File size: 4,906 Bytes
6075d60
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
import argparse
from glob import glob
import os
import xml.etree.ElementTree as ET
import pandas as pd
import sys
from process_csvs import process_csvs

# Ensure you have natsort installed: pip install natsort
from natsort import natsorted

def flatten_macula_xml(xml_data: str) -> pd.DataFrame:
    """
    Parses Macula GNT XML data to produce a flattened, duplicate-free DataFrame.
    It corrects for "broken clauses" by re-assigning clause-linking conjunctions
    to the clause they introduce. The final output is sorted by biblical text order.

    Args:
        xml_data: A string containing the Macula XML data.

    Returns:
        A pandas DataFrame with the flattened syntactic information.
    """
    all_words_data = []

    def _traverse_and_collect_words(node, sentence_id, active_clause_id, active_phrase_id):
        """
        Recursively performs a single top-down traversal of the tree, collecting
        word data and propagating the correct IDs.
        """
        current_cat = node.get('Cat', '').lower()
        
        is_word_node = (node.text and node.text.strip()) and not node.findall('Node')
        if is_word_node:
            word_info = {
                'sentence_id': sentence_id,
                'clause_id': active_clause_id,
                'phrase_id': active_phrase_id,
                'word_id': node.get('{http://www.w3.org/XML/1998/namespace}id'),
                'ref': node.get('ref'),
                'text': node.text.strip(),
                'lemma': node.get('UnicodeLemma'),
                'gloss': node.get('English'),
                'strong': node.get('StrongNumber'),
                'morph': node.get('FunctionalTag'),
            }
            all_words_data.append(word_info)
            return

        if current_cat == 'cl':
            new_clause_id = node.get('nodeId')
            for child_phrase in node.findall('./Node'):
                new_phrase_id = child_phrase.get('nodeId')
                _traverse_and_collect_words(child_phrase, sentence_id, new_clause_id, new_phrase_id)
        else:
            for child in node.findall('./Node'):
                _traverse_and_collect_words(child, sentence_id, active_clause_id, active_phrase_id)

    # --- Pass 1: Collect all word data and sort into linear text order ---
    root = ET.fromstring(xml_data)

    for sentence in root.findall('Sentence'):
        sentence_ref = sentence.get('ref')
        tree_root = sentence.find('.//Tree/Node')
        if tree_root is not None:
            _traverse_and_collect_words(tree_root, sentence_ref, active_clause_id=None, active_phrase_id=None)

    sorted_words_data = natsorted(all_words_data, key=lambda x: x['ref'])
    
    # --- Pass 2: Post-processing to fix broken clauses ---
    # Iterate through the sorted list to re-assign clause-linking conjunctions.
    for i in range(len(sorted_words_data) - 1):
        current_word = sorted_words_data[i]
        next_word = sorted_words_data[i+1]

        # A word is a clause-linking conjunction if its morph tag is CONJ
        # and its clause differs from the word immediately following it.
        is_conjunction = current_word.get('morph') == 'CONJ'
        
        if is_conjunction and current_word['clause_id'] != next_word['clause_id']:
            # Re-assign this conjunction to the next word's clause and phrase.
            current_word['clause_id'] = next_word['clause_id']
            current_word['phrase_id'] = next_word['phrase_id']
            
    return pd.DataFrame(sorted_words_data)


# --- Main execution block ---
if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Flatten Macula GNT XML files.")
    parser.add_argument("root_folder", help="Glob pattern for input XML files (e.g., '../macula-greek/SBLGNT/nodes/*')")
    args = parser.parse_args()

    root_folder = args.root_folder
    
    # I'm using `grass/` because it's no longer a tree :)
    os.makedirs("grass", exist_ok=True)
    
    for file_path in glob(root_folder):
        filename = file_path.split("/")[-1]
        try:
            with open(file_path, 'r', encoding='utf-8') as f:
                xml_content = f.read()
            
            flat_df = flatten_macula_xml(xml_content)
            
            print(f"Successfully processed '{file_path}'.")
            print("--- Flattened Macula Data Writing to File ---\n")

            out_file = "grass/" + filename.replace(".xml", "_flat_corrected.csv")
            with open(out_file, "w", encoding="utf-8") as out_file:
                out_file.write(flat_df.to_csv())

        except ET.ParseError as e:
            print(f"Error: Could not parse the XML file. It may be malformed.", file=sys.stderr)
            print(f"Details: {e}", file=sys.stderr)
        except Exception as e:
            print(f"An unexpected error occurred: {e}", file=sys.stderr)

    process_csvs("macula_grass.csv")