pyxpdf 0.1.1

Powerful and Pythonic PDF processing library based on xpdf-4.02

0.1.1 (2020-05-10)

• FIX: bug where default Config.text_encoding value i.e UTF-8 does not persist Config.reset() and changes to Latin1
• pdftotext: remove all parameters that change global Config properties

pyxpdf

Fast Python PDF parser module based on xpdf-reader sources.

Quickstart

from pyxpdf import Document, Page, Config
from pyxpdf.xpdf import TextControl

doc = Document("samples/nonfree/mandarin.pdf")
# or
# load pdf from file like object
with open("samples/nonfree/mandarin.pdf", 'rb') as fp:
    doc = Document(fp)

# get pdf metadata dict
print(doc.info())
# >>> doc.info()
# {'CreationDate': "D:20080721141207-04'00'", 
#  'Subject': 'Chinese Version of Universal PCXR8 ...', 
#  'Author': 'SKC Inc.', 
#  'Creator': 'PScript5.dll
#   .....

# get all text
all_text = doc.text()

# iter first 10 pages
for page in doc[:10]:
    # get page label if any
    print(page.label)

# get page by page label
label_page = doc['1']

# get text in table layout without discarding clipped
# text.
text_control = TextControl("table", clip_text=True)
text = label_page.text(control=text_control)

# find case sensitive text within [x_min, y_min, x_max, y_max]
res_box = label_page.find_text('操作说明', search_box=[0, 0, 400, 400],
                                case_sensitive=True)
# >>> print(res_box)
# (281.88, 269.718, 354.05819999999994, 287.7)

# load xpdfrc
Config.load_file('my_xpdfrc')
# suppress stderr output for xpdf error log.
Config.error_quiet = False

pdftotext

If you are familiar with pdftotext binary then this is it's python port with almost native binary speed.

from pyxpdf import pdftotext

file = "sample.pdf"
# Get text from first two pages of pdf
pdf_text = pdftotext(file, start=1, end=2, layout="table",
                     userpass="1234", ownerpass="1234", 
                     cfg_file="~/.xpdfrc")

Note:-

• pdftotext returns Unicode encoded string, so if your PDF contain characters outside of utf-8 then they will be ignored [decode('utf-8', errors='ignore')].
• If you are working with different encoding then you can use pdftotext_raw which has same function signature but returns bytes object. You can then decode it yourself but make sure to set Config.text_encoding to your encoding so that xpdf can properly extract text. Currently only 'UTF-8', 'Latin1', 'ASCII7', 'Symbol', 'ZapfDingbats' and 'UCS-2' encodings are predefined. To add additional encodings you can provide Unicode CMaps for your encoding through xpdfrc.

Install

pip install pyxpdf

Note (Windows):-

To build this in windows you will need Visual C++ compiler which you can get by installing Visual Studio Build Tools

Build Instructions

Requirements:-

• (CPython) Python 3.4+
• A recent enough C/C++ build environment

First clone the pyxpdf git repository:

$ git clone https://github.com/ashutoshvarma/pyxpdf.git
$ cd pyxpdf

Optionally create a virtualenv (recommended):

$ python -m venv <directory>
$ source <directory>/bin/activate

Then install the dependencies:

$ pip install -r test_requirements.txt

Build wheel

$ pip install wheel
$ python setup.py bdist_wheel --with-cython

Install wheel package

$ pip install dist/*.whl

Now you can run the tests

$ python runtests.py -v

License

pyxpdf is licensed under the GNU General Public License (GPL), version 3. See the LICENSE

It uses following third party sources :-

• Xpdf Reader [https://www.xpdfreader.com/] by Derek Noonburg