TabulaPro: Pro-version of Tabula-py

TabulaPro is a layer on the tabula-py library to extract tables from Scan PDFs and Images.

TabulaPro vs Tabula

TabulaPro is no different from the original Tabula to code. Turn your current tabula-py code to TabulaPro compatible with flavor="TabulaPro" or tabulapro=True in read_pdf() to process images or scanned PDFs".

Installation

💡 ProTip: ExtractTable-py is the official library, FASTER than this wrapper, has NO software dependencies.

As the library itself is dependent on Tabula which has software dependencies, the developer is expected to install them*, to use the regular Tabula flavors ("stream", "lattice") along with "TabulaPro".

Using pip

After installing software dependencies, you can simply use pip to install TabulaPro:

$ pip install -U TabulaPro  

Prerequisites

The developer needs an api_key (free credits here) to use TabulaPro. Each Image file or one PDF page consumes one credit to trigger the process.

api_key should be passed through pro_kwargs, a dict type argument that accepts api_key, job_id, dup_check, wait_for_output as keys, can be used as below

{
    "api_key": str,
    Mandatory, to trigger "TabulaPro" flavor, to process Scan PDFs and images, also text PDF files

    "job_id": str,
        optional, if processing a new file
        Mandatory, to retrieve the result of the already submitted file

    "dup_check": bool, default: False - to bypass the duplicate check
        Useful to handle duplicate requests, check based on the FileName

    "max_wait_time": int, default: 300
        Checks for the output every 15 seconds until successfully processed or for a maximum of 300 seconds.
}

Let's code

Quickly validate the API key and see the number of credits attached to it

api_key = YOUR_API_KEY_HERE

from tabula_pro import check_usage
print(check_usage(api_key))

No error from the above code snippet run implies API Key is valid

Here's how you can extract tables from Image files.

The example image (tabula-data-page-1.PNG) used in the code below, can be found here. Notice that tabula-data-page-1.PNG is the image version of the first page of Tabula's PDF example, data.pdf.

from tabula_pro import read_pdf
pro_tables = read_pdf(
    'foo-image.jpg', 
    flavor="tabulapro", 
    pro_kwargs={"api_key": api_key}
)

# To process PDF, make use of pages ("1", "1,3-4", "all") params in the read_pdf function
# pro_tables = read_pdf('foo-image.PDF', flavor="tabulaPro", pages="1,3-4", pro_kwargs={'api_key': api_key})

pro_tables is a list of dataframes that are found in the file

pro_tables[0]

Most of the image files are processed under 5 seconds. At times a blurry/big/bad image processing may take up to 15 seconds and the PDF file depends on the page count. In these cases, the process waits for a maximum of 300 seconds to check the job status every 15 seconds until a process ends successfully to return a final response.

ProTip: To have more control on the process wait time checkout ExtractTable-py

Pull Requests & Rewards

Pull requests are most welcome and greatly appreciated with API credits.

License

This project is licensed under the Apache License 2.0, see the LICENSE file for details.

Credits

Last but not least, we want to be thankful to the contributors of tabula-py

Social Media

Follow us on Social media for library updates and free credits.