sxs 2024.0.33__tar.gz → 2024.0.35__tar.gz

{sxs-2024.0.33 → sxs-2024.0.35}/CITATION.cff

@@ -10,5 +10,5 @@ authors:
 title: "The sxs package"
 license: MIT
 doi: 10.5281/zenodo.4034006
-version: 2024.0.33
+version: 2024.0.35
 date-released: 2025-02-19

{sxs-2024.0.33 → sxs-2024.0.35}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sxs
-Version: 2024.0.33
+Version: 2024.0.35
 Summary: Interface to data produced by the Simulating eXtreme Spacetimes collaboration
 Project-URL: Homepage, https://github.com/sxs-collaboration/sxs
 Project-URL: Documentation, https://sxs.readthedocs.io/

sxs-2024.0.35/sxs/__version__.py

@@ -0,0 +1 @@
+__version__ = "2024.0.35"

{sxs-2024.0.33 → sxs-2024.0.35}/sxs/simulations/local.py

@@ -65,7 +65,7 @@ def extract_id_from_common_metadata(file, annex_dir):
     return key
 
 
-def local_simulations(annex_dir, compute_md5=False):
+def local_simulations(annex_dir, compute_md5=False, show_progress=False):
     """
     Walk the annex directory to find and process all simulations
 
@@ -91,6 +91,8 @@ def local_simulations(annex_dir, compute_md5=False):
     compute_md5 : bool, optional
         Whether to compute the MD5 hash of each file.  Default is
         False.
+    show_progress : bool, optional
+        Whether to show a progress bar.  Default is False.
 
     Returns
     -------
@@ -99,10 +101,24 @@ def local_simulations(annex_dir, compute_md5=False):
     """
     from os import walk
     from ..utilities import md5checksum
+    from tqdm import tqdm
 
     simulations = {}
     annex_dir = Path(annex_dir).resolve()
 
+    if show_progress:  # Count the number of common-metadata.txt files
+        num_files = 0
+        for dirpath, dirnames, filenames in walk(annex_dir, topdown=True):
+            if Path(dirpath).name.startswith("."):
+                dirnames[:] = []
+                continue
+            if "common-metadata.txt" in filenames:
+                if not any(d.startswith("Lev") for d in dirnames):
+                    continue
+                num_files += 1
+                dirnames[:] = []
+        progress_bar = tqdm(total=num_files, desc="Processing simulations")
+
     # The `walk` method can be made *much* faster than the `glob` method
     for dirpath, dirnames, filenames in walk(annex_dir, topdown=True):
         dirpath = Path(dirpath)
@@ -116,40 +132,47 @@ def local_simulations(annex_dir, compute_md5=False):
             if not any(d.startswith("Lev") for d in dirnames):
                 continue
 
-            key = extract_id_from_common_metadata(dirpath / "common-metadata.txt", annex_dir)
+            if show_progress:
+                progress_bar.update(1)
 
-            # Find the highest Lev directory and extract the metadata
-            highest_lev = sorted(
-                [d for d in dirnames if d.startswith("Lev")]
-            )[-1]
-            metadata = Metadata.load(dirpath / highest_lev / "metadata")
-            metadata = metadata.add_standard_parameters()
+            try:
+                key = extract_id_from_common_metadata(dirpath / "common-metadata.txt", annex_dir)
 
-            metadata["directory"] = str(dirpath.relative_to(annex_dir))
+                # Find the highest Lev directory and extract the metadata
+                highest_lev = sorted(
+                    [d for d in dirnames if d.startswith("Lev")]
+                )[-1]
+                metadata = Metadata.load(dirpath / highest_lev / "metadata")
+                metadata = metadata.add_standard_parameters()
 
-            metadata["files"] = {
-                path_to_invenio(file.relative_to(dirpath)): {
-                    "link": str(file),
-                    "size": file.stat().st_size,
-                    "checksum": md5checksum(file) if compute_md5 else "",
+                metadata["directory"] = str(dirpath.relative_to(annex_dir))
+
+                metadata["files"] = {
+                    path_to_invenio(file.relative_to(dirpath)): {
+                        "link": str(file),
+                        "size": file.stat().st_size,
+                        "checksum": md5checksum(file) if compute_md5 else "",
+                    }
+                    for file in files_to_upload(dirpath, annex_dir)
                 }
-                for file in files_to_upload(dirpath, annex_dir)
-            }
 
-            simulations[key] = metadata
+                simulations[key] = metadata
+            except KeyboardInterrupt:
+                raise
+            except Exception as e:
+                print(f"Error processing {dirpath}: {e}")
 
             dirnames[:] = []  # Don't keep looking for common-metadata.txt files under this directory
 
     return simulations
 
 
-def write_local_simulations(annex_dir, output_file=None):
+def write_local_simulations(annex_dir, output_file=None, compute_md5=False, show_progress=False):
     """Write the local simulations to a file for use when loading `Simulations`
 
     This function calls `local_simulations` to obtain the dictionary,
     but also writes the dictionary to a JSON file.
 
-
     Parameters
     ----------
     annex_dir : (str or Path)
@@ -159,6 +182,11 @@ def write_local_simulations(annex_dir, output_file=None):
         written to `sxs_directory("cache") / "local_simulations.json"`.
         N.B.: If you specify a different file, `sxs.load` will not
         automatically find it.
+    compute_md5 : bool, optional
+        Whether to compute the MD5 hash of each file.  Default is
+        False.
+    show_progress : bool, optional
+        Whether to show a progress bar.  Default is False.
 
     Returns
     -------
@@ -168,7 +196,7 @@ def write_local_simulations(annex_dir, output_file=None):
     from json import dump
 
     # Process the annex directory to find all simulations
-    simulations = local_simulations(annex_dir)
+    simulations = local_simulations(annex_dir, compute_md5=compute_md5, show_progress=show_progress)
 
     # Write the simulations to file
     if output_file is None:

sxs-2024.0.33/sxs/__version__.py

@@ -1 +0,0 @@
-__version__ = "2024.0.33"