Get text out of a pdf-file?

[Rapid1898-code]

Hello - i try to get the text out of the attached pdf-file using the following code:

import pymupdf
doc = pymupdf.open("test.pdf")
cat = doc.pdf_catalog()
doc.xref_set_key(cat, "StructTreeRoot", "null")
for page in doc:
  text = page.get_text()
  print(text)
  print(type(text))
  input("Press!")

But i only get this text:

pdf) C:\DEV\Fiverr2026\TRY\ad_corcoran>python test2.py
☺☻♥♦♣♠☻ ♠

☻
☻♠
☻►◄►♠↕►☻‼‼¶‼§¶▬↨☻↑↑↑☻↓→☻∟↔▲↔
►↓↔▼♠◄☻∟
☻♦
☺
↔♠◄∟
♠
▲∟►
☻‼‼‼

123'(,4/+543./65137 8623.+59+:1
;<=>?@=ABC<A>=DE>FAGBC<=HIJBKLBMDENO<ABP>=@@>JCDA><AJBQRBSTSTU
VKUWJUSSISS
XF@YBZF[@
\QBC<AO>=]E>F<AJBU^UBQ_O>FEBR`@JQ@ab<=aJBQRBSTU^^
VK^LJSSSISS
XF@YBZF[@
c@A@=D[Bd@`@[<He@A>BfAEJBWTgBYD>@=BO>JYDN@bF@[aJBQRBSULLS
VK^LJhSSISS
XF@YBZF[@
QDOO@_BC<AO>=]E>F<ABC<=HIJBUWWBP>@>O<ABR`@IJPYDeHOE<>>JBQRBSUhSg
VKL^JKW^ISS
XF@YBZF[@
RiCBd@`@[<He@A>Bc=<]HBfAEJBUUK^BjDO?FAG><ABP>=@@>JkDA<`@=JBQRBSTllhVKhWJl^WISS
XF@YBZF[@
\?<eDOBmBPA<Ya@AJBUUgB;BQDFABP>JQFaa[@><AJBQRBSUhWh
VgTLJU^KISS
XF@YBZF[@
cXjBf;CJBlLBn@AA@>>BP>Jo_AAJBQRBSUhS^
VglWJWl^ISS
XF@YBZF[@
P<]>?BC<DO>BfeH=<`@e@A>BC<eHDA_JBUlBQD=E<AFBoDA@JQD=F<AJBQRBSTglL
VgghJLLhISS
XF@YBZF[@
\=FH[@BC<AO>=]E>F<AJBTlgBdDAF@[Bj@pO>@=BkFG?YD_JQ@==FeDENJB;kBSlS^W
VghLJLgLISS
XF@YBZF[@
qqrqsrtuvwtxyzw{|
}~r}{~qqq
}xrr}~r}{~qqq
qrq

<class 'str'>

Is there any way to get the text out of this pdf-file?

test.pdf

[JorjMcKie]

The pdf misses proper back-translation information from the glyphs to the originating Unicodes.

Some background: what you see on any PDF page are glyphs only = graphical representations of the Unicode characters that were used to select them. In the end, glyph numbers are all you have - the connection to the Unicode is lost ... except when the font explicitly or implicitly provide a backward mapping.

This is obviously not the case here. So all you can do is using OCR.
You can more obviously confirm the situation by using flags=0 for extraction. If you then see lots of � characters then you have the assertion.

[JorjMcKie]

Sorry: I have to take back my previous comment:
The situation is different here: using different flags will not change the behavior. The following appears when copy/pasting with Acrobat reader.
You have to convert the page to a picture which you then can treat with OCR.

􀀀����������	�
��
����
���������������������������������
����������
�� ��
�!!�"�����"
�#$�����
�%&��%�
'()*+,- ./(*(0123'(,4/+543./65137 8623.+59+:1
;<=>?@=ABC<A>=DE>FAGBC<=HIJBKLBMDENO<ABP>=@@>JCDA><AJBQRBSTSTU VKUWJUSSISS XF@YBZF[@
\QBC<AO>=]E>F<AJBU^UBQ_O>FEBR'@JQ@ab<=aJBQRBSTU^^ VK^LJSSSISS XF@YBZF[@
c@A@=D[Bd@'@[<He@A>BfAEJBWTgBYD>@=BO>JYDN@bF@[aJBQRBSULLS VK^LJhSSISS XF@YBZF[@
QDOO@_BC<AO>=]E>F<ABC<=HIJBUWWBP>@>O<ABR'@IJPYDeHOE<>>JBQRBSUhSg VKL^JKW^ISS XF@YBZF[@
RiCBd@'@[<He@A>Bc=<]HBfAEJBUUK^BjDO?FAG><ABP>=@@>JkDA<'@=JBQRBSTllhVKhWJl^WISS XF@YBZF[@
\?<eDOBmBPA<Ya@AJBUUgB;BQDFABP>JQFaa[@><AJBQRBSUhWh VgTLJU^KISS XF@YBZF[@
cXjBf;CJBlLBn@AA@>>BP>Jo_AAJBQRBSUhS^ VglWJWl^ISS XF@YBZF[@
P<]>?BC<DO>BfeH=<'@e@A>BC<eHDA_JBUlBQD=E<AFBoDA@JQD=F<AJBQRBSTglL VgghJLLhISS XF@YBZF[@
\=FH[@BC<AO>=]E>F<AJBTlgBdDAF@[Bj@pO>@=BkFG?YD_JQ@==FeDENJB;kBSlS^W VghLJLgLISS XF@YBZF[@
qqrqsrtuvwtxyzw{| }~������������r��������������������}��{~����������qq�q
���