seqchromloader 0.6.0__tar.gz → 0.6.2__tar.gz

{seqchromloader-0.6.0 → seqchromloader-0.6.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: seqchromloader
-Version: 0.6.0
+Version: 0.6.2
 Summary: Sequence and chromatin dataloader for deep learning
 Home-page: https://github.com/yztxwd/seqchromloader
 Author-email: yztxwd@gmail.com

{seqchromloader-0.6.0 → seqchromloader-0.6.2}/seqchromloader/loader.py

@@ -116,7 +116,8 @@ class _SeqChromDatasetByDataFrame(Dataset):
                  bigwig_filelist:list, 
                  target_bam=None, 
                  transforms:dict=None, 
-                 initialize_first=False):
+                 initialize_first=False,
+                 return_region=False):
         
         self.dataframe = dataframe        
         self.genome_fasta = genome_fasta
@@ -129,6 +130,8 @@ class _SeqChromDatasetByDataFrame(Dataset):
         self.transforms = transforms
 
         if initialize_first: self.initialize()
+
+        self.return_region = return_region
     
     def initialize(self):
         # create the stream handler after child processes spawned to enable parallel reading
@@ -158,7 +161,10 @@ class _SeqChromDatasetByDataFrame(Dataset):
         except utils.BigWigInaccessible as e:
             raise e
 
-        return feature['seq'], feature['chrom'], feature['target'], feature['label']
+        if not self.return_region:
+            return feature['seq'], feature['chrom'], feature['target'], feature['label']
+        else:
+            return f'{item.chrom}:{item.start}-{item.end}', feature['seq'], feature['chrom'], feature['target'], feature['label']
     
 SeqChromDatasetByDataFrame = seqChromLoaderCurry(_SeqChromDatasetByDataFrame)
 
@@ -175,14 +181,15 @@ class _SeqChromDatasetByBed(_SeqChromDatasetByDataFrame):
     :param transforms: A dictionary of functions to transform the output data, accepted keys are *["seq", "chrom", "target", "label"]*
     :type transforms: dict of functions
     """
-    def __init__(self, bed: str, genome_fasta: str, bigwig_filelist:list, target_bam=None, transforms:dict=None, initialize_first=False):
+    def __init__(self, bed: str, genome_fasta: str, bigwig_filelist:list, target_bam=None, transforms:dict=None, initialize_first=False, return_region=False):
         dataframe = pd.read_table(bed, header=None, names=['chrom', 'start', 'end', 'label', 'score', 'strand' ])
         super().__init__(dataframe,
                          genome_fasta,
                          bigwig_filelist,
                          target_bam,
                          transforms,
-                         initialize_first)
+                         initialize_first,
+                         return_region)
 
 SeqChromDatasetByBed = seqChromLoaderCurry(_SeqChromDatasetByBed)
 

{seqchromloader-0.6.0 → seqchromloader-0.6.2}/seqchromloader/utils.py

@@ -354,9 +354,14 @@ def extract_target(chrom, start, end, strand, target):
     if isinstance(target, pysam.AlignmentFile):
         target_array = np.array(target.count(chrom, start, end), dtype=np.float32)[np.newaxis]
     elif isinstance(target, pyBigWig.pyBigWig):
-        target_array = np.nan_to_num(target.values(chrom, start, end)).astype(np.float32)
-        if strand=="-":
-            target_array = target_array[::-1]
+        try:
+            target_array = np.nan_to_num(target.values(chrom, start, end)).astype(np.float32)
+            if strand=="-":
+                target_array = target_array[::-1]
+        except RuntimeError as e:
+            logging.warning(e)
+            logging.warning(f"RuntimeError happened when accessing {chrom}:{start}-{end}, it's probably due to at least one chromatin track bigwig doesn't have information in this region")
+            raise BigWigInaccessible(chrom, start, end)
     else:
         target_array = None
     return target_array
@@ -385,4 +390,4 @@ def extract_info(chrom, start, end, label, genome_pyfaidx, bigwigs, target, stra
         for k,t in transforms.items(): 
             feature[k] = t(feature[k])
     
-    return feature
+    return feature

{seqchromloader-0.6.0 → seqchromloader-0.6.2}/seqchromloader.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: seqchromloader
-Version: 0.6.0
+Version: 0.6.2
 Summary: Sequence and chromatin dataloader for deep learning
 Home-page: https://github.com/yztxwd/seqchromloader
 Author-email: yztxwd@gmail.com

{seqchromloader-0.6.0 → seqchromloader-0.6.2}/setup.py

@@ -20,7 +20,7 @@ setup(
     # eg: 1.0.0, 1.0.1, 3.0.2, 5.0-beta, etc.
     # You CANNOT upload two versions of your package with the same version number
     # This field is REQUIRED
-    version="0.6.0",
+    version="0.6.2",
 
     # The packages that constitute your project.
     # For my project, I have only one - "pydash".

{seqchromloader-0.6.0 → seqchromloader-0.6.2}/tests/test_writer_loader.py

@@ -245,6 +245,23 @@ class Test(unittest.TestCase):
         self.assertEqual(target[0].item(), 6.0)
         self.assertEqual(label[1].item(), 1)
 
+    def test_bed_loader_return_region(self):
+
+        it = iter(SeqChromDatasetByBed(
+            bed="data/sample.bed",
+            genome_fasta="data/sample.fa",
+            bigwig_filelist=["data/sample.bw"],
+            target_bam="data/sample.bam",
+            transforms={"seq": test_seq_transform,
+                        "chrom": test_chrom_transform,
+                        "target": test_target_transform},
+            dataloader_kws={"batch_size":2,
+                            "shuffle":False},
+            return_region=True
+        ))
+        region, seq, chrom, target, label = next(it)
+        self.assertEqual(region[0], "chr19:0-5")
+
     def test_lightning_datamodule(self):
         dm = SeqChromDataModule(
             train_wds="data/test_0.tar.gz",
@@ -295,4 +312,4 @@ def test_target_transform(target):
     return target * 3
 
 if __name__ == "__main__":
-    unittest.main(verbosity=2)
+    unittest.main(verbosity=2)