Based on #17. Currently we have sped it up to about a second, but only if the difference between each required datapoint is the same. Additionally, it is not completely bug-free, since it can still crash if no data is available. It currently assumes that 4 x as many samples as necessary should be enough, even if some drop out due to being on a weekend or during the night. This is not generally true, so should be improved on.
Ideas:
Check whether a no-trade time is within the required range, and if necessary increase the number of retrieved samples or the start time of retrieval.
Based on #17. Currently we have sped it up to about a second, but only if the difference between each required datapoint is the same. Additionally, it is not completely bug-free, since it can still crash if no data is available. It currently assumes that 4 x as many samples as necessary should be enough, even if some drop out due to being on a weekend or during the night. This is not generally true, so should be improved on. Ideas: