summaryrefslogtreecommitdiff
path: root/mammoth/raw_text.py
blob: 5ffe59f4b82409d437bd0cbda5a4362f21a89f52 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
from . import documents


def extract_raw_text_from_element(element):
    if isinstance(element, documents.Text):
        return element.value
    elif isinstance(element, documents.Tab):
        return "\t"
    else:
        text = "".join(map(extract_raw_text_from_element, getattr(element, "children", [])))
        if isinstance(element, documents.Paragraph):
            return text + "\n\n"
        else:
            return text