words extracted from pdf getting split

[nischithac]

Please provide all mandatory information!

Describe the bug (mandatory)

Few words are getting split into half . Example beneficiaries word is split into 'benefi', 'ciaries' . But there is no space between this words in pdf

To Reproduce (mandatory)

doc = fitz.open(input_file)
page = doc[0]
page.getText()
Output: The discounts are not available to benefi ciaries of \nMedicare, Medicaid or other federal or state healthcare programs or residents of Massachusetts, Puerto Rico and other US territories.

Expected behavior (optional)

words shouldn't get split.
Expected Output: The discounts are not available to beneficiaries of \nMedicare, Medicaid or other federal or state healthcare programs or residents of Massachusetts, Puerto Rico and other US territories.

Screenshots (optional)

Your configuration (mandatory)

• Operating system, potentially version and bitness: Ubuntu
• Python version, bitness: Python 3.6
• PyMuPDF version, installation method (wheel or generated from source). Pymupdf 1.18.9

For example, the output of print(sys.version, "\n", sys.platform, "\n", fitz.__doc__) would be sufficient (for the first two bullets).

Additional context (optional)

Add any other context about the problem here.

[JorjMcKie]

Please:

• upgrade to the current version
• provide minimal example file to reproduce
• read the documentation concerning text extraction flag TEXT_INHIBIT_SPACES and friends.
• try this layout-preserving script.

[nischithac]

I tried all the solutions which you mentioned but its not working. This is a sample pdf file where the words 'first', 'notifications' are getting split.
Code:
doc = fitz.open(pr_input_file)
page = doc[0]
blocks = page.getText('rawdict', flags=fitz.TEXT_INHIBIT_SPACES)
for block in blocks['blocks']:
if block['type'] == 0:
for line in block['lines']:
for span in line['spans']:
text = ''
for char in span['chars']:
text+=char['c']
print(text)
Output:
Fingersticks are required if your glucose alarms and readings do not match symptoms or when you see Check Blood Glucose symbol during the fi rst 12 hours.
**
PA may be required.
†
Notifi cations will only be received when alarms are turned on and the sensor is within 20 feet of the reading device.
Split_word_issue.pdf

[JorjMcKie]

This is a badly constructed PDF:
Invisibly for the PDF viewer, space characters are specified which (partly) overlap their preceeding character. E.g. "Notifi" ends at some x-coordinate, say 25.5, then a space character follows which sparts at 25.0 ending at 26.0, followed by "cations" ...
I have modified the script textlayout.py such that it detects the situation and ignores those spaces. I have attached it here.
Play with it until you see acceptable results.
It definitely is not a bug of PyMuPDF, but the PDF maker screwed up the file.
textlayout.zip