eval.py

# Copyright (c) 2025 Bytedance Ltd. and/or its affiliates
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

from Levenshtein import distance as levenshtein_distance
import json

import argparse
def calculate_metrics(pred, truth):

    pred_clean = pred.replace(" ", "")
    truth_clean = truth.replace(" ", "")
    
    correct_chars = sum(c1 == c2 for c1, c2 in zip(pred_clean, truth_clean))
    char_acc = correct_chars / len(truth_clean) if len(truth_clean) > 0 else 0
    
    edit_dist = levenshtein_distance(pred_clean, truth_clean)
    max_len = max(len(pred_clean), len(truth_clean))
    similarity = 1 - (edit_dist / max_len) if max_len > 0 else 1
    
    return {
        "char_accuracy": char_acc,
        "edit_distance": edit_dist,
        "similarity": similarity
    }
def compare_three_values(all_samples):
    clear_stats = {"total_acc": 0, "total_sim": 0, "count": 0}
    notclear_stats = {"total_acc": 0, "total_sim": 0, "count": 0}
    final_stats = {"total_acc": 0, "total_sim": 0, "count": 0}
    
    for sample in all_samples:
        clear_metrics = calculate_metrics(sample[0]["clear Char-level OCR"], sample[1]["clear Char-level OCR"])
        clear_stats["total_acc"] += clear_metrics["char_accuracy"]
        clear_stats["total_sim"] += clear_metrics["similarity"]
        clear_stats["count"] += 1

        notclear_metrics = calculate_metrics(sample[0]["not clear enough Char-level OCR"], sample[1]["not clear enough Char-level OCR"])
        notclear_stats["total_acc"] += notclear_metrics["char_accuracy"]
        notclear_stats["total_sim"] += notclear_metrics["similarity"]
        notclear_stats["count"] += 1

        final_metrics = calculate_metrics(sample[0]["Final OCR"],sample[1]["Final OCR"])
        final_stats["total_acc"] += final_metrics["char_accuracy"]
        final_stats["total_sim"] += final_metrics["similarity"]
        final_stats["count"] += 1
    print({
        "Clear": {
            "avg_accuracy": clear_stats["total_acc"] / clear_stats["count"],
            "avg_similarity": clear_stats["total_sim"] / clear_stats["count"]
        },
        "NotClear": {
            "avg_accuracy": notclear_stats["total_acc"] / notclear_stats["count"],
            "avg_similarity": notclear_stats["total_sim"] / notclear_stats["count"]
        },
        "Final": {
            "avg_accuracy": final_stats["total_acc"] / final_stats["count"],
            "avg_similarity": final_stats["total_sim"] / final_stats["count"]
        }
    })
def parse_arguments():
    parser = argparse.ArgumentParser()
    parser.add_argument("--input_file", type=str)

    args = parser.parse_args()

    return args
def main():
    args = parse_arguments()

    with open(args.input_file, 'r') as f_in:
        now = []
        for line in f_in:
            data = json.loads(line.strip())
            text = data['answer']
            start_tag = "<answer>"
            end_tag = "</answer>"
            start_index = text.find(start_tag) + len(start_tag)
            end_index = text.find(end_tag)
            answer_content = text[start_index:end_index].strip()
            parsed_json_gt = json.loads(answer_content)
            q = data['response']

            start_tag = "<answer>"
            end_tag = "</answer>"
            start_index = q.find(start_tag) + len(start_tag)
            end_index = q.find(end_tag)
            answer_content = q[start_index:end_index].strip()
            answer_content = q

            answer_content_json = answer_content.replace("'", '"').replace("```json","").replace("```","").replace("\n","")

            parsed_json_pred = json.loads(answer_content_json)
            now.append([parsed_json_gt,parsed_json_pred])
    compare_three_values(now)

if __name__ == "__main__":
    main()