advance next to qa

.gitignore

@@ -12,7 +12,6 @@
 /build
 /dist
 /python/imars3d.egg-info
-conda.recipe/
 
 # temp files and dirs
 _tmp

.pre-commit-config.yaml

@@ -3,7 +3,7 @@
 
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.5.0
+    rev: v4.6.0
     hooks:
       - id: trailing-whitespace
       - id: check-docstring-first
@@ -25,7 +25,7 @@ repos:
       - id: end-of-file-fixer
       - id: sort-simple-yaml
   - repo: https://github.com/psf/black
-    rev: 24.3.0
+    rev: 24.8.0
     hooks:
       - id: black
         args: ['--line-length=119']

.readthedocs.yaml

@@ -1,9 +1,9 @@
 version: 2
 
 build:
-  os: ubuntu-20.04
+  os: "ubuntu-22.04"
   tools:
-    python: "mambaforge-4.10"
+    python: "mambaforge-22.9"
   jobs:
     pre_build:
       - mkdir ~/tmp

conda.recipe/meta.yaml

@@ -47,6 +47,7 @@ requirements:
     - bokeh
     - datashader
     - hvplot
+    - numpy<2
 
 test:
   imports:

environment.yml

@@ -11,11 +11,13 @@ dependencies:
   - astropy
   - tomopy
   - algotom
+  - numpy < 2
   # plot
   - holoviews
   - bokeh
   - datashader
   - hvplot
+  - dask
   # GUI
   - panel
   - param

src/imars3d/backend/corrections/beam_hardening.py

@@ -4,7 +4,7 @@
 import logging
 import param
 import numpy as np
-from imars3d.backend.util.functions import clamp_max_workers
+from imars3d.backend.util.functions import clamp_max_workers, calculate_chunksize
 from multiprocessing.managers import SharedMemoryManager
 from functools import partial
 from tqdm.contrib.concurrent import process_map
@@ -83,6 +83,7 @@ def __call__(self, **params):
                 # mp
                 kwargs = {
                     "max_workers": self.max_workers,
+                    "chunksize": calculate_chunksize(params.arrays.shape[0], self.max_workers),
                     "desc": "denoise_by_bilateral",
                 }
                 if self.tqdm_class:

src/imars3d/backend/corrections/denoise.py

@@ -3,7 +3,7 @@
 """Image noise reduction (denoise) module."""
 import logging
 import param
-from imars3d.backend.util.functions import clamp_max_workers
+from imars3d.backend.util.functions import clamp_max_workers, calculate_chunksize
 import numpy as np
 import tomopy
 from multiprocessing.managers import SharedMemoryManager
@@ -153,6 +153,7 @@ def denoise_by_bilateral(
             # mp
             kwargs = {
                 "max_workers": max_workers,
+                "chunksize": calculate_chunksize(arrays.shape[0], max_workers),
                 "desc": "denoise_by_bilateral",
             }
             if tqdm_class:

src/imars3d/backend/corrections/intensity_fluctuation_correction.py

@@ -2,7 +2,7 @@
 # -*- coding: utf-8 -*-
 """iMars3D's intensity fluctuation correction module."""
 import logging
-from imars3d.backend.util.functions import clamp_max_workers
+from imars3d.backend.util.functions import clamp_max_workers, calculate_chunksize
 import numpy as np
 import param
 import tomopy
@@ -93,6 +93,7 @@ def _intensity_fluctuation_correction(self, ct, air_pixels, sigma, max_workers,
                 # map the multiprocessing calls
                 kwargs = {
                     "max_workers": max_workers,
+                    "chunksize": calculate_chunksize(ct.shape[0], max_workers),
                     "desc": "intensity_fluctuation_correction",
                 }
                 if tqdm_class:

src/imars3d/backend/corrections/ring_removal.py

@@ -3,7 +3,7 @@
 """iMars3D's ring artifact correction module."""
 import logging
 import param
-from imars3d.backend.util.functions import clamp_max_workers
+from imars3d.backend.util.functions import clamp_max_workers, calculate_chunksize
 import scipy
 import numpy as np
 
@@ -238,6 +238,7 @@ def _remove_ring_artifact(
             # invoke mp via tqdm wrapper
             kwargs = {
                 "max_workers": max_workers,
+                "chunksize": calculate_chunksize(arrays.shape[1], max_workers),
                 "desc": "Removing ring artifact",
             }
             if tqdm_class:

src/imars3d/backend/dataio/data.py

@@ -3,7 +3,7 @@
 
 # package imports
 from imars3d.backend.dataio.metadata import MetaData
-from imars3d.backend.util.functions import clamp_max_workers, to_time_str
+from imars3d.backend.util.functions import clamp_max_workers, to_time_str, calculate_chunksize
 
 # third party imports
 import numpy as np
@@ -100,7 +100,7 @@ class load_data(param.ParameterizedFunction):
     dc_fnmatch: Optional[str]
         Unix shells-style wild card (``*.tiff``) for selecting dark current
     max_workers: Optional[int]
-        maximum number of processes allowed during loading, default to use as many as possible.
+        maximum number of processes allowed during loading, default to use a single core.
     tqdm_class: panel.widgets.Tqdm
         Class to be used for rendering tqdm progress
 
@@ -125,6 +125,9 @@ class load_data(param.ParameterizedFunction):
 
         Currently, we are using a forgiving reader to load the image where a corrupted file
         will not block reading other data.
+
+        The rotation angles are extracted from the filenames if possible, otherwise from the
+        metadata embedded in the tiff files. If both failed, the angle will be set to None.
     """
 
     #
@@ -296,7 +299,17 @@ def _load_images(filelist: List[str], desc: str, max_workers: int, tqdm_class) -
     file_ext = Path(filelist[0]).suffix.lower()
     if file_ext in (".tif", ".tiff"):
         # use tifffile directly for a faster loading
-        reader = partial(tifffile.imread, out="memmap")
+        # NOTE: Test conducted on 09-05-2024 on bl10-analysis1 shows that using
+        #       memmap is faster, which contradicts the observation from the instrument
+        #       team.
+        #                       | Method | Time (s) |
+        #                       |--------|----------|
+        #                       | `imread(out="memmap")` | 2.62 s ± 24.6 ms |
+        #                       | `imread()` | 3.59 s ± 13.6 ms |
+        #       The `memmap` option is removed until we have a better understanding of the
+        #       discrepancy.
+        # reader = partial(tifffile.imread, out="memmap")
+        reader = tifffile.imread
     elif file_ext == ".fits":
         reader = dxchange.read_fits
     else:
@@ -316,6 +329,7 @@ def _load_images(filelist: List[str], desc: str, max_workers: int, tqdm_class) -
         #       - there are a lot of cores available
         kwargs = {
             "max_workers": max_workers,
+            "chunksize": calculate_chunksize(len(filelist), max_workers),
             "desc": desc,
         }
         rst = process_map(partial(_forgiving_reader, reader=reader), filelist, **kwargs)
@@ -534,7 +548,7 @@ def _get_filelist_by_dir(
 def _extract_rotation_angles(
     filelist: List[str],
     metadata_idx: int = 65039,
-) -> np.ndarray:
+) -> Optional[np.ndarray]:
     """
     Extract rotation angles in degrees from filename or metadata.
 
@@ -548,40 +562,106 @@ def _extract_rotation_angles(
     Returns
     -------
         rotation_angles
+            Array of rotation angles if successfully extracted, None otherwise.
     """
     # sanity check
-    if filelist == []:
+    if not filelist:
         logger.error("filelist is [].")
         raise ValueError("filelist cannot be empty list.")
 
-    # extract rotation angles from file names
+    # process one file at a time
+    rotation_angles = []
+    for filename in filelist:
+        file_ext = Path(filename).suffix.lower()
+        angle = None
+        if file_ext == ".tiff":
+            # first, let's try to extract the angle from the filename
+            angle = extract_rotation_angle_from_filename(filename)
+            if angle is None:
+                # if failed, try to extract from metadata
+                angle = extract_rotation_angle_from_tiff_metadata(filename, metadata_idx)
+            if angle is None:
+                # if failed, log a warning and move on
+                logger.warning(f"Failed to extract rotation angle from {filename}.")
+        elif file_ext in (".tif", ".fits"):
+            # for tif and fits, we can only extract from filename as the metadata is not reliable
+            angle = extract_rotation_angle_from_filename(filename)
+            if angle is None:
+                # if failed, log a warning and move on
+                logger.warning(f"Failed to extract rotation angle from {filename}.")
+        else:
+            # if the file type is not supported, raise value error
+            logger.error(f"Unsupported file type: {file_ext}")
+            raise ValueError(f"Unsupported file type: {file_ext}")
+
+        rotation_angles.append(angle)
+
+    # this means we have a list of None
+    if all(angle is None for angle in rotation_angles):
+        logger.warning("Failed to extract any rotation angles.")
+        return None
+
+    # warn users if some angles are missing
+    if any(angle is None for angle in rotation_angles):
+        logger.warning("Some rotation angles are missing. You will see nan in the rotation angles array.")
+
+    return np.array(rotation_angles, dtype=float)
+
+
+def extract_rotation_angle_from_filename(filename: str) -> Optional[float]:
+    """
+    Extract rotation angle in degrees from filename.
+
+    Parameters
+    ----------
+    filename:
+        Filename to extract rotation angle from.
+
+    Returns
+    -------
+        rotation_angle
+            Rotation angle in degrees if successfully extracted, None otherwise.
+    """
+    # extract rotation angle from file names
     # Note
     # ----
     #   For the following file
-    #       20191030_ironman_small_0070_300_440_0520.tiff
+    #       20191030_ironman_small_0070_300_440_0520.tif(f)
+    #       20191030_ironman_small_0070_300_440_0520.fits
     #   the rotation angle is 300.44 degrees
-    # If all given filenames follows the pattern, we will use the angles from
-    # filenames. Otherwise, we will use the angles from metadata.
-    regex = r"\d{8}_\S*_\d{4}_(?P<deg>\d{3})_(?P<dec>\d{3})_\d*\.tiff"
-    matches = [re.match(regex, Path(f).name) for f in filelist]
-    if all(matches):
-        logger.info("Using rotation angles from filenames.")
-        rotation_angles = np.array([float(".".join(m.groups())) for m in matches])
+    regex = r"\d{8}_\S*_\d{4}_(?P<deg>\d{3})_(?P<dec>\d{3})_\d*\.(?:tiff?|fits)"
+    match = re.match(regex, Path(filename).name)
+    if match:
+        rotation_angle = float(".".join(match.groups()))
     else:
-        # extract rotation angles from metadata
-        file_ext = set([Path(f).suffix for f in filelist])
-        if file_ext != {".tiff"}:
-            logger.error("Only tiff files are supported.")
-            raise ValueError("Rotation angle from metadata is only supported for Tiff.")
+        rotation_angle = None
+    return rotation_angle
+
+
+def extract_rotation_angle_from_tiff_metadata(filename: str, metadata_idx: int = 65039) -> Optional[float]:
+    """
+    Extract rotation angle in degrees from metadata of a tiff file.
+
+    Parameters
+    ----------
+    filename:
+        Filename to extract rotation angle from.
+    metadata_idx:
+        Index of metadata to extract rotation angle from, default is 65039.
+
+    Returns
+    -------
+        rotation_angle
+            Rotation angle in degrees if successfully extracted, None otherwise.
+    """
+    try:
         # -- read metadata
         # img = tifffile.TiffFile("test_with_metadata_0.tiff")
         # img.pages[0].tags[65039].value
         # >> 'RotationActual:0.579840'
-        rotation_angles = np.array(
-            [float(tifffile.TiffFile(f).pages[0].tags[metadata_idx].value.split(":")[-1]) for f in filelist],
-            dtype="float",
-        )
-    return rotation_angles
+        return float(tifffile.TiffFile(filename).pages[0].tags[metadata_idx].value.split(":")[-1])
+    except Exception:
+        return None
 
 
 def _save_data(filename: Path, data: np.ndarray, rot_angles: np.ndarray = None) -> None:

src/imars3d/backend/diagnostics/rotation.py

@@ -5,7 +5,7 @@
 import numpy as np
 
 import param
-from imars3d.backend.util.functions import clamp_max_workers
+from imars3d.backend.util.functions import clamp_max_workers, calculate_chunksize
 from multiprocessing.managers import SharedMemoryManager
 from tqdm.contrib.concurrent import process_map
 from tomopy.recon.rotation import find_center_pc
@@ -139,6 +139,7 @@ def _find_rotation_center(
             # map the multiprocessing calls
             kwargs = {
                 "max_workers": max_workers,
+                "chunksize": calculate_chunksize(len(idx_low), max_workers),
                 "desc": "Finding rotation center",
             }
             if tqdm_class: