Remove false positive on en:eu-agriculture and en:non-eu-agriculture tags

[raphael0202]

We detect "en:eu-agriculture" and "en:non-eu-agriculture" using regex on OCRs, in labels.py. This label can be found just below the european agriculture label (en:eu-organic).

We have a significant fraction of these detections that are false positive, as "eu/non-eu agriculture" is also possible, and the negative lookbehind use in the regex doesn't alway detect the "eu" part. Only "en:eu-non-eu-agriculture" is trustworthy.

These insights are particularly tricky to validate, as you have to look closely at the packaging.
We temporarily disabled the generation of these insights, until we find a proper way to handle it.