
border: none !important;
|
615 |
-
color: white !important;
|
616 |
-
font-weight: bold !important;
|
617 |
-
background-color: blue !important;}
|
618 |
-
.process-button:hover {
|
619 |
background-color: darkblue !important;
|
620 |
-
transform: translateY(-2px) !important;
|
621 |
box-shadow: 0 4px 8px rgba(0,0,0,0.2) !important; }
|
622 |
.info-box { border: 1px solid #dee2e6; border-radius: 8px; padding: 15px; margin: 10px 0; }
|
623 |
.page-info { text-align: center; padding: 8px 16px; border-radius: 20px; font-weight: bold; margin: 10px 0; }
|
@@ -633,248 +459,91 @@ def create_gradio_interface():
|
|
633 |
</p>
|
634 |
</div>
|
635 |
""")
|
636 |
-
|
637 |
-
# Main interface
|
638 |
with gr.Row():
|
639 |
-
# Left column - Input and controls
|
640 |
with gr.Column(scale=1):
|
641 |
-
|
642 |
-
# Model selection
|
643 |
-
model_choice = gr.Dropdown(
|
644 |
choices=["Camel-Doc-OCR-062825", "MonkeyOCR-Recognition", "Nanonets-OCR-s", "Megalodon-OCR-Sync-0713"],
|
645 |
label="Select Model",
|
646 |
value="Camel-Doc-OCR-062825"
|
647 |
)
|
648 |
-
|
649 |
-
# File input
|
650 |
file_input = gr.File(
|
651 |
label="Upload Image or PDF",
|
652 |
file_types=[".jpg", ".jpeg", ".png", ".bmp", ".tiff", ".pdf"],
|
653 |
type="filepath"
|
654 |
)
|
655 |
-
|
656 |
-
# Image preview
|
657 |
-
image_preview = gr.Image(
|
658 |
-
label="Preview",
|
659 |
-
type="pil",
|
660 |
-
interactive=False,
|
661 |
-
height=300
|
662 |
-
)
|
663 |
-
|
664 |
-
# Page navigation for PDFs
|
665 |
with gr.Row():
|
666 |
prev_page_btn = gr.Button("◀ Previous", size="md")
|
667 |
page_info = gr.HTML('<div class="page-info">No file loaded</div>')
|
668 |
next_page_btn = gr.Button("Next ▶", size="md")
|
669 |
-
|
670 |
-
# Advanced settings
|
671 |
with gr.Accordion("Advanced Settings", open=False):
|
672 |
-
max_new_tokens = gr.Slider(
|
673 |
-
|
674 |
-
|
675 |
-
|
676 |
-
step=1000,
|
677 |
-
label="Max New Tokens",
|
678 |
-
info="Maximum number of tokens to generate"
|
679 |
-
)
|
680 |
-
|
681 |
-
min_pixels = gr.Number(
|
682 |
-
value=MIN_PIXELS,
|
683 |
-
label="Min Pixels",
|
684 |
-
info="Minimum image resolution"
|
685 |
-
)
|
686 |
-
|
687 |
-
max_pixels = gr.Number(
|
688 |
-
value=MAX_PIXELS,
|
689 |
-
label="Max Pixels",
|
690 |
-
info="Maximum image resolution"
|
691 |
-
)
|
692 |
-
|
693 |
-
# Process button
|
694 |
-
process_btn = gr.Button(
|
695 |
-
"🚀 Process Document",
|
696 |
-
variant="primary",
|
697 |
-
elem_classes=["process-button"],
|
698 |
-
size="lg"
|
699 |
-
)
|
700 |
-
|
701 |
-
# Clear button
|
702 |
clear_btn = gr.Button("🗑️ Clear All", variant="secondary")
|
703 |
-
|
704 |
-
# Right column - Results
|
705 |
with gr.Column(scale=2):
|
706 |
-
|
707 |
-
# Results tabs
|
708 |
with gr.Tabs():
|
709 |
-
# Processed image tab
|
710 |
with gr.Tab("🖼️ Processed Image"):
|
711 |
-
processed_image = gr.Image(
|
712 |
-
label="Image with Layout Detection",
|
713 |
-
type="pil",
|
714 |
-
interactive=False,
|
715 |
-
height=500
|
716 |
-
)
|
717 |
-
# Markdown output tab
|
718 |
with gr.Tab("📝 Extracted Content"):
|
719 |
-
markdown_output = gr.Markdown(
|
720 |
-
value="Click 'Process Document' to see extracted content...",
|
721 |
-
height=500
|
722 |
-
)
|
723 |
-
# JSON layout tab
|
724 |
with gr.Tab("📋 Layout JSON"):
|
725 |
-
json_output = gr.JSON(
|
726 |
-
label="Layout Analysis Results",
|
727 |
-
value=None
|
728 |
-
)
|
729 |
-
|
730 |
-
# Event handlers
|
731 |
def process_document(model_name, file_path, max_tokens, min_pix, max_pix):
|
732 |
-
"""Process the uploaded document"""
|
733 |
global pdf_cache
|
734 |
-
|
735 |
try:
|
736 |
if not file_path:
|
737 |
return None, "Please upload a file first.", None
|
738 |
-
|
739 |
-
# This function now correctly returns a single image for preview
|
740 |
-
# and populates the cache for multi-page processing.
|
741 |
-
preview_img, page_info_str = load_file_for_preview(file_path)
|
742 |
-
if preview_img is None:
|
743 |
-
return None, page_info_str, None
|
744 |
-
|
745 |
-
# Process the image(s)
|
746 |
if pdf_cache["file_type"] == "pdf":
|
747 |
-
# Process all pages for PDF from the cache
|
748 |
all_results = []
|
749 |
all_markdown = []
|
750 |
-
|
751 |
for i, img in enumerate(pdf_cache["images"]):
|
752 |
-
result = process_image(
|
753 |
-
model_name,
|
754 |
-
img,
|
755 |
-
min_pixels=int(min_pix) if min_pix else None,
|
756 |
-
max_pixels=int(max_pix) if max_pix else None
|
757 |
-
)
|
758 |
all_results.append(result)
|
759 |
if result.get('markdown_content'):
|
760 |
all_markdown.append(f"## Page {i+1}\n\n{result['markdown_content']}")
|
761 |
-
|
762 |
pdf_cache["results"] = all_results
|
763 |
pdf_cache["is_parsed"] = True
|
764 |
-
|
765 |
-
# Show results for first page
|
766 |
-
first_result = all_results
|
767 |
combined_markdown = "\n\n---\n\n".join(all_markdown)
|
768 |
-
|
769 |
-
# Check if the combined markdown contains mostly Arabic text
|
770 |
if is_arabic_text(combined_markdown):
|
771 |
markdown_update = gr.update(value=combined_markdown, rtl=True)
|
772 |
else:
|
773 |
markdown_update = combined_markdown
|
774 |
-
|
775 |
-
return (
|
776 |
-
first_result['processed_image'],
|
777 |
-
markdown_update,
|
778 |
-
first_result['layout_result']
|
779 |
-
)
|
780 |
else:
|
781 |
-
|
782 |
-
result = process_image(
|
783 |
-
model_name,
|
784 |
-
preview_img, # Use the single loaded image
|
785 |
-
min_pixels=int(min_pix) if min_pix else None,
|
786 |
-
max_pixels=int(max_pix) if max_pix else None
|
787 |
-
)
|
788 |
-
|
789 |
pdf_cache["results"] = [result]
|
790 |
pdf_cache["is_parsed"] = True
|
791 |
-
|
792 |
-
# Check if the content contains mostly Arabic text
|
793 |
content = result['markdown_content'] or "No content extracted"
|
794 |
if is_arabic_text(content):
|
795 |
markdown_update = gr.update(value=content, rtl=True)
|
796 |
else:
|
797 |
markdown_update = content
|
798 |
-
|
799 |
-
return (
|
800 |
-
result['processed_image'],
|
801 |
-
markdown_update,
|
802 |
-
result['layout_result']
|
803 |
-
)
|
804 |
-
|
805 |
except Exception as e:
|
806 |
error_msg = f"Error processing document: {str(e)}"
|
807 |
print(error_msg)
|
808 |
traceback.print_exc()
|
809 |
return None, error_msg, None
|
810 |
-
|
811 |
def handle_file_upload(file_path):
|
812 |
-
"""Handle file upload and show preview"""
|
813 |
if not file_path:
|
814 |
-
return None,
|
815 |
-
|
816 |
image, page_info = load_file_for_preview(file_path)
|
817 |
return image, page_info
|
818 |
-
|
819 |
def clear_all():
|
820 |
-
"""Clear all data and reset interface"""
|
821 |
global pdf_cache
|
822 |
-
|
823 |
-
|
824 |
-
|
825 |
-
|
826 |
-
|
827 |
-
|
828 |
-
|
829 |
-
None, # file_input
|
830 |
-
None, # image_preview
|
831 |
-
'<div class="page-info">No file loaded</div>', # page_info
|
832 |
-
None, # processed_image
|
833 |
-
"Click 'Process Document' to see extracted content...", # markdown_output
|
834 |
-
None, # json_output
|
835 |
-
)
|
836 |
-
|
837 |
-
# Wire up event handlers
|
838 |
-
file_input.change(
|
839 |
-
handle_file_upload,
|
840 |
-
inputs=[file_input],
|
841 |
-
outputs=[image_preview, page_info]
|
842 |
-
)
|
843 |
-
|
844 |
-
prev_page_btn.click(
|
845 |
-
lambda: turn_page("prev"),
|
846 |
-
outputs=[image_preview, page_info, markdown_output, processed_image, json_output]
|
847 |
-
)
|
848 |
-
|
849 |
-
next_page_btn.click(
|
850 |
-
lambda: turn_page("next"),
|
851 |
-
outputs=[image_preview, page_info, markdown_output, processed_image, json_output]
|
852 |
-
)
|
853 |
-
|
854 |
-
process_btn.click(
|
855 |
-
process_document,
|
856 |
-
inputs=[model_choice, file_input, max_new_tokens, min_pixels, max_pixels],
|
857 |
-
outputs=[processed_image, markdown_output, json_output]
|
858 |
-
)
|
859 |
-
|
860 |
-
clear_btn.click(
|
861 |
-
clear_all,
|
862 |
-
outputs=[
|
863 |
-
file_input, image_preview, page_info, processed_image,
|
864 |
-
markdown_output, json_output
|
865 |
-
]
|
866 |
-
)
|
867 |
-
|
868 |
return demo
|
869 |
|
870 |
-
|
871 |
if __name__ == "__main__":
|
872 |
-
# Create and launch the interface
|
873 |
demo = create_gradio_interface()
|
874 |
-
demo.queue(max_size=10).launch(
|
875 |
-
server_name="0.0.0.0",
|
876 |
-
server_port=7860,
|
877 |
-
share=False,
|
878 |
-
debug=True,
|
879 |
-
show_error=True
|
880 |
-
)
|
|
|
11 |
import gradio as gr
|
12 |
import requests
|
13 |
import torch
|
|
|
14 |
from PIL import Image, ImageDraw, ImageFont
|
15 |
from transformers import (
|
16 |
Qwen2_5_VLForConditionalGeneration,
|
|
|
45 |
5. Final Output: The entire output must be a single JSON object.
|
46 |
"""
|
47 |
|
48 |
+
# Load models
|
49 |
MODEL_ID_M = "prithivMLmods/Camel-Doc-OCR-062825"
|
50 |
processor_m = AutoProcessor.from_pretrained(MODEL_ID_M, trust_remote_code=True)
|
51 |
model_m = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
|
|
54 |
torch_dtype=torch.float16
|
55 |
).to(device).eval()
|
56 |
|
|
|
57 |
MODEL_ID_T = "prithivMLmods/Megalodon-OCR-Sync-0713"
|
58 |
processor_t = AutoProcessor.from_pretrained(MODEL_ID_T, trust_remote_code=True)
|
59 |
model_t = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
|
|
62 |
torch_dtype=torch.float16
|
63 |
).to(device).eval()
|
64 |
|
|
|
65 |
MODEL_ID_C = "nanonets/Nanonets-OCR-s"
|
66 |
processor_c = AutoProcessor.from_pretrained(MODEL_ID_C, trust_remote_code=True)
|
67 |
model_c = Qwen2_5_VLForConditionalGeneration.from_pretrained(
|
|
|
70 |
torch_dtype=torch.float16
|
71 |
).to(device).eval()
|
72 |
|
|
|
73 |
MODEL_ID_G = "echo840/MonkeyOCR"
|
74 |
SUBFOLDER = "Recognition"
|
75 |
processor_g = AutoProcessor.from_pretrained(
|
|
|
84 |
torch_dtype=torch.float16
|
85 |
).to(device).eval()
|
86 |
|
|
|
87 |
# Utility functions
|
88 |
def round_by_factor(number: int, factor: int) -> int:
|
|
|
89 |
return round(number / factor) * factor
|
90 |
|
|
|
91 |
def smart_resize(
|
92 |
height: int,
|
93 |
width: int,
|
|
|
95 |
min_pixels: int = 3136,
|
96 |
max_pixels: int = 11289600,
|
97 |
):
|
|
|
|
|
|
|
|
|
|
|
98 |
if max(height, width) / min(height, width) > 200:
|
99 |
+
raise ValueError(f"Aspect ratio too extreme: {max(height, width) / min(height, width)}")
|
|
|
|
|
100 |
h_bar = max(factor, round_by_factor(height, factor))
|
101 |
w_bar = max(factor, round_by_factor(width, factor))
|
|
|
102 |
if h_bar * w_bar > max_pixels:
|
103 |
beta = math.sqrt((height * width) / max_pixels)
|
104 |
h_bar = round_by_factor(height / beta, factor)
|
|
|
109 |
w_bar = round_by_factor(width * beta, factor)
|
110 |
return h_bar, w_bar
|
111 |
|
|
|
112 |
def fetch_image(image_input, min_pixels: int = None, max_pixels: int = None):
|
|
|
113 |
if isinstance(image_input, str):
|
114 |
if image_input.startswith(("http://", "https://")):
|
115 |
response = requests.get(image_input)
|
|
|
120 |
image = image_input.convert('RGB')
|
121 |
else:
|
122 |
raise ValueError(f"Invalid image input type: {type(image_input)}")
|
123 |
+
if min_pixels or max_pixels:
|
|
|
124 |
min_pixels = min_pixels or MIN_PIXELS
|
125 |
max_pixels = max_pixels or MAX_PIXELS
|
126 |
+
height, width = smart_resize(image.height, image.width, factor=IMAGE_FACTOR, min_pixels=min_pixels, max_pixels=max_pixels)
|
|
|
|
|
|
|
|
|
|
|
|
|
127 |
image = image.resize((width, height), Image.LANCZOS)
|
|
|
128 |
return image
|
129 |
|
|
|
130 |
def load_images_from_pdf(pdf_path: str) -> List[Image.Image]:
|
|
|
131 |
images = []
|
132 |
try:
|
133 |
pdf_document = fitz.open(pdf_path)
|
134 |
for page_num in range(len(pdf_document)):
|
135 |
page = pdf_document.load_page(page_num)
|
136 |
+
mat = fitz.Matrix(2.0, 2.0)
|
|
|
137 |
pix = page.get_pixmap(matrix=mat)
|
138 |
img_data = pix.tobytes("ppm")
|
139 |
image = Image.open(BytesIO(img_data)).convert('RGB')
|
|
|
141 |
pdf_document.close()
|
142 |
except Exception as e:
|
143 |
print(f"Error loading PDF: {e}")
|
|
|
144 |
return images
|
145 |
|
|
|
146 |
def draw_layout_on_image(image: Image.Image, layout_data: List[Dict]) -> Image.Image:
|
|
|
147 |
img_copy = image.copy()
|
148 |
draw = ImageDraw.Draw(img_copy)
|
|
|
|
|
149 |
colors = {
|
150 |
+
'Caption': '#FF6B6B', 'Footnote': '#4ECDC4', 'Formula': '#45B7D1',
|
151 |
+
'List-item': '#96CEB4', 'Page-footer': '#FFEAA7', 'Page-header': '#DDA0DD',
|
152 |
+
'Picture': '#FFD93D', 'Section-header': '#6C5CE7', 'Table': '#FD79A8',
|
153 |
+
'Text': '#74B9FF', 'Title': '#E17055'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
154 |
}
|
|
|
155 |
try:
|
156 |
+
font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf", 12)
|
157 |
+
except Exception:
|
158 |
+
font = ImageFont.load_default()
|
159 |
+
for item in layout_data:
|
160 |
+
if 'bbox' in item and 'category' in item:
|
161 |
+
bbox = item['bbox']
|
162 |
+
category = item['category']
|
163 |
+
color = colors.get(category, '#000000')
|
164 |
+
draw.rectangle(bbox, outline=color, width=2)
|
165 |
+
label = category
|
166 |
+
label_bbox = draw.textbbox((0, 0), label, font=font)
|
167 |
+
label_width = label_bbox[2] - label_bbox[0]
|
168 |
+
label_height = label_bbox[3] - label_bbox[1]
|
169 |
+
label_x = bbox[0]
|
170 |
+
label_y = max(0, bbox[1] - label_height - 2)
|
171 |
+
draw.rectangle([label_x, label_y, label_x + label_width + 4, label_y + label_height + 2], fill=color)
|
172 |
+
draw.text((label_x + 2, label_y + 1), label, fill='white', font=font)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
173 |
return img_copy
|
174 |
|
|
|
175 |
def is_arabic_text(text: str) -> bool:
|
|
|
176 |
if not text:
|
177 |
return False
|
|
|
|
|
|
|
178 |
header_pattern = r'^#{1,6}\s+(.+)$'
|
179 |
paragraph_pattern = r'^(?!#{1,6}\s|!\[|```|\||\s*[-*+]\s|\s*\d+\.\s)(.+)$'
|
|
|
180 |
content_text = []
|
|
|
181 |
for line in text.split('\n'):
|
182 |
line = line.strip()
|
183 |
if not line:
|
184 |
continue
|
|
|
|
|
185 |
header_match = re.match(header_pattern, line, re.MULTILINE)
|
186 |
if header_match:
|
187 |
content_text.append(header_match.group(1))
|
188 |
continue
|
|
|
|
|
189 |
if re.match(paragraph_pattern, line, re.MULTILINE):
|
190 |
content_text.append(line)
|
|
|
191 |
if not content_text:
|
192 |
return False
|
|
|
|
|
193 |
combined_text = ' '.join(content_text)
|
|
|
|
|
194 |
arabic_chars = 0
|
195 |
total_chars = 0
|
|
|
196 |
for char in combined_text:
|
197 |
if char.isalpha():
|
198 |
total_chars += 1
|
|
|
199 |
if ('\u0600' <= char <= '\u06FF') or ('\u0750' <= char <= '\u077F') or ('\u08A0' <= char <= '\u08FF'):
|
200 |
arabic_chars += 1
|
201 |
+
return total_chars > 0 and (arabic_chars / total_chars) > 0.5
|
|
|
|
|
|
|
|
|
|
|
|
|
202 |
|
203 |
def layoutjson2md(image: Image.Image, layout_data: List[Dict], text_key: str = 'text') -> str:
|
|
|
204 |
import base64
|
205 |
from io import BytesIO
|
|
|
206 |
markdown_lines = []
|
|
|
207 |
try:
|
208 |
+
sorted_items = sorted(layout_data, key=lambda x: (x.get('bbox', [0, 0, 0, 0])[1], x.get('bbox', [0, 0, 0, 0])[0]))
|
|
|
|
|
209 |
for item in sorted_items:
|
210 |
category = item.get('category', '')
|
211 |
text = item.get(text_key, '')
|
212 |
bbox = item.get('bbox', [])
|
|
|
213 |
if category == 'Picture':
|
|
|
214 |
if bbox and len(bbox) == 4:
|
215 |
try:
|
|
|
216 |
x1, y1, x2, y2 = bbox
|
|
|
217 |
x1, y1 = max(0, int(x1)), max(0, int(y1))
|
218 |
x2, y2 = min(image.width, int(x2)), min(image.height, int(y2))
|
|
|
219 |
if x2 > x1 and y2 > y1:
|
220 |
cropped_img = image.crop((x1, y1, x2, y2))
|
|
|
|
|
221 |
buffer = BytesIO()
|
222 |
cropped_img.save(buffer, format='PNG')
|
223 |
img_data = base64.b64encode(buffer.getvalue()).decode()
|
|
|
|
|
224 |
markdown_lines.append(f"\n")
|
225 |
else:
|
226 |
markdown_lines.append("\n")
|
|
|
240 |
elif category == 'List-item':
|
241 |
markdown_lines.append(f"- {text}\n")
|
242 |
elif category == 'Table':
|
|
|
243 |
if text.strip().startswith('<'):
|
244 |
markdown_lines.append(f"{text}\n")
|
245 |
else:
|
246 |
markdown_lines.append(f"**Table:** {text}\n")
|
247 |
elif category == 'Formula':
|
|
|
248 |
if text.strip().startswith('$') or '\\' in text:
|
249 |
markdown_lines.append(f"$$\n{text}\n$$\n")
|
250 |
else:
|
|
|
254 |
elif category == 'Footnote':
|
255 |
markdown_lines.append(f"^{text}^\n")
|
256 |
elif category in ['Page-header', 'Page-footer']:
|
|
|
257 |
continue
|
258 |
else:
|
259 |
markdown_lines.append(f"{text}\n")
|
260 |
+
markdown_lines.append("")
|
|
|
|
|
261 |
except Exception as e:
|
262 |
print(f"Error converting to markdown: {e}")
|
263 |
return str(layout_data)
|
|
|
264 |
return "\n".join(markdown_lines)
|
265 |
|
|
|
266 |
# PDF handling state
|
267 |
pdf_cache = {
|
268 |
"images": [],
|
|
|
272 |
"is_parsed": False,
|
273 |
"results": []
|
274 |
}
|
275 |
+
|
276 |
@spaces.GPU
|
277 |
+
def inference(model_name: str, image: Image.Image, prompt: str, max_new_tokens: int = 1024) -> str:
|
|
|
278 |
try:
|
279 |
if model_name == "Camel-Doc-OCR-062825":
|
280 |
processor = processor_m
|
|
|
295 |
{
|
296 |
"role": "user",
|
297 |
"content": [
|
298 |
+
{"type": "image", "image": image},
|
299 |
+
{"type": "text", "text": prompt}
|
300 |
]
|
301 |
}
|
302 |
]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
303 |
|
304 |
+
text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
305 |
+
inputs = processor(text=[text], images=[image], return_tensors="pt").to(device)
|
306 |
|
307 |
with torch.no_grad():
|
308 |
generated_ids = model.generate(
|
|
|
312 |
temperature=0.1
|
313 |
)
|
314 |
|
315 |
+
generated_ids_trimmed = [
|
316 |
out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
|
317 |
]
|
318 |
+
output_text = processor.batch_decode(generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False)
|
319 |
+
return output_text[0] if output_text else ""
|
320 |
|
321 |
except Exception as e:
|
322 |
print(f"Error during inference: {e}")
|
323 |
traceback.print_exc()
|
324 |
return f"Error during inference: {str(e)}"
|
325 |
|
|
|
326 |
def process_image(
|
327 |
model_name: str,
|
328 |
image: Image.Image,
|
329 |
min_pixels: Optional[int] = None,
|
330 |
max_pixels: Optional[int] = None
|
331 |
) -> Dict[str, Any]:
|
|
|
332 |
try:
|
333 |
+
if min_pixels or max_pixels:
|
|
|
334 |
image = fetch_image(image, min_pixels=min_pixels, max_pixels=max_pixels)
|
|
|
|
|
335 |
raw_output = inference(model_name, image, prompt)
|
|
|
|
|
336 |
result = {
|
337 |
'original_image': image,
|
338 |
'raw_output': raw_output,
|
|
|
340 |
'layout_result': None,
|
341 |
'markdown_content': None
|
342 |
}
|
|
|
|
|
343 |
try:
|
|
|
|
|
344 |
json_match = re.search(r'```json\s*([\s\S]+?)\s*```', raw_output)
|
345 |
+
json_str = json_match.group(1) if json_match else raw_output
|
|
|
|
|
|
|
|
|
346 |
layout_data = json.loads(json_str)
|
347 |
result['layout_result'] = layout_data
|
|
|
|
|
348 |
try:
|
349 |
processed_image = draw_layout_on_image(image, layout_data)
|
350 |
result['processed_image'] = processed_image
|
351 |
except Exception as e:
|
352 |
print(f"Error drawing layout: {e}")
|
|
|
|
|
|
|
353 |
try:
|
354 |
markdown_content = layoutjson2md(image, layout_data, text_key='text')
|
355 |
result['markdown_content'] = markdown_content
|
356 |
except Exception as e:
|
357 |
print(f"Error generating markdown: {e}")
|
358 |
result['markdown_content'] = raw_output
|
|
|
359 |
except json.JSONDecodeError:
|
360 |
print("Failed to parse JSON output, using raw output")
|
361 |
result['markdown_content'] = raw_output
|
|
|
362 |
return result
|
|
|
363 |
except Exception as e:
|
364 |
print(f"Error processing image: {e}")
|
365 |
traceback.print_exc()
|
|
|
371 |
'markdown_content': f"Error processing image: {str(e)}"
|
372 |
}
|
373 |
|
|
|
374 |
def load_file_for_preview(file_path: str) -> Tuple[Optional[Image.Image], str]:
|
|
|
375 |
global pdf_cache
|
|
|
376 |
if not file_path or not os.path.exists(file_path):
|
377 |
return None, "No file selected"
|
378 |
+
file_ext = os.path.splitext(file_path)[1].lower()
|
|
|
|
|
|
|
379 |
try:
|
380 |
if file_ext == '.pdf':
|
|
|
381 |
images = load_images_from_pdf(file_path)
|
382 |
if not images:
|
383 |
return None, "Failed to load PDF"
|
|
|
384 |
pdf_cache.update({
|
385 |
"images": images,
|
386 |
"current_page": 0,
|
|
|
389 |
"is_parsed": False,
|
390 |
"results": []
|
391 |
})
|
392 |
+
return images[0], f"Page 1 / {len(images)}"
|
|
|
|
|
|
|
393 |
elif file_ext in ['.jpg', '.jpeg', '.png', '.bmp', '.tiff']:
|
|
|
394 |
image = Image.open(file_path).convert('RGB')
|
|
|
395 |
pdf_cache.update({
|
396 |
"images": [image],
|
397 |
"current_page": 0,
|
|
|
400 |
"is_parsed": False,
|
401 |
"results": []
|
402 |
})
|
|
|
403 |
return image, "Page 1 / 1"
|
404 |
else:
|
405 |
return None, f"Unsupported file format: {file_ext}"
|
|
|
406 |
except Exception as e:
|
407 |
print(f"Error loading file: {e}")
|
408 |
return None, f"Error loading file: {str(e)}"
|
409 |
|
|
|
410 |
def turn_page(direction: str) -> Tuple[Optional[Image.Image], str, Any, Optional[Image.Image], Optional[Dict]]:
|
|
|
411 |
global pdf_cache
|
|
|
412 |
if not pdf_cache["images"]:
|
413 |
return None, '<div class="page-info">No file loaded</div>', "No results yet", None, None
|
|
|
414 |
if direction == "prev":
|
415 |
pdf_cache["current_page"] = max(0, pdf_cache["current_page"] - 1)
|
416 |
elif direction == "next":
|
417 |
+
pdf_cache["current_page"] = min(pdf_cache["total_pages"] - 1, pdf_cache["current_page"] + 1)
|
|
|
|
|
|
|
|
|
418 |
index = pdf_cache["current_page"]
|
419 |
current_image_preview = pdf_cache["images"][index]
|
420 |
page_info_html = f'<div class="page-info">Page {index + 1} / {pdf_cache["total_pages"]}</div>'
|
|
|
|
|
421 |
markdown_content = "Page not processed yet"
|
422 |
processed_img = None
|
423 |
layout_json = None
|
424 |
+
if pdf_cache["is_parsed"] and index < len(pdf_cache["results"]) and pdf_cache["results"][index]:
|
|
|
|
|
|
|
|
|
|
|
425 |
result = pdf_cache["results"][index]
|
426 |
markdown_content = result.get('markdown_content') or result.get('raw_output', 'No content available')
|
427 |
+
processed_img = result.get('processed_image', None)
|
428 |
+
layout_json = result.get('layout_result', None)
|
|
|
|
|
429 |
if is_arabic_text(markdown_content):
|
430 |
markdown_update = gr.update(value=markdown_content, rtl=True)
|
431 |
else:
|
432 |
markdown_update = markdown_content
|
|
|
433 |
return current_image_preview, page_info_html, markdown_update, processed_img, layout_json
|
434 |
|
|
|
435 |
def create_gradio_interface():
|
|
|
|
|
436 |
css = """
|
437 |
.main-container { max-width: 1400px; margin: 0 auto; }
|
438 |
.header-text { text-align: center; color: #2c3e50; margin-bottom: 20px; }
|
439 |
+
.process-button {
|
440 |
+
border: none !important;
|
441 |
+
color: white !important;
|
442 |
+
font-weight: bold !important;
|
443 |
+
background-color: blue !important;}
|
444 |
+
.process-button:hover {
|
445 |
background-color: darkblue !important;
|
446 |
+
transform: translateY(-2px) !important;
|
447 |
box-shadow: 0 4px 8px rgba(0,0,0,0.2) !important; }
|
448 |
.info-box { border: 1px solid #dee2e6; border-radius: 8px; padding: 15px; margin: 10px 0; }
|
449 |
.page-info { text-align: center; padding: 8px 16px; border-radius: 20px; font-weight: bold; margin: 10px 0; }
|
|
|
459 |
</p>
|
460 |
</div>
|
461 |
""")
|
|
|
|
|
462 |
with gr.Row():
|
|
|
463 |
with gr.Column(scale=1):
|
464 |
+
model_choice = gr.Radio(
|
|
|
|
|
465 |
choices=["Camel-Doc-OCR-062825", "MonkeyOCR-Recognition", "Nanonets-OCR-s", "Megalodon-OCR-Sync-0713"],
|
466 |
label="Select Model",
|
467 |
value="Camel-Doc-OCR-062825"
|
468 |
)
|
|
|
|
|
469 |
file_input = gr.File(
|
470 |
label="Upload Image or PDF",
|
471 |
file_types=[".jpg", ".jpeg", ".png", ".bmp", ".tiff", ".pdf"],
|
472 |
type="filepath"
|
473 |
)
|
474 |
+
image_preview = gr.Image(label="Preview", type="pil", interactive=False, height=300)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
475 |
with gr.Row():
|
476 |
prev_page_btn = gr.Button("◀ Previous", size="md")
|
477 |
page_info = gr.HTML('<div class="page-info">No file loaded</div>')
|
478 |
next_page_btn = gr.Button("Next ▶", size="md")
|
|
|
|
|
479 |
with gr.Accordion("Advanced Settings", open=False):
|
480 |
+
max_new_tokens = gr.Slider(minimum=1000, maximum=32000, value=24000, step=1000, label="Max New Tokens")
|
481 |
+
min_pixels = gr.Number(value=MIN_PIXELS, label="Min Pixels")
|
482 |
+
max_pixels = gr.Number(value=MAX_PIXELS, label="Max Pixels")
|
483 |
+
process_btn = gr.Button("🚀 Process Document", variant="primary", elem_classes=["process-button"], size="lg")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
484 |
clear_btn = gr.Button("🗑️ Clear All", variant="secondary")
|
|
|
|
|
485 |
with gr.Column(scale=2):
|
|
|
|
|
486 |
with gr.Tabs():
|
|
|
487 |
with gr.Tab("🖼️ Processed Image"):
|
488 |
+
processed_image = gr.Image(label="Image with Layout Detection", type="pil", interactive=False, height=500)
|
|
|
|
|
|
|
|
|
|
|
|
|
489 |
with gr.Tab("📝 Extracted Content"):
|
490 |
+
markdown_output = gr.Markdown(value="Click 'Process Document' to see extracted content...", height=500)
|
|
|
|
|
|
|
|
|
491 |
with gr.Tab("📋 Layout JSON"):
|
492 |
+
json_output = gr.JSON(label="Layout Analysis Results", value=None)
|
|
|
|
|
|
|
|
|
|
|
493 |
def process_document(model_name, file_path, max_tokens, min_pix, max_pix):
|
|
|
494 |
global pdf_cache
|
|
|
495 |
try:
|
496 |
if not file_path:
|
497 |
return None, "Please upload a file first.", None
|
498 |
+
load_file_for_preview(file_path)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
499 |
if pdf_cache["file_type"] == "pdf":
|
|
|
500 |
all_results = []
|
501 |
all_markdown = []
|
|
|
502 |
for i, img in enumerate(pdf_cache["images"]):
|
503 |
+
result = process_image(model_name, img, min_pixels=int(min_pix) if min_pix else None, max_pixels=int(max_pix) if max_pix else None)
|
|
|
|
|
|
|
|
|
|
|
504 |
all_results.append(result)
|
505 |
if result.get('markdown_content'):
|
506 |
all_markdown.append(f"## Page {i+1}\n\n{result['markdown_content']}")
|
|
|
507 |
pdf_cache["results"] = all_results
|
508 |
pdf_cache["is_parsed"] = True
|
509 |
+
first_result = all_results[0]
|
|
|
|
|
510 |
combined_markdown = "\n\n---\n\n".join(all_markdown)
|
|
|
|
|
511 |
if is_arabic_text(combined_markdown):
|
512 |
markdown_update = gr.update(value=combined_markdown, rtl=True)
|
513 |
else:
|
514 |
markdown_update = combined_markdown
|
515 |
+
return first_result['processed_image'], markdown_update, first_result['layout_result']
|
|
|
|
|
|
|
|
|
|
|
516 |
else:
|
517 |
+
result = process_image(model_name, pdf_cache["images"][0], min_pixels=int(min_pix) if min_pix else None, max_pixels=int(max_pix) if max_pix else None)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
518 |
pdf_cache["results"] = [result]
|
519 |
pdf_cache["is_parsed"] = True
|
|
|
|
|
520 |
content = result['markdown_content'] or "No content extracted"
|
521 |
if is_arabic_text(content):
|
522 |
markdown_update = gr.update(value=content, rtl=True)
|
523 |
else:
|
524 |
markdown_update = content
|
525 |
+
return result['processed_image'], markdown_update, result['layout_result']
|
|
|
|
|
|
|
|
|
|
|
|
|
526 |
except Exception as e:
|
527 |
error_msg = f"Error processing document: {str(e)}"
|
528 |
print(error_msg)
|
529 |
traceback.print_exc()
|
530 |
return None, error_msg, None
|
|
|
531 |
def handle_file_upload(file_path):
|
|
|
532 |
if not file_path:
|
533 |
+
return None, "No file loaded"
|
|
|
534 |
image, page_info = load_file_for_preview(file_path)
|
535 |
return image, page_info
|
|
|
536 |
def clear_all():
|
|
|
537 |
global pdf_cache
|
538 |
+
pdf_cache = {"images": [], "current_page": 0, "total_pages": 0, "file_type": None, "is_parsed": False, "results": []}
|
539 |
+
return None, None, '<div class="page-info">No file loaded</div>', None, "Click 'Process Document' to see extracted content...", None
|
540 |
+
file_input.change(handle_file_upload, inputs=[file_input], outputs=[image_preview, page_info])
|
541 |
+
prev_page_btn.click(lambda: turn_page("prev"), outputs=[image_preview, page_info, markdown_output, processed_image, json_output])
|
542 |
+
next_page_btn.click(lambda: turn_page("next"), outputs=[image_preview, page_info, markdown_output, processed_image, json_output])
|
543 |
+
process_btn.click(process_document, inputs=[model_choice, file_input, max_new_tokens, min_pixels, max_pixels], outputs=[processed_image, markdown_output, json_output])
|
544 |
+
clear_btn.click(clear_all, outputs=[file_input, image_preview, page_info, processed_image, markdown_output, json_output])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
545 |
return demo
|
546 |
|
|
|
547 |
if __name__ == "__main__":
|
|
|
548 |
demo = create_gradio_interface()
|
549 |
+
demo.queue(max_size=10).launch(server_name="0.0.0.0", server_port=7860, share=False, debug=True, show_error=True)
|
|
|
|
|
|
|
|
|
|
|
|