Text Extraction Yields cid and Fails on Mixed Content Pages in PDF

[hrhktkbzyy]

Issue:

When attempting to extract text from the attached PDF, several pages return cid values instead of readable text. Additionally, pages containing mixed content (text and images) do not return any text at all.

Affected PDF:

The Phantom of the Opera.pdf

Code Sample:

from pdfminer.high_level import extract_text

def get_text_from_pdf_by_pdfminer(file_path):
    try:
        text = extract_text(file_path.absolute())
        number_of_pages = text.count('\f')
        return text, number_of_pages

    except Exception as e:
        print(e)

Output:

The extracted content includes cid values such as:

(cid:11550)(cid:450)(cid:5509)
(cid:12720)(cid:450)(cid:1275)
(cid:20)(cid:450)(cid:55)(cid:75)(cid:72)(cid:3)(cid:71)(cid:68)(cid:81)(cid:70)(cid:72)(cid:85)(cid:86)
(cid:20)(cid:714)(cid:14414)(cid:17528)(cid:9540)(cid:2696)(cid:1308)
(cid:21)(cid:450)(cid:55)(cid:75)(cid:72)(cid:3)(cid:71)(cid:76)(cid:85)(cid:72)(cid:70)(cid:87)(cid:82)(cid:85)(cid:86)(cid:3)(cid:82)(cid:73)(cid:3)(cid:87)(cid:75)(cid:72)(cid:3)(cid:50)(cid:83)(cid:72)(cid:85)(cid:68)(cid:3)(cid:43)(cid:82)(cid:88)(cid:86)(cid:72)
...

[dhdaines]

This PDF has completely arbitrary and corrupt ToUnicode character mappings, it's unlikely that pdfminer can do much about it. You can see the problem by trying to copy and paste text out of it from your browser's PDF viewer (in my case Chrome). Even the English text is corrupted, for example, "The dancers" on page 3 comes out as:

7KHGDQFHUV