Scanned invoices are extracted using image processing to reduce the non-reliability and man-power in calculating payments and invoice bills. Additionally, the whole payment process is automated with less turn around time and more flexibility in the invoice templates.
- Accurate Prediction
- Reducing cost of time and space
- Automating changes and cycles
- Reliability of data is ensured after every updation
- Ease of access
- Standard Invoice Format ensured
- Scalable to any structured and unstructured documents
A detailed explanation of the CODE and the EXECUTION is available at E-Invoicing
RECOMMENDED: Running in GOOGLE COLAB provides faster results and interactivity
/flask.py
/pdfconvert.py
/digital_processing.py
/table_extraction.py
/text_extraction.py
/templates
/index.html
/static
/index.css
/output
/dataset1/
invoice.pdf
static
&templates
are required for FLASK operation.- Place the TEST INVOICE inside
output/dataset1
. .py
files in the root folder are responsible for the extraction processes.- At any time during the entire running of the program, intermediate O/P's are present at
output/dataset1
. - Every
function
present in the CODE has a clearDOCSTRING
attached to it which can be called usinghelp(function-name)
(or)function-name.__doc__
You can either directly mount this Colab drive link in your colab, change path according to the project structure from your GDrive in the code and run flask.py or follow the below mentioned steps.
- Ensure the above directory structure is maintained by moving all the 5
.ipynb
modules and 2 foldersstatic
andtemplates
. - Connect to the runtime environment and mount the GDrive.
RUN ALL Cells
inflask.ipynb
to fire up the WEB SERVER ๐- Upload the invoice
- Initially the
CONVERT
button is blocked and afterpre-processing
it is enabled. - The final O/P's are available at
output/dataset1
including pre-processing steps. (In case the process is slow please refer to this directory) - Finally
.zip
containing all the required information is available atoutput/dt1.zip
- Clone the repo and extract the
localhost
code to a seperate folder. pip install -r requirements.txt
located here- Run
python index.py
inside the directory and you're good to go ๐ - Upload the invoice
- Initially the
CONVERT
button is blocked and afterpre-processing
it is enabled. - The final O/P's are available at
output/dataset1
including pre-processing steps. (In case the process is slow please refer to this directory) - Finally
.zip
containing all the required information is available atoutput/dt1.zip
If there are any issues, I request you to mailto