Update app.py
Browse files
app.py
CHANGED
|
@@ -340,6 +340,7 @@ def run_single_pdf_preprocessing(pdf_path: str) -> Tuple[int, int, int, str, flo
|
|
| 340 |
global GLOBAL_FIGURE_COUNT, GLOBAL_EQUATION_COUNT
|
| 341 |
start_time = time.time()
|
| 342 |
log_messages = []
|
|
|
|
| 343 |
all_base64_images: List[str] = []
|
| 344 |
|
| 345 |
# Dictionary to store {page_number (int): equation_count (int)}
|
|
@@ -399,7 +400,7 @@ def run_single_pdf_preprocessing(pdf_path: str) -> Tuple[int, int, int, str, flo
|
|
| 399 |
detect_start = time.time()
|
| 400 |
# page_equations, _ = run_yolo_detection_and_count(original_img, model, page_num)
|
| 401 |
page_equations, _, page_images = run_yolo_detection_and_count(original_img, model, page_num)
|
| 402 |
-
|
| 403 |
|
| 404 |
detect_time = time.time() - detect_start
|
| 405 |
|
|
@@ -437,7 +438,8 @@ def run_single_pdf_preprocessing(pdf_path: str) -> Tuple[int, int, int, str, flo
|
|
| 437 |
|
| 438 |
# Return the dictionary with string keys
|
| 439 |
# return total_pages, GLOBAL_EQUATION_COUNT, GLOBAL_FIGURE_COUNT, report, total_execution_time, equation_counts_per_page_str_keys, []
|
| 440 |
-
return total_pages, GLOBAL_EQUATION_COUNT, GLOBAL_FIGURE_COUNT, report, total_execution_time, equation_counts_per_page_str_keys,
|
|
|
|
| 441 |
|
| 442 |
|
| 443 |
# ============================================================================
|
|
@@ -459,8 +461,10 @@ def gradio_process_pdf(pdf_file) -> Tuple[str, str, str, str, Dict[str, int], Li
|
|
| 459 |
# num_pages, num_equations, num_figures, report, total_time, equation_counts_per_page, _ = run_single_pdf_preprocessing(
|
| 460 |
# pdf_path
|
| 461 |
# )
|
|
|
|
| 462 |
num_pages, num_equations, num_figures, report, total_time, equation_counts_per_page, images = run_single_pdf_preprocessing(pdf_path)
|
| 463 |
|
|
|
|
| 464 |
|
| 465 |
# Return results (6 items now)
|
| 466 |
# return str(num_pages), str(num_equations), str(num_figures), report, equation_counts_per_page, []
|
|
|
|
| 340 |
global GLOBAL_FIGURE_COUNT, GLOBAL_EQUATION_COUNT
|
| 341 |
start_time = time.time()
|
| 342 |
log_messages = []
|
| 343 |
+
all_saved_images = []
|
| 344 |
all_base64_images: List[str] = []
|
| 345 |
|
| 346 |
# Dictionary to store {page_number (int): equation_count (int)}
|
|
|
|
| 400 |
detect_start = time.time()
|
| 401 |
# page_equations, _ = run_yolo_detection_and_count(original_img, model, page_num)
|
| 402 |
page_equations, _, page_images = run_yolo_detection_and_count(original_img, model, page_num)
|
| 403 |
+
all_saved_images.extend(page_images)
|
| 404 |
|
| 405 |
detect_time = time.time() - detect_start
|
| 406 |
|
|
|
|
| 438 |
|
| 439 |
# Return the dictionary with string keys
|
| 440 |
# return total_pages, GLOBAL_EQUATION_COUNT, GLOBAL_FIGURE_COUNT, report, total_execution_time, equation_counts_per_page_str_keys, []
|
| 441 |
+
return total_pages, GLOBAL_EQUATION_COUNT, GLOBAL_FIGURE_COUNT, report, total_execution_time, equation_counts_per_page_str_keys, all_saved_images
|
| 442 |
+
|
| 443 |
|
| 444 |
|
| 445 |
# ============================================================================
|
|
|
|
| 461 |
# num_pages, num_equations, num_figures, report, total_time, equation_counts_per_page, _ = run_single_pdf_preprocessing(
|
| 462 |
# pdf_path
|
| 463 |
# )
|
| 464 |
+
# num_pages, num_equations, num_figures, report, total_time, equation_counts_per_page, images = run_single_pdf_preprocessing(pdf_path)
|
| 465 |
num_pages, num_equations, num_figures, report, total_time, equation_counts_per_page, images = run_single_pdf_preprocessing(pdf_path)
|
| 466 |
|
| 467 |
+
|
| 468 |
|
| 469 |
# Return results (6 items now)
|
| 470 |
# return str(num_pages), str(num_equations), str(num_figures), report, equation_counts_per_page, []
|