setiastrosuitepro 1.6.2.post1__py3-none-any.whl

setiastro/saspro/parallel_utils.py

@@ -0,0 +1,554 @@
+"""
+Parallelization utilities for Seti Astro Suite Pro.
+
+This module provides standardized parallelization helpers to ensure consistent
+behavior across the codebase. It addresses the optimization recommendation #4
+from OTTIMIZZAZIONI_PROPOSTE.md.
+
+Key features:
+- Centralized worker count calculation
+- Consistent ThreadPoolExecutor/ProcessPoolExecutor usage
+- Memory-aware parallelization
+- Task-type specific worker allocation
+
+Usage:
+    from setiastro.saspro.parallel_utils import (
+        get_optimal_workers,
+        get_io_workers,
+        get_cpu_workers,
+        run_parallel,
+        run_in_thread_pool,
+        run_in_process_pool,
+    )
+    
+    # For CPU-bound tasks
+    workers = get_cpu_workers()
+    
+    # For I/O-bound tasks
+    workers = get_io_workers()
+    
+    # Run parallel tasks
+    results = run_in_thread_pool(process_image, images, max_workers=workers)
+"""
+
+from __future__ import annotations
+
+import os
+import sys
+import gc
+import logging
+from concurrent.futures import (
+    ThreadPoolExecutor,
+    ProcessPoolExecutor,
+    Future,
+    as_completed,
+)
+from typing import (
+    Callable,
+    TypeVar,
+    Iterable,
+    Iterator,
+    Optional,
+    Any,
+    List,
+    Tuple,
+    Union,
+)
+from functools import wraps
+from contextlib import contextmanager
+
+logger = logging.getLogger(__name__)
+
+T = TypeVar('T')
+R = TypeVar('R')
+
+
+# ---------------------------------------------------------------------------
+# Constants and Defaults
+# ---------------------------------------------------------------------------
+
+# Maximum workers to prevent over-subscription
+MAX_WORKERS_CAP = 32
+
+# Default I/O worker multiplier (I/O-bound can use more threads)
+IO_WORKER_MULTIPLIER = 2
+
+# Minimum workers
+MIN_WORKERS = 1
+
+# Memory threshold for reducing workers (in bytes, ~1GB)
+MEMORY_THRESHOLD_LOW = 1 * 1024 * 1024 * 1024
+
+
+# ---------------------------------------------------------------------------
+# Worker Count Calculation
+# ---------------------------------------------------------------------------
+
+def get_cpu_count() -> int:
+    """
+    Get the number of available CPU cores.
+    
+    Falls back to a conservative value if detection fails.
+    
+    Returns:
+        Number of CPU cores (at least 1)
+    """
+    try:
+        count = os.cpu_count()
+        return max(1, count) if count else 4
+    except Exception:
+        return 4
+
+
+def get_available_memory() -> int:
+    """
+    Get available system memory in bytes.
+    
+    Returns:
+        Available memory in bytes, or 0 if unable to determine
+    """
+    try:
+        import psutil
+        return psutil.virtual_memory().available
+    except ImportError:
+        return 0
+    except Exception:
+        return 0
+
+
+def is_memory_constrained(threshold: int = MEMORY_THRESHOLD_LOW) -> bool:
+    """
+    Check if the system is running low on memory.
+    
+    Args:
+        threshold: Memory threshold in bytes
+        
+    Returns:
+        True if available memory is below threshold
+    """
+    available = get_available_memory()
+    if available == 0:
+        return False  # Can't determine, assume OK
+    return available < threshold
+
+
+def get_optimal_workers(
+    task_count: Optional[int] = None,
+    max_cap: int = MAX_WORKERS_CAP,
+    min_workers: int = MIN_WORKERS,
+    memory_aware: bool = True,
+    io_bound: bool = False,
+) -> int:
+    """
+    Calculate the optimal number of workers for parallel execution.
+    
+    This function considers:
+    - Available CPU cores
+    - Task count (no point having more workers than tasks)
+    - System memory constraints
+    - Task type (I/O-bound vs CPU-bound)
+    
+    Args:
+        task_count: Number of tasks to process (None for default)
+        max_cap: Maximum workers cap
+        min_workers: Minimum workers to use
+        memory_aware: Reduce workers if memory is low
+        io_bound: True for I/O-bound tasks (allows more workers)
+        
+    Returns:
+        Optimal number of workers
+        
+    Example:
+        >>> workers = get_optimal_workers(task_count=100, io_bound=True)
+        >>> with ThreadPoolExecutor(max_workers=workers) as pool:
+        ...     results = list(pool.map(process, items))
+    """
+    cpu_count = get_cpu_count()
+    
+    # Start with CPU count
+    if io_bound:
+        # I/O-bound tasks can use more threads (they spend time waiting)
+        workers = min(cpu_count * IO_WORKER_MULTIPLIER, max_cap)
+    else:
+        # CPU-bound tasks should not exceed core count
+        workers = min(cpu_count, max_cap)
+    
+    # Don't exceed task count
+    if task_count is not None and task_count > 0:
+        workers = min(workers, task_count)
+    
+    # Reduce if memory constrained
+    if memory_aware and is_memory_constrained():
+        workers = max(min_workers, workers // 2)
+        logger.debug("Memory constrained, reduced workers to %d", workers)
+    
+    # Ensure within bounds
+    return max(min_workers, min(workers, max_cap))
+
+
+def get_cpu_workers(task_count: Optional[int] = None) -> int:
+    """
+    Get optimal worker count for CPU-bound tasks.
+    
+    Args:
+        task_count: Number of tasks to process
+        
+    Returns:
+        Optimal worker count
+    """
+    return get_optimal_workers(
+        task_count=task_count,
+        max_cap=MAX_WORKERS_CAP,
+        io_bound=False,
+    )
+
+
+def get_io_workers(task_count: Optional[int] = None) -> int:
+    """
+    Get optimal worker count for I/O-bound tasks.
+    
+    Args:
+        task_count: Number of tasks to process
+        
+    Returns:
+        Optimal worker count (typically higher than CPU workers)
+    """
+    return get_optimal_workers(
+        task_count=task_count,
+        max_cap=MAX_WORKERS_CAP * 2,  # Allow more for I/O
+        io_bound=True,
+    )
+
+
+def get_image_processing_workers(
+    image_count: int,
+    image_size_mb: float = 100.0,
+) -> int:
+    """
+    Get optimal worker count for image processing tasks.
+    
+    Takes into account typical memory usage per image.
+    
+    Args:
+        image_count: Number of images to process
+        image_size_mb: Estimated size per image in MB
+        
+    Returns:
+        Optimal worker count
+    """
+    cpu_count = get_cpu_count()
+    
+    # Estimate memory needed per worker
+    available_mb = get_available_memory() / (1024 * 1024)
+    if available_mb > 0:
+        # Reserve 2GB for system, rest for workers
+        usable_mb = max(0, available_mb - 2048)
+        memory_limited_workers = max(1, int(usable_mb / image_size_mb))
+    else:
+        memory_limited_workers = cpu_count
+    
+    # Use the more restrictive limit
+    workers = min(cpu_count, memory_limited_workers, image_count)
+    
+    return max(1, workers)
+
+
+# ---------------------------------------------------------------------------
+# Parallel Execution Helpers
+# ---------------------------------------------------------------------------
+
+def run_in_thread_pool(
+    func: Callable[[T], R],
+    items: Iterable[T],
+    max_workers: Optional[int] = None,
+    progress_callback: Optional[Callable[[int, int], None]] = None,
+) -> List[R]:
+    """
+    Execute a function across items using ThreadPoolExecutor.
+    
+    Best for I/O-bound tasks (file operations, network requests).
+    
+    Args:
+        func: Function to apply to each item
+        items: Items to process
+        max_workers: Max parallel workers (auto-calculated if None)
+        progress_callback: Optional callback(completed, total)
+        
+    Returns:
+        List of results in original order
+        
+    Example:
+        >>> def load_image(path):
+        ...     return Image.open(path)
+        >>> images = run_in_thread_pool(load_image, file_paths, max_workers=8)
+    """
+    items_list = list(items)
+    total = len(items_list)
+    
+    if total == 0:
+        return []
+    
+    if max_workers is None:
+        max_workers = get_io_workers(total)
+    
+    results: List[Optional[R]] = [None] * total
+    completed = 0
+    
+    with ThreadPoolExecutor(max_workers=max_workers) as executor:
+        # Submit all tasks with their indices
+        future_to_idx = {
+            executor.submit(func, item): idx
+            for idx, item in enumerate(items_list)
+        }
+        
+        for future in as_completed(future_to_idx):
+            idx = future_to_idx[future]
+            try:
+                results[idx] = future.result()
+            except Exception as e:
+                logger.error("Task %d failed: %s", idx, e)
+                raise
+            
+            completed += 1
+            if progress_callback:
+                progress_callback(completed, total)
+    
+    return results  # type: ignore
+
+
+def run_in_process_pool(
+    func: Callable[[T], R],
+    items: Iterable[T],
+    max_workers: Optional[int] = None,
+    progress_callback: Optional[Callable[[int, int], None]] = None,
+) -> List[R]:
+    """
+    Execute a function across items using ProcessPoolExecutor.
+    
+    Best for CPU-bound tasks that benefit from true parallelism.
+    Note: func and items must be picklable.
+    
+    Args:
+        func: Function to apply to each item (must be picklable)
+        items: Items to process (must be picklable)
+        max_workers: Max parallel workers (auto-calculated if None)
+        progress_callback: Optional callback(completed, total)
+        
+    Returns:
+        List of results in original order
+        
+    Example:
+        >>> def heavy_compute(data):
+        ...     return np.fft.fft2(data)
+        >>> results = run_in_process_pool(heavy_compute, data_arrays)
+    """
+    items_list = list(items)
+    total = len(items_list)
+    
+    if total == 0:
+        return []
+    
+    if max_workers is None:
+        max_workers = get_cpu_workers(total)
+    
+    results: List[Optional[R]] = [None] * total
+    completed = 0
+    
+    with ProcessPoolExecutor(max_workers=max_workers) as executor:
+        future_to_idx = {
+            executor.submit(func, item): idx
+            for idx, item in enumerate(items_list)
+        }
+        
+        for future in as_completed(future_to_idx):
+            idx = future_to_idx[future]
+            try:
+                results[idx] = future.result()
+            except Exception as e:
+                logger.error("Task %d failed: %s", idx, e)
+                raise
+            
+            completed += 1
+            if progress_callback:
+                progress_callback(completed, total)
+    
+    return results  # type: ignore
+
+
+@contextmanager
+def thread_pool(max_workers: Optional[int] = None, io_bound: bool = True):
+    """
+    Context manager for ThreadPoolExecutor with automatic worker count.
+    
+    Args:
+        max_workers: Max workers (auto-calculated if None)
+        io_bound: Whether tasks are I/O-bound
+        
+    Yields:
+        ThreadPoolExecutor instance
+        
+    Example:
+        >>> with thread_pool(io_bound=True) as pool:
+        ...     futures = [pool.submit(load_file, f) for f in files]
+        ...     results = [f.result() for f in futures]
+    """
+    if max_workers is None:
+        max_workers = get_io_workers() if io_bound else get_cpu_workers()
+    
+    executor = ThreadPoolExecutor(max_workers=max_workers)
+    try:
+        yield executor
+    finally:
+        executor.shutdown(wait=True)
+
+
+@contextmanager
+def process_pool(max_workers: Optional[int] = None):
+    """
+    Context manager for ProcessPoolExecutor with automatic worker count.
+    
+    Args:
+        max_workers: Max workers (auto-calculated if None)
+        
+    Yields:
+        ProcessPoolExecutor instance
+        
+    Example:
+        >>> with process_pool() as pool:
+        ...     futures = [pool.submit(compute, data) for data in datasets]
+        ...     results = [f.result() for f in futures]
+    """
+    if max_workers is None:
+        max_workers = get_cpu_workers()
+    
+    executor = ProcessPoolExecutor(max_workers=max_workers)
+    try:
+        yield executor
+    finally:
+        executor.shutdown(wait=True)
+
+
+# ---------------------------------------------------------------------------
+# Batch Processing
+# ---------------------------------------------------------------------------
+
+def batch_items(items: List[T], batch_size: int) -> Iterator[List[T]]:
+    """
+    Split items into batches of specified size.
+    
+    Args:
+        items: List of items to batch
+        batch_size: Maximum items per batch
+        
+    Yields:
+        Batches of items
+        
+    Example:
+        >>> for batch in batch_items(images, batch_size=10):
+        ...     process_batch(batch)
+    """
+    for i in range(0, len(items), batch_size):
+        yield items[i:i + batch_size]
+
+
+def run_in_batches(
+    func: Callable[[List[T]], List[R]],
+    items: List[T],
+    batch_size: int,
+    max_workers: Optional[int] = None,
+    progress_callback: Optional[Callable[[int, int], None]] = None,
+) -> List[R]:
+    """
+    Process items in batches using parallel execution.
+    
+    Useful for reducing overhead when processing many small items.
+    
+    Args:
+        func: Function that processes a batch and returns list of results
+        items: All items to process
+        batch_size: Items per batch
+        max_workers: Max parallel workers
+        progress_callback: Optional callback(completed_items, total_items)
+        
+    Returns:
+        Flattened list of all results
+    """
+    batches = list(batch_items(items, batch_size))
+    total_items = len(items)
+    completed_items = 0
+    all_results: List[R] = []
+    
+    if max_workers is None:
+        max_workers = get_cpu_workers(len(batches))
+    
+    with ThreadPoolExecutor(max_workers=max_workers) as executor:
+        future_to_batch = {
+            executor.submit(func, batch): batch
+            for batch in batches
+        }
+        
+        for future in as_completed(future_to_batch):
+            batch = future_to_batch[future]
+            try:
+                batch_results = future.result()
+                all_results.extend(batch_results)
+            except Exception as e:
+                logger.error("Batch processing failed: %s", e)
+                raise
+            
+            completed_items += len(batch)
+            if progress_callback:
+                progress_callback(completed_items, total_items)
+    
+    return all_results
+
+
+# ---------------------------------------------------------------------------
+# Cleanup and Memory Management
+# ---------------------------------------------------------------------------
+
+def cleanup_after_parallel(force_gc: bool = True) -> None:
+    """
+    Clean up after parallel processing.
+    
+    Should be called after large parallel operations to free memory.
+    
+    Args:
+        force_gc: Whether to force garbage collection
+    """
+    if force_gc:
+        gc.collect()
+
+
+# ---------------------------------------------------------------------------
+# Module exports
+# ---------------------------------------------------------------------------
+
+__all__ = [
+    # Constants
+    'MAX_WORKERS_CAP',
+    'IO_WORKER_MULTIPLIER',
+    'MIN_WORKERS',
+    
+    # Worker calculation
+    'get_cpu_count',
+    'get_available_memory',
+    'is_memory_constrained',
+    'get_optimal_workers',
+    'get_cpu_workers',
+    'get_io_workers',
+    'get_image_processing_workers',
+    
+    # Execution helpers
+    'run_in_thread_pool',
+    'run_in_process_pool',
+    'thread_pool',
+    'process_pool',
+    
+    # Batch processing
+    'batch_items',
+    'run_in_batches',
+    
+    # Cleanup
+    'cleanup_after_parallel',
+]

setiastro/saspro/pedestal.py

@@ -0,0 +1,121 @@
+# pro/pedestal.py
+from __future__ import annotations
+import numpy as np
+
+try:
+    import cv2  # not required
+    # just for mask helpers if you later want it
+except Exception:
+    cv2 = None
+
+
+def _as_float01(img: np.ndarray) -> np.ndarray:
+    """Return float32 image; compress crazy-high float ranges to ~[0..1]."""
+    a = np.asarray(img)
+    if a.dtype != np.float32:
+        a = a.astype(np.float32, copy=False)
+    # If upstream keeps images normalized, this is a noop.
+    # If someone slipped a >1.0 range, compress gently so 'min-subtract' behaves.
+    if a.size:
+        mx = float(a.max())
+        if mx > 5.0:
+            a = a / mx
+    return a
+
+
+def _remove_pedestal_array(a: np.ndarray) -> np.ndarray:
+    """
+    Subtract per-channel minimum. Preserves shape (2D or 3D).
+    Result is clipped to [0,1] for safety.
+    """
+    a = _as_float01(a)
+
+    if a.ndim == 2:
+        mn = float(a.min()) if a.size else 0.0
+        out = a - mn
+    else:
+        # H, W, C
+        out = np.empty_like(a, dtype=np.float32)
+        for c in range(a.shape[2]):
+            ch = a[..., c]
+            mn = float(ch.min()) if ch.size else 0.0
+            out[..., c] = ch - mn
+
+    return np.clip(out, 0.0, 1.0).astype(np.float32, copy=False)
+
+def remove_pedestal(main, target_doc=None):
+    """
+    Subtract per-channel minimum on the *currently active* content:
+    - If a Preview/ROI tab is active, operate on that ROI document.
+    - Otherwise, operate on the full base document.
+    Creates a single undo step via doc.apply_edit (ROI docs manage their
+    own preview undo; base docs go into full undo stack).
+    """
+    doc = target_doc
+
+    # Prefer DocManager's view-aware resolution
+    dm = getattr(main, "doc_manager", None) or getattr(main, "docman", None)
+
+    if doc is None and dm is not None:
+        # 1) Try to resolve from the currently active view (ROI-aware)
+        vw = None
+        try:
+            if hasattr(dm, "_active_view_widget") and callable(dm._active_view_widget):
+                vw = dm._active_view_widget()
+        except Exception:
+            vw = None
+
+        if vw is not None and hasattr(dm, "get_document_for_view"):
+            try:
+                candidate = dm.get_document_for_view(vw)
+                if candidate is not None:
+                    doc = candidate
+            except Exception:
+                doc = None
+
+        # 2) If that failed, fall back to whatever DM thinks is active
+        if doc is None and hasattr(dm, "get_active_document"):
+            try:
+                doc = dm.get_active_document()
+            except Exception:
+                doc = None
+
+    # 3) Last resort: legacy _active_doc on the main window
+    if doc is None:
+        ad = getattr(main, "_active_doc", None)
+        if callable(ad):
+            doc = ad()
+        else:
+            doc = ad
+
+    if doc is None or getattr(doc, "image", None) is None:
+        # Quiet exit if nothing to do.
+        return
+
+    # Optional debug so you can confirm it hits ROI when Preview is active
+    try:
+        is_roi = bool(getattr(doc, "_roi", None)) or bool(getattr(doc, "_roi_info", None))
+        print(f"[Pedestal] target_doc={doc!r}, type={type(doc)}, is_roi={is_roi}")
+    except Exception:
+        pass
+
+    # ---- actual operation ----
+    try:
+        src = np.asarray(doc.image)
+        out = _remove_pedestal_array(src)
+
+        meta = {
+            "step_name": "Pedestal Removal",
+            "bit_depth": "32-bit floating point",
+            "is_mono": (out.ndim == 2),
+        }
+        doc.apply_edit(out, metadata=meta, step_name="Pedestal Removal")
+
+        if hasattr(main, "_log"):
+            main._log("Pedestal Removal")
+    except Exception as e:
+        try:
+            from PyQt6.QtWidgets import QMessageBox
+            QMessageBox.critical(main, "Pedestal Removal", f"Failed:\n{e}")
+        except Exception:
+            pass