Stack Update 2.7 - Performance focus

About the V2.7

In continuous improvement of our services, we inform you that an update is planned for the 4th of June 2020. There should not be any interruption in our current system. 

Among many improvements, this 2.7 version redesign the semantic extraction process. This improve our performances on several points and fixes some issues you reported. It also opens the door to adding new features in a much simpler and faster way.

• Improvements:
  • The detection of a product line in a receipt is more precise
  • Better extraction of the buying time
  • Better detection of the total amount
  • The database of points of sale in Spain has been largely widened
  • Store's sign are more efficiently detected
• Bugs correction:
  • Detection of anomaly in quantity: a pack of 200 tissues could be marked as a bundle of 200 packs. Now this value is at 1 per default if the detected value seems harebrained.
  • Date: Months cannot have more than 31 days now.
  • If there is no product in a receipt, it is considered as not exploitable.

As we have entirely rebuilt the semantic extraction process, there may be some unexpected behaviors over the next few weeks. Do not hesitate to report any issue or unexpected behavior at [e-mail].

The data output remain the same, but we still updated the API documentation to correct some mistakes.

Check the routes you use here.