docs: remove C# Implementation Considerations sections, clean up temp scripts, reorganize test files

- Remove 'C# Implementation Considerations' sections from 34 indicator .md files - Delete 29 temp PowerShell scripts (_fix_mojibake.ps1, _hex_scan.ps1, etc.) - Move test files into tests/ subdirectories for consistent project structure - Add trader-focused bullet points to indicator documentation
2026-08-25 22:08:05 +00:00 · 2026-03-12 12:34:16 -07:00
parent 8937b0c0fa
commit 060649192f
1149 changed files with 1780 additions and 3316 deletions
@@ -91,111 +91,3 @@ Each SMA component benefits from SIMD prefix-sum optimization:
 | **Skender** | ✅ | Matches composite `SMA(SMA)` logic. |
 | **Tulip** | ✅ | Matches `trima` exactly. |
 | **Ooples** | N/A | Not implemented. |
-
-## C# Implementation Considerations
-
-QuanTAlib's TRIMA uses cascaded SMA composition, achieving O(1) streaming updates by leveraging the O(1) nature of each internal SMA. The implementation demonstrates clean indicator composition:
-
-### Composition Architecture
-
-```csharp
-[SkipLocalsInit]
-public sealed class Trima : AbstractBase
-{
-    private readonly Sma _sma1;
-    private readonly Sma _sma2;
-
-    public Trima(int period)
-    {
-        int p1 = (period + 1) / 2;
-        int p2 = period / 2 + 1;
-
-        _sma1 = new Sma(p1);
-        _sma2 = new Sma(p2);
-    }
-}
-```
-
-TRIMA delegates all complexity to its internal SMA instances. Each SMA maintains its own O(1) running sum, so the cascade is also O(1).
-
-### Key Optimizations
-
-| Technique | Implementation | Benefit |
-| :--- | :--- | :--- |
-| **SMA delegation** | Two internal `Sma` instances | O(1) streaming via running sums |
-| **Zero state** | No additional fields beyond SMAs | Minimal memory footprint |
-| **Inline cascade** | `_sma2.Update(_sma1.Update(input))` | No intermediate allocation |
-| **ArrayPool** | Batch uses rented buffer for SMA1 output | Zero allocation in batch mode |
-| **Warmup composition** | `WarmupPeriod = p1 + p2 - 1` | Correct cascaded warmup |
-
-### Streaming Update
-
-```csharp
-[MethodImpl(MethodImplOptions.AggressiveInlining)]
-public override TValue Update(TValue input, bool isNew = true)
-{
-    _isNew = isNew;
-    TValue v1 = _sma1.Update(input, isNew);
-    TValue v2 = _sma2.Update(v1, isNew);
-
-    Last = v2;
-    PubEvent(Last, isNew);
-    return Last;
-}
-```
-
-The `isNew` flag propagates through both SMAs, enabling bar correction at both levels.
-
-### Memory Layout
-
-| Field | Type | Size | Purpose |
-| :--- | :--- | :---: | :--- |
-| `_period` | int | 4 bytes | Original period |
-| `_sma1` | Sma | ~48 + 8×P₁ bytes | First smoothing stage |
-| `_sma2` | Sma | ~48 + 8×P₂ bytes | Second smoothing stage |
-| `_handler` | delegate | 8 bytes | Event handler reference |
-| `_isNew` | bool | 1 byte | Current bar state |
-| **Instance total** | | **~110 + 8N bytes** | N = period |
-
-### Batch Processing
-
-```csharp
-public static void Batch(ReadOnlySpan<double> source, Span<double> output, int period)
-{
-    int p1 = (period + 1) / 2;
-    int p2 = period / 2 + 1;
-
-    double[] tempArray = ArrayPool<double>.Shared.Rent(source.Length);
-    Span<double> tempSpan = tempArray.AsSpan(0, source.Length);
-
-    try
-    {
-        Sma.Batch(source, tempSpan, p1);   // First SMA pass
-        Sma.Batch(tempSpan, output, p2);   // Second SMA pass
-    }
-    finally
-    {
-        ArrayPool<double>.Shared.Return(tempArray);
-    }
-}
-```
-
-Uses ArrayPool for the intermediate buffer to avoid heap allocation per batch.
-
-### Bar Correction Propagation
-
-The `isNew` flag propagates through both internal SMAs:
-
-```csharp
-// isNew=false triggers rollback in BOTH SMAs
-TValue v1 = _sma1.Update(input, isNew);  // SMA1 rolls back its running sum
-TValue v2 = _sma2.Update(v1, isNew);     // SMA2 rolls back based on corrected SMA1 output
-```
-
-This ensures consistent bar correction across the entire cascade.
-
-### Common Pitfalls
-
-1. **Lag**: TRIMA has more lag than SMA, EMA, or WMA. It is a lagging indicator, not a leading one.
-2. **Signal Generation**: Due to its lag, TRIMA is poor for crossover signals. It is best used for visual trend identification or as a baseline for envelopes (e.g., TMA Bands).
-3. **Even/Odd Periods**: The exact calculation of $P_1$ and $P_2$ differs slightly between implementations for even periods. QuanTAlib matches the standard definition used by TA-Lib.