scripts/win_pull_and_run_next_batch.ps1

param(
    [switch]$DryRun,
    [int]$Epochs = 20,
    [int]$BatchSizePerGpu = 600,
    [string]$HfCacheMode = "auto",
    [string]$HfCacheDir = ".hf_cache",
    [string]$MachineTag = "win"
)

$ErrorActionPreference = "Stop"

$repoRoot = Resolve-Path (Join-Path $PSScriptRoot "..")
Set-Location $repoRoot

git pull --ff-only origin main

$prepared = Get-ChildItem -Path "data" -Directory -ErrorAction SilentlyContinue |
    Where-Object { $_.Name -match '^sinhala_tts_batch(\d+)_custom$' } |
    Where-Object { Test-Path (Join-Path $_.FullName "raw.arrow") } |
    ForEach-Object {
        [PSCustomObject]@{
            Number = [int]([regex]::Match($_.Name, '^sinhala_tts_batch(\d+)_custom$').Groups[1].Value)
            DatasetName = "sinhala_tts_batch$([regex]::Match($_.Name, '^sinhala_tts_batch(\d+)_custom$').Groups[1].Value)"
        }
    } |
    Sort-Object Number

if (-not $prepared) {
    throw "No prepared batches found under data\\sinhala_tts_batch*_custom"
}

$lockFile = Join-Path $repoRoot "ACTIVE_RUNS.md"
$lockContent = if (Test-Path $lockFile) { Get-Content $lockFile -Raw } else { "" }

$selected = $null
foreach ($p in $prepared) {
    if ($lockContent -match [regex]::Escape($p.DatasetName)) {
        continue
    }
    $selected = $p
    break
}

if (-not $selected) {
    throw "All available prepared batches appear claimed in ACTIVE_RUNS.md"
}

$datasetName = $selected.DatasetName
$expName = "F5TTS_v1_Base_${MachineTag}_batch$($selected.Number)"
$startedUtc = (Get-Date).ToUniversalTime().ToString("yyyy-MM-ddTHH:mm:ssZ")

if (-not (Test-Path $lockFile)) {
    @(
        "# ACTIVE RUNS",
        "# machine | dataset_name | exp_name | started_utc"
    ) | Set-Content -LiteralPath $lockFile -Encoding UTF8
}

Add-Content -LiteralPath $lockFile -Value ("| {0} | {1} | {2} | {3} |" -f $MachineTag, $datasetName, $expName, $startedUtc)

$cmd = @(
    ".\scripts\run_finetune_safe.ps1",
    "-DatasetName", $datasetName,
    "-ExpName", $expName,
    "-Epochs", "$Epochs",
    "-BatchSizePerGpu", "$BatchSizePerGpu",
    "-HfCacheMode", $HfCacheMode,
    "-HfCacheDir", $HfCacheDir
)

Write-Output ("Selected batch: {0}" -f $datasetName)
Write-Output ("Command: {0}" -f ($cmd -join " "))

if ($DryRun) {
    Write-Output "Dry run only; not launching."
    exit 0
}

& .\scripts\run_finetune_safe.ps1 `
    -DatasetName $datasetName `
    -ExpName $expName `
    -Epochs $Epochs `
    -BatchSizePerGpu $BatchSizePerGpu `
    -HfCacheMode $HfCacheMode `
    -HfCacheDir $HfCacheDir