Did these arise from some common source predating both (PIE?), or was the Graeco-Roman week imported into India at some later point in time?
The ancient Hindus did not regularly use "days of the week" (although they are attested). The reason for this was that the Hindu calendar before 1100 AD used mean times (called madhyama) and this can shift days from one month to another. They did have a division into days assigned as one day to each planet as follows:
This is only one possible naming scheme, and many others can be found in ancient Indian writings.
More important among the ancient Hindus was the lunisolar calendar in which each day was numbered, similar to the method of the Romans. Like the Romans the month was divided into two 15-day fortnights, the first being suddha (waxing) and the second being bahula (waning).
Note that there is no regular calendar in India, but many of them in common use for various purposes. It is estimated that today there are about 30 different commonly used calendars in India.
To the extent that the Indians use the Norse names of the week is due to the influence of the Portuguese and English.
One look at that chart you link to makes clear that the author is deficient in understanding Norse mythology:
That's three of the seven days with a serious issue on the claimed equivalence. The Sun and Moon are often seen as celestial book-ends, so putting them at the start and end of the week may simply be an extension of that. I don't see any need to invoke more human rationalization for the other claimed equivalences.