blob: 3df61dd2b84a074b774aaa8f869231fe5940cc95 (
plain)
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
|
from io import StringIO
from typing import List, Dict, Optional
import xml.etree.ElementTree as ET
def djvu_extract_leaf_texts(
blob: StringIO, only_leaves: Optional[List[int]] = None
) -> Dict[int, str]:
"""
Takes an in-memory djvu XML string (note: not an actual djvu file, just the
IA XML file type), and iterates throug
"""
leaf_text = dict()
max_leaf = None
if only_leaves:
max_leaf = max(only_leaves)
elem_iter = ET.iterparse(blob, ["start", "end"])
for (event, element) in elem_iter:
if event == "start":
continue
if not (element.tag == "OBJECT" and event == "end"):
continue
# <OBJECT data="file://localhost//tmp/derive/ERIC_ED441501//ERIC_ED441501.djvu" height="6545" type="image/x.djvu" usemap="ERIC_ED441501_0002.djvu" width="5048">
usemap = element.get("usemap")
if not usemap:
continue
leaf_num = None
try:
leaf_num = int(usemap.replace(".djvu", "").split("_")[-1])
except Exception:
continue
if only_leaves is not None and leaf_num is not None:
if leaf_num not in only_leaves:
continue
if max_leaf is not None and leaf_num > max_leaf:
break
paragraph_texts = []
for paragraph in element.iter("PARAGRAPH"):
tokens = [r.strip() for r in paragraph.itertext()]
tokens = [t for t in tokens if t]
p_text = " ".join(tokens)
if p_text:
paragraph_texts.append(p_text)
page_text = "\n".join(paragraph_texts)
# print(f"### {leaf_num}\n{page_text}\n")
if page_text:
leaf_text[leaf_num] = page_text
element.clear()
return leaf_text
|
