Use SeriesBuffer everywhere

[ferranbt]

All the historical time series are capped at MAX_LOOKBACK (around 5000 bars). Most of the time series use the SeriesBuffer struct which uses an efficient circular buffer with Vecqueue. However, some other parts still used a plain Vec<> and inefficient functions to cap the vector.

This PR uses SeriesBuffer to represent all historical time series. The benchmark CI test might fail since I have updated the bench test cases and increase the number of bars being used to MAX_LOOPBACK*2 so it takes longer now. All in all, the improvement is around 5x.

[codspeed-hq]

Merging this PR will not alter performance

✅ 10 untouched benchmarks
🆕 5 new benchmarks
⏩ 64 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
🆕	interpreter/single_bar[complex]	N/A	728.1 ms	N/A
🆕	interpreter/single_bar[macd]	N/A	444.8 ms	N/A
🆕	interpreter/single_bar[moving_averages]	N/A	427.2 ms	N/A
🆕	interpreter/single_bar[rsi]	N/A	299 ms	N/A
🆕	interpreter/single_bar[simple]	N/A	238 ms	N/A

---

_{Comparing series-with-max-loopback (74358ac) with main (89631d6)}

Footnotes

1. 64 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩