| """ |
| Copyright (c) 2025 Bytedance Ltd. and/or its affiliates |
| SPDX-License-Identifier: MIT |
| """ |
|
|
| import argparse |
| import glob |
| import os |
|
|
| import cv2 |
| import torch |
| from PIL import Image |
| from transformers import AutoProcessor, VisionEncoderDecoderModel |
|
|
| from utils.utils import * |
|
|
|
|
| class DOLPHIN: |
| def __init__(self, model_id_or_path): |
| """Initialize the Hugging Face model |
| |
| Args: |
| model_id_or_path: Path to local model or Hugging Face model ID |
| """ |
| |
| self.processor = AutoProcessor.from_pretrained(model_id_or_path) |
| self.model = VisionEncoderDecoderModel.from_pretrained(model_id_or_path) |
| self.model.eval() |
| |
| |
| self.device = "cuda" if torch.cuda.is_available() else "cpu" |
| self.model.to(self.device) |
| self.model = self.model.half() |
| |
| |
| self.tokenizer = self.processor.tokenizer |
| |
| def chat(self, prompt, image): |
| """Process an image or batch of images with the given prompt(s) |
| |
| Args: |
| prompt: Text prompt or list of prompts to guide the model |
| image: PIL Image or list of PIL Images to process |
| |
| Returns: |
| Generated text or list of texts from the model |
| """ |
| |
| is_batch = isinstance(image, list) |
| |
| if not is_batch: |
| |
| images = [image] |
| prompts = [prompt] |
| else: |
| |
| images = image |
| prompts = prompt if isinstance(prompt, list) else [prompt] * len(images) |
| |
| |
| batch_inputs = self.processor(images, return_tensors="pt", padding=True) |
| batch_pixel_values = batch_inputs.pixel_values.half().to(self.device) |
| |
| |
| prompts = [f"<s>{p} <Answer/>" for p in prompts] |
| batch_prompt_inputs = self.tokenizer( |
| prompts, |
| add_special_tokens=False, |
| return_tensors="pt" |
| ) |
|
|
| batch_prompt_ids = batch_prompt_inputs.input_ids.to(self.device) |
| batch_attention_mask = batch_prompt_inputs.attention_mask.to(self.device) |
| |
| |
| outputs = self.model.generate( |
| pixel_values=batch_pixel_values, |
| decoder_input_ids=batch_prompt_ids, |
| decoder_attention_mask=batch_attention_mask, |
| min_length=1, |
| max_length=4096, |
| pad_token_id=self.tokenizer.pad_token_id, |
| eos_token_id=self.tokenizer.eos_token_id, |
| use_cache=True, |
| bad_words_ids=[[self.tokenizer.unk_token_id]], |
| return_dict_in_generate=True, |
| do_sample=False, |
| num_beams=1, |
| repetition_penalty=1.1 |
| ) |
| |
| |
| sequences = self.tokenizer.batch_decode(outputs.sequences, skip_special_tokens=False) |
| |
| |
| results = [] |
| for i, sequence in enumerate(sequences): |
| cleaned = sequence.replace(prompts[i], "").replace("<pad>", "").replace("</s>", "").strip() |
| results.append(cleaned) |
| |
| |
| if not is_batch: |
| return results[0] |
| return results |
|
|
|
|
| def process_page(image_path, model, save_dir, max_batch_size=None): |
| """Parse document images with two stages""" |
| |
| pil_image = Image.open(image_path).convert("RGB") |
| layout_output = model.chat("Parse the reading order of this document.", pil_image) |
|
|
| |
| padded_image, dims = prepare_image(pil_image) |
| recognition_results = process_elements(layout_output, padded_image, dims, model, max_batch_size) |
|
|
| |
| json_path = save_outputs(recognition_results, image_path, save_dir) |
|
|
| return json_path, recognition_results |
|
|
|
|
| def process_elements(layout_results, padded_image, dims, model, max_batch_size=None): |
| """Parse all document elements with parallel decoding""" |
| layout_results = parse_layout_string(layout_results) |
|
|
| |
| text_elements = [] |
| table_elements = [] |
| figure_results = [] |
| previous_box = None |
| reading_order = 0 |
|
|
| |
| for bbox, label in layout_results: |
| try: |
| |
| x1, y1, x2, y2, orig_x1, orig_y1, orig_x2, orig_y2, previous_box = process_coordinates( |
| bbox, padded_image, dims, previous_box |
| ) |
|
|
| |
| cropped = padded_image[y1:y2, x1:x2] |
| if cropped.size > 0: |
| if label == "fig": |
| |
| figure_results.append( |
| { |
| "label": label, |
| "bbox": [orig_x1, orig_y1, orig_x2, orig_y2], |
| "text": "", |
| "reading_order": reading_order, |
| } |
| ) |
| else: |
| |
| pil_crop = Image.fromarray(cv2.cvtColor(cropped, cv2.COLOR_BGR2RGB)) |
| element_info = { |
| "crop": pil_crop, |
| "label": label, |
| "bbox": [orig_x1, orig_y1, orig_x2, orig_y2], |
| "reading_order": reading_order, |
| } |
| |
| |
| if label == "tab": |
| table_elements.append(element_info) |
| else: |
| text_elements.append(element_info) |
|
|
| reading_order += 1 |
|
|
| except Exception as e: |
| print(f"Error processing bbox with label {label}: {str(e)}") |
| continue |
|
|
| |
| recognition_results = figure_results.copy() |
| |
| |
| if text_elements: |
| text_results = process_element_batch(text_elements, model, "Read text in the image.", max_batch_size) |
| recognition_results.extend(text_results) |
| |
| |
| if table_elements: |
| table_results = process_element_batch(table_elements, model, "Parse the table in the image.", max_batch_size) |
| recognition_results.extend(table_results) |
|
|
| |
| recognition_results.sort(key=lambda x: x.get("reading_order", 0)) |
|
|
| return recognition_results |
|
|
|
|
| def process_element_batch(elements, model, prompt, max_batch_size=None): |
| """Process elements of the same type in batches""" |
| results = [] |
| |
| |
| batch_size = len(elements) |
| if max_batch_size is not None and max_batch_size > 0: |
| batch_size = min(batch_size, max_batch_size) |
| |
| |
| for i in range(0, len(elements), batch_size): |
| batch_elements = elements[i:i+batch_size] |
| crops_list = [elem["crop"] for elem in batch_elements] |
| |
| |
| prompts_list = [prompt] * len(crops_list) |
| |
| |
| batch_results = model.chat(prompts_list, crops_list) |
| |
| |
| for j, result in enumerate(batch_results): |
| elem = batch_elements[j] |
| results.append({ |
| "label": elem["label"], |
| "bbox": elem["bbox"], |
| "text": result.strip(), |
| "reading_order": elem["reading_order"], |
| }) |
| |
| return results |
|
|
|
|
| def main(): |
| parser = argparse.ArgumentParser(description="Document processing tool using DOLPHIN model") |
| parser.add_argument("--input_path", type=str, default="./demo", help="Path to input image or directory of images") |
| parser.add_argument( |
| "--save_dir", |
| type=str, |
| default=None, |
| help="Directory to save parsing results (default: same as input directory)", |
| ) |
| parser.add_argument( |
| "--max_batch_size", |
| type=int, |
| default=16, |
| help="Maximum number of document elements to parse in a single batch (default: 16)", |
| ) |
| args = parser.parse_args() |
|
|
| |
| model = DOLPHIN("ByteDance/Dolphin") |
|
|
| |
| if os.path.isdir(args.input_path): |
| image_files = [] |
| for ext in [".jpg", ".jpeg", ".png", ".JPG", ".JPEG", ".PNG"]: |
| image_files.extend(glob.glob(os.path.join(args.input_path, f"*{ext}"))) |
| image_files = sorted(image_files) |
| else: |
| if not os.path.exists(args.input_path): |
| raise FileNotFoundError(f"Input path {args.input_path} does not exist") |
| image_files = [args.input_path] |
|
|
| save_dir = args.save_dir or ( |
| args.input_path if os.path.isdir(args.input_path) else os.path.dirname(args.input_path) |
| ) |
| setup_output_dirs(save_dir) |
|
|
| total_samples = len(image_files) |
| print(f"\nTotal samples to process: {total_samples}") |
|
|
| |
| for image_path in image_files: |
| print(f"\nProcessing {image_path}") |
| try: |
| json_path, recognition_results = process_page( |
| image_path=image_path, |
| model=model, |
| save_dir=save_dir, |
| max_batch_size=args.max_batch_size, |
| ) |
|
|
| print(f"Processing completed. Results saved to {save_dir}") |
|
|
| except Exception as e: |
| print(f"Error processing {image_path}: {str(e)}") |
| continue |
|
|
|
|
| if __name__ == "__main__": |
| main() |