batchalign 0.8.1.post3__tar.gz → 0.8.1.post4__tar.gz

{batchalign-0.8.1.post3/batchalign.egg-info → batchalign-0.8.1.post4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: batchalign
-Version: 0.8.1.post3
+Version: 0.8.1.post4
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu
@@ -48,6 +48,9 @@ Requires-Dist: onnxruntime
 Requires-Dist: certifi>=2025.10.5
 Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"
+Requires-Dist: mypy; extra == "dev"
+Requires-Dist: pyrefly; extra == "dev"
+Requires-Dist: ty; extra == "dev"
 Provides-Extra: train
 Requires-Dist: accelerate~=0.27; extra == "train"
 Provides-Extra: docs

batchalign-0.8.1.post4/batchalign/__main__.py

@@ -0,0 +1,4 @@
+from batchalign.cli.cli import batchalign
+
+if __name__ == "__main__":
+    batchalign()

batchalign-0.8.1.post4/batchalign/cli/bench.py

@@ -0,0 +1,48 @@
+import time
+import os
+from pathlib import Path
+
+import rich_click as click
+from rich.console import Console
+
+from batchalign.cli.dispatch import _dispatch
+
+
+@click.command()
+@click.argument("command", type=click.Choice(["align", "transcribe", "transcribe_s", "morphotag", "translate", "utseg", "benchmark", "opensmile", "coref"]))
+@click.argument("in_dir", type=click.Path(exists=True, file_okay=False))
+@click.argument("out_dir", type=click.Path(exists=True, file_okay=False))
+@click.option("--runs", type=int, default=1, show_default=True, help="Number of benchmark runs.")
+@click.option("--no-pool", is_flag=True, default=False, help="Disable pooled execution for this benchmark run.")
+@click.option("--no-lazy-audio", is_flag=True, default=False, help="Disable lazy audio loading for this benchmark run.")
+@click.option("--no-adaptive-workers", is_flag=True, default=False, help="Disable adaptive worker caps for this benchmark run.")
+@click.option("--workers", type=int, default=None, help="Number of workers to use (defaults to CPU count).")
+@click.pass_context
+def bench(ctx, command, in_dir, out_dir, runs, no_pool, no_lazy_audio, no_adaptive_workers, workers):
+    """Benchmark Batchalign command performance on a dataset."""
+    console = Console()
+    durations = []
+    for idx in range(runs):
+        run_ctx = type("Ctx", (), {"obj": dict(ctx.obj)})()
+        if no_pool:
+            run_ctx.obj["pool"] = False
+        if no_lazy_audio:
+            run_ctx.obj["lazy_audio"] = False
+        if no_adaptive_workers:
+            run_ctx.obj["adaptive_workers"] = False
+        if workers is not None:
+            run_ctx.obj["workers"] = workers
+        start = time.time()
+        if command in ["align", "morphotag", "translate", "utseg", "coref"]:
+            extensions = ["cha"]
+        elif command in ["transcribe", "transcribe_s", "benchmark", "opensmile"]:
+            extensions = ["wav", "mp3", "mp4"]
+        else:
+            extensions = ["cha"]
+        _dispatch(command, "eng", 1, extensions, run_ctx,
+                  in_dir, out_dir, None, None, console)
+        durations.append(time.time() - start)
+        console.print(f"[dim]Run {idx+1}/{runs}:[/dim] {durations[-1]:.2f}s")
+    if durations:
+        avg = sum(durations) / len(durations)
+        console.print(f"\n[bold]Average:[/bold] {avg:.2f}s over {len(durations)} run(s)")

{batchalign-0.8.1.post3 → batchalign-0.8.1.post4}/batchalign/cli/cli.py

@@ -16,6 +16,7 @@ from rich.console import Console
 from rich.logging import RichHandler
 
 from batchalign.cli.dispatch import _dispatch
+from batchalign.cli.bench import bench
 from batchalign.models.training.run import cli as train
 
 import pyfiglet
@@ -68,8 +69,17 @@ def handle_verbosity(verbosity):
 @click.pass_context
 @click.version_option(VERSION_NUMBER)
 @click.option("-v", "--verbose", type=int, count=True, default=0, help="How loquacious Batchalign should be.")
-@click.option("--workers", type=int, default=os.cpu_count(), help="Number of worker processes to use.")
-def batchalign(ctx, verbose, workers):
+@click.option("--workers", type=int, default=os.cpu_count(), help="Number of workers (threads when pooled, processes otherwise).")
+@click.option("--memlog", is_flag=True, default=False, help="Write memory telemetry log to OUT_DIR.")
+@click.option("--mem-guard", is_flag=True, default=False, help="Abort early if available memory is too low to start a new worker.")
+@click.option("--adaptive-workers/--no-adaptive-workers", default=True, help="Adaptively cap workers based on observed memory.")
+@click.option("--pool/--no-pool", default=True, help="Allow pooled model execution for multi-file runs.")
+@click.option("--lazy-audio/--no-lazy-audio", default=True, help="Enable lazy audio loading for alignment/ASR.")
+@click.option("--adaptive-safety-factor", type=float, default=1.35, show_default=True, help="Safety factor applied to observed worker RSS peaks.")
+@click.option("--adaptive-warmup", type=int, default=2, show_default=True, help="Initial worker count before adaptive cap kicks in.")
+@click.option("--force-cpu/--no-force-cpu", default=False, help="Disable MPS/CUDA and force CPU-only models.")
+@click.option("--shared-models/--no-shared-models", default=False, help="Preload models and fork workers to share read-only memory (advanced; ignored in pooled mode).")
+def batchalign(ctx, verbose, workers, memlog, mem_guard, adaptive_workers, adaptive_safety_factor, adaptive_warmup, pool, lazy_audio, force_cpu, shared_models):
     """process .cha and/or audio files in IN_DIR and dumps them to OUT_DIR using recipe COMMAND"""
 
     ## setup commands ##
@@ -82,6 +92,15 @@ def batchalign(ctx, verbose, workers):
     # add to arguments
     ctx.obj["verbose"] = verbose
     ctx.obj["workers"] = workers
+    ctx.obj["memlog"] = memlog
+    ctx.obj["mem_guard"] = mem_guard
+    ctx.obj["adaptive_workers"] = adaptive_workers
+    ctx.obj["adaptive_safety_factor"] = adaptive_safety_factor
+    ctx.obj["adaptive_warmup"] = adaptive_warmup
+    ctx.obj["pool"] = pool
+    ctx.obj["lazy_audio"] = lazy_audio
+    ctx.obj["force_cpu"] = force_cpu
+    ctx.obj["shared_models"] = shared_models
     # setup config
     from batchalign.utils import config
     ctx.obj["config"] = config.config_read(True)
@@ -91,6 +110,7 @@ def batchalign(ctx, verbose, workers):
     install()
 
 batchalign.add_command(train, "models")
+batchalign.add_command(bench, "bench")
 
 from batchalign.cli.cache import cache
 batchalign.add_command(cache, "cache")