Project Overview
What problem does this app solve?
This app was developed when I saw an opportunity to automate part of a company that was being outsourced to a third party.
This company outsources Material Safety Data Sheets(MSDS) that are hard to read or input into the data base like pdfs that are locked or copies of copies...
By using this app instead of outsourcing this work the company you can do the job of 3 people in a matter of a Seconds!
Can this be implemented into a bigger pipeline?
Yes! Due the the modular design of the product it can easily be implemented in a bigger pipeline to serve a bigger purpouse!
Technicalities
How does this app work?
To get started with this app you place all the documents you want to get transcribed into a folder and press start! It doesn't matter if it is a pdf, png, jpeg, etc.. because starting off the API function will parse thru the documents and send them out to the appropriate function to be transcribed. If the doc is an image it will be sent straight to the Tesseract Image scanner, If the doc is a pdf it will check if it can be copied and pasted to save time if not It will use the python library called pdf2image to convert each page of the pdf into images, then end each of those to the Tesseract Image scanner. The the returned results from these functions can be displayed or turned into txt documents.
How can this tool be improved?
The slowest part of this pipeline is the pdf2image function which converts pdfs to images at the rate of about 1 page per second. Finding a faster way to get these pdf pages into the Tesseract model will improve the app's run time during large data inputs.
