Packaging (tracker bug)

[teolemon]

Autoselect with Regex

• Auto-select recycling photos based on OCR #260
  • Auto-select recycling images (closes #260) #301 (Deployed, but does not seem to work ??)

Extract with Regex

• Extract full recycling instructions with Regex
  • Extract packaging shape with Regex
  • Extract packaging material with Regex
  • Extract packaging shapes and materials with Regex
• Copy producer recycling instructions to our own recycling instructions field
  • Add support for writing to the new packaging field #257

More advanced solutions (openfoodfacts-ai)

• Test detecting packaging with CircularNet models openfoodfacts-ai#207
  • Those are 3 models, probably solving the 3 issues below
  • Detect "shiny" or reflective glass, plastic packaging openfoodfacts-ai#56
  • Recognize packaging type (bottle, bag…) openfoodfacts-ai#29
  • Recognize packaging material openfoodfacts-ai#23
• Detect non-written packaging instructions openfoodfacts-ai#72
  • Those have probably been detected by Logos & Labels, can we make a Hunger Games dashboard out of them ?
• Infer packaging from related products in the category openfoodfacts-ai#55

Part of

• 🎯 What can I work on ? #374
• 📦 Packaging (tracker) openfoodfacts-server#6374
• Extract packaging-related information (Tracker) openfoodfacts-ai#97

[teolemon]

Issues with Regex extraction

https://world.openfoodfacts.org/product/3700007900439/les-tomates-cerises

Packaging: Barquette en plastique, Film en plastique

NET-PLASTIC recycle with carrier\nbags at larger stores- not at kerbside\nTAG-MIXED MATERIAL\nnot currently recycled\n19 Sept\nGrown by\nAntonio Peregrin, Spain\nCount: X4\nR0466B\nL-81-06\n7500735\n0914\nAt home, store in a cool, dry place.

https://world.openfoodfacts.org/images/products/01795257/1.jpg

[raphael0202]

The first bug is fixed: ee71a66.
Can you tell me more about the second image?