epstein-imessage-viewer/convert_to_json.py

import json
import sys

def parse_log_file(file_path, labels):
    with open(file_path, 'r') as f:
        lines = f.readlines()

    data = {"Messages": []}
    message_section = False
    current_message = {}

    # Create a regex pattern to find any of the labels
    import re
    label_pattern = re.compile(f"({'|'.join(re.escape(label) for label in labels)}):")

    i = 0
    while i < len(lines):
        line = lines[i].strip()

        if line == "Messages -":
            message_section = True
            i += 1
            continue

        if not message_section:
            match = label_pattern.match(line)
            if match:
                key = match.group(1).strip()
                value = line[len(match.group(0)):].strip()
                # Handle multi-line values for the initial fields
                while i + 1 < len(lines) and not label_pattern.match(lines[i+1]) and lines[i+1].strip() != "Messages -":
                    value += " " + lines[i+1].strip()
                    i += 1
                data[key] = value
        else:
            if label_pattern.match(line):
                if current_message:
                    # Before starting a new message, save the previous one if it exists
                    if any(current_message.values()): # Save only if not empty
                        data["Messages"].append(current_message)
                    current_message = {}

                key = label_pattern.match(line).group(1).strip()
                value_part = line[len(label_pattern.match(line).group(0)):].strip()

                # Multi-line message content
                if key == "Message":
                    message_content = [value_part]
                    while i + 1 < len(lines) and not label_pattern.match(lines[i+1]):
                        message_content.append(lines[i+1].strip())
                        i += 1
                    current_message[key] = " ".join(message_content).strip()
                else:
                    current_message[key] = value_part
            elif line.strip() and not label_pattern.match(line):
                 # This handles cases where a message might not have a "Message:" label
                 # and is just free text between other labeled fields.
                 if "Message" in current_message:
                     current_message["Message"] += " " + line
                 else:
                     current_message["Message"] = line


        i += 1
    
    # Add the last message
    if current_message and any(current_message.values()):
        data["Messages"].append(current_message)


    return data

if __name__ == "__main__":
    log_file_path = sys.argv[1]
    
    with open('data_labels.json', 'r') as f:
        labels = json.load(f)

    parsed_data = parse_log_file(log_file_path, labels)
    print(json.dumps(parsed_data, indent=4))