A fairly accurate description of the problem and its solving, when same word stay in text in various forms, using an example:
1. addon for chrome named vocabsplitter (www.vocabsplitter.com )
https://www.reddit.com/r/languagelearning/comments/dbm6w2/need_some_opinions_on_vocab_splitter_a_tool/
…"fais", "fait", "faisons", " faites" as different words, although their lemma is the same ("faire"). So if we have already added "fais" to our vocabulary, and then later when we encounter the word "faisons" in a text, the tool will treat "faisons" as a different word from "fais" and mark it as "unknown" (red colour), and we have to add "faisons" to our vocabulary again…”
Solving.
“…Finally, I decided to add both "Variant" and "Lemma" modes for the words. In "Lemma" mode, variants like "fais", "fait", "faisons", " faites" will be treated as one word ("faire"), while in "Variant" mode, they will be treated as different words. In order to choose between "Variant" and "Lemma" modes, you will need to go the the "Profile/Settings" page to set the Vocabulary Mode to either "Use Variant" or "Use Lemma" (the default mode is "Use Variant") ..:
2. Solving in addon for chrome named wordology ( https://geajack.github.io/Wordology/):
Use Smart Matching modes (prefix and suffix analysis) - heuristic algorithm.
May be not 100 % accuracy but these words in modified forms are marked with an additional color. The user independently decides whether to change the translation.
PS It is quite exhausting and annoying to assign status and color to the same word in various forms.
I prefer "Lemma" mode. (Polish language)