Open tomcruise81 opened 3 years ago
Merging #33 (61a09a5) into main (e87ec2e) will increase coverage by
0.02%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## main #33 +/- ##
==========================================
+ Coverage 99.29% 99.31% +0.02%
==========================================
Files 35 35
Lines 566 587 +21
==========================================
+ Hits 562 583 +21
Misses 2 2
Partials 2 2
Impacted Files | Coverage Δ | |
---|---|---|
indicator_basic.go | 100.00% <ø> (ø) |
|
indicator_cci.go | 100.00% <ø> (ø) |
|
indicator_constant.go | 100.00% <ø> (ø) |
|
indicator_fixed.go | 100.00% <ø> (ø) |
|
indicator_aroon.go | 93.54% <100.00%> (+0.21%) |
:arrow_up: |
indicator_bollinger_band.go | 100.00% <100.00%> (ø) |
|
indicator_derivative.go | 100.00% <100.00%> (ø) |
|
indicator_difference.go | 100.00% <100.00%> (ø) |
|
indicator_maximum_drawdown.go | 100.00% <100.00%> (ø) |
|
indicator_maximum_value.go | 100.00% <100.00%> (ø) |
|
... and 8 more |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update e87ec2e...61a09a5. Read the comment docs.
Hey @tomcruise81! Love the idea. Check out #30 which I just merged. I think if we export the new cachedIndicator
type, and you add this method there, and then change the return types of all cached indicators to the new type, we can accomplish this without changing the base type or breaking the API. Lmk what you think
@sdcoffey - I think we'd still need to augment the other types due to the need for recursively removing the cached calculations from inner indicators. So even if I do something like:
type CachedIndicator interface {
Indicator
cache() resultCache
setCache(cache resultCache)
windowSize() int
}
func RemoveCachedEntry(indicator Indicator, index int) {
cachedIndicator, ok := indicator.(CachedIndicator)
if ok {
cacheResult(cachedIndicator, index, nil)
}
for _, subIndicator := range indicator.subIndicators() {
RemoveCachedEntry(subIndicator, index)
}
}
we'd need to change to something like indicator.subIndicators()
for the recursive part...
...although I may be able to create another interface:
type NestedIndicator interface {
Indicator
nestedIndicators() []Indicator
}
func RemoveCachedEntry(indicator Indicator, index int) {
cachedIndicator, ok := indicator.(CachedIndicator)
if ok {
cacheResult(cachedIndicator, index, nil)
}
// Recursion
nestedIndicator, ok := indicator.(NestedIndicator)
if ok {
for _, nestedIndicator := range nestedIndicator.nestedIndicators() {
cachedNestedIndicator, ok := nestedIndicator.(CachedIndicator)
if ok {
RemoveCachedEntry(cachedNestedIndicator, index)
}
}
}
}
Not sure if this would be something that you're interested in, but I've found benefit from using the MACD Histogram with minute candles, where the last candle is updated based on sub-minute streamed values - i.e. the last candle is an approximation, but still provides value. It requires the ability to re-calculate (i.e. not use the cached value).