Challenges

1. First we faced problem for creating ML model which can extract text from bill invoice we used ocr technology but had problem with accuracy. So study how Indian government allot GST number found the stratergy by using regular expression we can extract number of 14 digit following certain pattern. All over India 14th character will be Z only many more like this.

2. Using api only government has database of all user with valid gst number. We joined to GST Developer Site which was basicly created for developers like us. But they take 15 days for verification so search for another alternative like webscrapping, third party api. We used api to do that.

3. For hosting we don't have interanational credit card so used Python sever where we can freely deploy ml model and all python flask app. Hosted using git actions.

4. Data quality issues: GST validation relies heavily on accurate and complete data, and any errors or inconsistencies in the data can affect the accuracy of the validation process. Data quality issues can include missing or incorrect data, duplicate entries, and inconsistencies in data formatting.

5. Incomplete or delayed data: In some cases, the data required for GST validation may be incomplete or delayed, which can affect the accuracy of the validation process. This can happen if data is not entered in a timely manner, or if data is missing due to technical issues or other factors.