Skip to main content

Python parser to extract data from pdf invoice

Project description

Data extractor for PDF invoices - ninvoice2data

|Circle CI|

This project has been selected for `GSoC
2018 <>`__. Read more
`here <>`__.

A modular Python library to support your accounting process. Tested on
Python 2.7 and 3.4+. Main steps:

1. extracts text from PDF files using different techniques, like
``pdftotext``, ``pdfminer`` or OCR – ``tesseract``, ``tesseract4`` or
``gvision`` (Google Cloud Vision).
2. searches for regex in the result using a YAML-based template system
3. saves results as CSV, JSON or XML or renames PDF files to match the

With the flexible template system you can:

- precisely match content PDF files
- plugins available to match line items and tables
- define static fields that are the same for every invoice
- define custom fields needed in your organisation or process
- have multiple regex per field (if layout or wording changes)
- define currency
- extract invoice-items using the ``lines``-plugin developed by `Holger
Brunn <>`__

Go from PDF files to this:


{'date': (2014, 5, 7), 'invoice_number': '30064443', 'amount': 34.73, 'desc': 'Invoice 30064443 from QualityHosting', 'lines': [{'price': 42.0, 'desc': u'Small Business StandardExchange 2010\nGrundgeb\xfchr pro Einheit\nDienst: OUDJQ_office\n01.05.14-31.05.14\n', 'pos': u'7', 'qty': 1.0}]}
{'date': (2014, 6, 4), 'invoice_number': 'EUVINS1-OF5-DE-120725895', 'amount': 35.24, 'desc': 'Invoice EUVINS1-OF5-DE-120725895 from Amazon EU'}
{'date': (2014, 8, 3), 'invoice_number': '42183017', 'amount': 4.11, 'desc': 'Invoice 42183017 from Amazon Web Services'}
{'date': (2015, 1, 28), 'invoice_number': '12429647', 'amount': 101.0, 'desc': 'Invoice 12429647 from Envato'}


1. Install pdftotext

If possible get the latest
`xpdf/poppler-utils <>`__ version. It’s
included with macOS Homebrew, Debian and Ubuntu. Without it,
``pdftotext`` won’t parse tables in PDF correctly.

2. Install ``ninvoice2data`` using pip


pip install ninvoice2data


Basic usage. Process PDF files and write result to CSV.

- ``ninvoice2data invoice.pdf``
- ``ninvoice2data *.pdf``

Choose any of the following input readers:

- pdftotext ``ninvoice2data --input-reader pdftotext invoice.pdf``
- tesseract ``ninvoice2data --input-reader tesseract invoice.pdf``
- pdf miner ``ninvoice2data --input-reader pdfminer invoice.pdf``
- tesseract4 ``ninvoice2data --input-reader tesseract4 invoice.pdf``
- gvision ``ninvoice2data --input-reader gvision invoice.pdf`` (needs ``GOOGLE_APPLICATION_CREDENTIALS`` env var)

Choose any of the following output formats:

- csv ``ninvoice2data --output-format csv invoice.pdf``
- json ``ninvoice2data --output-format json invoice.pdf``
- xml ``ninvoice2data --output-format xml invoice.pdf``

Save output file with custom name or a specific folder
``ninvoice2data --output-format csv --output-name myinvoices/invoices.csv invoice.pdf``

**Note:** You must specify the ``output-format`` in order to create

Specify folder with yml templates. (e.g. your suppliers)
``ninvoice2data --template-folder ACME-templates invoice.pdf``

Only use your own templates and exclude built-ins
``ninvoice2data --exclude-built-in-templates --template-folder ACME-templates invoice.pdf``

Processes a folder of invoices and copies renamed invoices to new
folder. ``ninvoice2data --copy new_folder folder_with_invoices/*.pdf``

Processes a single file and dumps whole file for debugging (useful when
adding new templates in
``ninvoice2data --debug my_invoice.pdf``

Recognize test invoices:
``ninvoice2data ninvoice2data/test/pdfs/* --debug``

If you want to use it as a lib just do


from ninvoice2data import extract_data

result = extract_data('path/to/my/file.pdf')

Template system

See ``ninvoice2data/extract/templates`` for existing templates. Just extend the
list to add your own. If deployed by a bigger organisation, there should
be an interface to edit templates for new suppliers. 80-20 rule. For a
short tutorial on how to add new templates, see
`TUTORIAL.rst <TUTORIAL.rst>`__.

Templates are based on Yaml. They define one or more keywords to find
the right template and regexp for fields to be extracted. They could
also be a static value, like the full company name.

Template files are tried in alphabetical order.

We may extend them to feature options to be used during invoice



issuer: Amazon Web Services, Inc.
- Amazon Web Services
amount: TOTAL AMOUNT DUE ON.*\$(\d+\.\d+)
amount_untaxed: TOTAL AMOUNT DUE ON.*\$(\d+\.\d+)
date: Invoice Date:\s+([a-zA-Z]+ \d+ , \d+)
invoice_number: Invoice Number:\s+(\d+)
partner_name: (Amazon Web Services, Inc\.)
remove_whitespace: false
currency: HKD
- '%d/%m/%Y'
start: Detail
end: \* May include estimated US sales tax
first_line: ^ (?P<description>\w+.*)\$(?P<price_unit>\d+\.\d+)
line: (.*)\$(\d+\.\d+)
last_line: VAT \*\*


If you are interested in improving this project, have a look at our
`developer guide <DEVELOP.rst>`__ to get you started quickly.

Roadmap and open tasks

- integrate with online OCR?
- try to ‘guess’ parameters for new invoice formats.
- can apply machine learning to guess new parameters?


- `Manuel Riel <>`__
- `Alexis de Lattre <>`__


- `Harshit Joshi <>`__: As Google Summer of
Code student.
- `Holger Brunn <>`__: Add support for parsing
invoice items.

Related Projects

- `OCR-Invoice <>`__ (FOSS \| C#)
- `Docparser <>`__ (Commercial \| Web Service)
- `A-PDF <>`__
- `PDFdeconstruct <>`__
- `CVision <>`__

.. |Circle CI| image::

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ninvoice2data-0.4.16.tar.gz (757.5 kB view hashes)

Uploaded Source

Built Distribution

ninvoice2data-0.4.16-py2.7.egg (93.8 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page