This is the key:
used to be part
It is important to look at this from a historical point of view. Up until the 1950s and early 1960s, there was no agreement to how granites form. This became the granite controversy.
On the one hand, experimental work showed that granites crystallise from magma, but this magma had to be derived from basalt-like magmas. This led to two problems:
- Where is all the basalt from which the granite was supposed to form?
- What happened to all the rock that was there before the granite (aka the "room problem")?
This led people to think that granites formed by metasomatic replacement of other rocks. Essentially, you have other things than were not granite and then they were transformed to granite through various geological process, without ever crystallising from magmas. This is what people say when they talk about granites as metamorphic rocks.
But, this is not the case and later research demonstrated that granites from by crystallisation of magma and the process of metamorphic formation of granites (by metasomatism, or "granitisation") does not actually exists. There were (and still are) many questions left unanswered on exactly how this happens, but we know that this does happen and granites are indeed igneous rocks. In fact, there's a great paper published just this week in the journal Nature, showing modelling of how some of these things can form.
Nevertheless, granite formed can be sometimes related to metamorphism. Extreme cases of hot metamorphic can cause melting of (previously sedimentary and then metamorphic) rocks to form magma that eventually crystallises as granites. This is known as migmatisation, and is commonly used to explain the widespread occurrence of "leucogranites" in the Himalayas for example.
Additionally, some high-grade metamorphic rocks contain quartz, feldspars, micas, and amphiboles. The metamorphic texture (foliation and lineation) is not always obvious, making it easy to mistake metamorphic gneisses as igneous granites.