geney 1.4.7__py2.py3-none-any.whl → 1.4.8__py2.py3-none-any.whl

geney/Gene.py

@@ -1,6 +1,6 @@
 import copy
 # import random
-from . import config
+from . import H
 from typing import Any, Dict, List, Tuple, Optional, Iterator, Union, TYPE_CHECKING
 from collections import Counter
 from .utils.utils import unload_pickle

geney/utils/SeqMats.py

@@ -79,6 +79,7 @@ class SeqMat:
         self.seq_array["cons"] = np.nan if conservation is None else conservation
         self.seq_array["valid_mask"] = self.seq_array["nt"] != b"-"
         self.insertion_counters = defaultdict(int)
+        self._pos_to_idx = {pos: i for i, pos in enumerate(self.seq_array["index"])}
 
         self.source = source if source is not None else "Unknown"
         self.notes = notes if notes is not None else {}
@@ -177,27 +178,35 @@ class SeqMat:
             self.reverse_complement()
 
         # Normalize shared prefix (similar to left-alignment in VCFs)
-        if ref and alt and ref[0] == alt[0]:
+        while ref and alt and ref[0] == alt[0]:
             pos += 1
             ref = ref[1:] or "-"
             alt = alt[1:] or "-"
 
         # Case 1: SNP or multi-base substitution
         if ref != "-" and alt != "-":
-            if len(ref) == len(alt):
-                # print('Inserting SNP')
-                pos_idx = np.searchsorted(self.seq_array["index"], pos)
-                end_idx = pos_idx + len(ref)
-                if end_idx > len(self.seq_array):
-                    raise ValueError(f"Substitution range exceeds sequence length at position {pos}.")
-                segment = self.seq_array["ref"][pos_idx:end_idx].tobytes().decode()
-                if segment != ref:
-                    raise ValueError(f"Reference mismatch at position {pos}: expected '{ref}', found '{segment}'")
-                for i, nt in enumerate(alt):
-                    self.seq_array["nt"][pos_idx + i] = nt.encode()
-            else:
+            if len(ref) != len(alt):
                 raise ValueError("Substitution mutations must have alleles of equal length.")
 
+            # print('Inserting SNP')
+            # pos_idx = np.searchsorted(self.seq_array["index"], pos)
+            pos_idx = self._pos_to_idx.get(pos)
+            if pos_idx is None:
+                raise ValueError(f"Position {pos} not found in index")
+            
+            end_idx = pos_idx + len(ref)
+            if end_idx > len(self.seq_array):
+                raise ValueError(f"Substitution range exceeds sequence length at position {pos}.")
+            # segment = self.seq_array["ref"][pos_idx:end_idx].tobytes().decode()
+            # if segment != ref:
+            #     raise ValueError(f"Reference mismatch at position {pos}: expected '{ref}', found '{segment}'")
+            ref_segment = self.seq_array["ref"][pos_idx:end_idx]
+            if not np.all(ref_segment == np.frombuffer(ref.encode(), dtype='S1')):
+                raise ValueError(f"Reference mismatch at position {pos}")
+            self.seq_array["nt"][pos_idx:end_idx] = np.frombuffer(alt.encode(), dtype='S1')
+            # for i, nt in enumerate(alt):
+            #     self.seq_array["nt"][pos_idx + i] = nt.encode()
+
         # Case 2: Insertion (ref is '-' means nothing was present, and we need to add bases)
         elif ref == "-" and alt != "-":
             if only_snps:

{geney-1.4.7.dist-info → geney-1.4.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: geney
-Version: 1.4.7
+Version: 1.4.8
 Summary: A Python package for gene expression modeling.
 Home-page: https://github.com/nicolaslynn/geney
 Author: Nicolas Lynn

{geney-1.4.7.dist-info → geney-1.4.8.dist-info}/RECORD

@@ -1,5 +1,5 @@
 geney/Fasta_segment.py,sha256=99HxNGNh_MfdVW6hhtlb1vOn7eSmT7oFoEfHDFMxG8w,11275
-geney/Gene.py,sha256=6x1sEZV50Il4oydegW6iHIF12EZTGexniG3YUD-3DfM,7036
+geney/Gene.py,sha256=QAST97ySepdZ3SnBzxIlVeJ3kQW8UEGkpNmxL_BwES0,7031
 geney/Oncosplice.py,sha256=ETAvMl_Oq6mEJQHPNwdDO5csX6Ahuped_om10KifCyM,17739
 geney/SeqMats.py,sha256=9-eJnfU2w3LGc0XvVvFEO_QrBneTkC6xkZKDfTcEw5o,19282
 geney/SpliceSimulator.py,sha256=iF6feVeSnsKFmn3WV60CgWLI0_rSLgpq5fVFL1IOv_4,18491
@@ -37,7 +37,7 @@ geney/translation_initiation/tis_utils.py,sha256=AF3siFjuQH-Rs44EV-80zHdbxRMvN4w
 geney/translation_initiation/resources/kozak_pssm.json,sha256=pcd0Olziutq-6H3mFWDCD9cujQ_AlZO-iiOvBl82hqE,1165
 geney/translation_initiation/resources/tis_regressor_model.joblib,sha256=IXb4DUDhJ5rBDKcqMk9zE3ECTZZcdj7Jixz3KpoZ7OA,2592025
 geney/utils/Fasta_segment.py,sha256=weB5NJ65P0XiyAJCiCHx4T9sHC1pWLpuQeOy0B85gyg,11364
-geney/utils/SeqMats.py,sha256=vjU0lTkB0s0RoLjNXLqt0kJQDni-it09-iAOv5QAYFs,17686
+geney/utils/SeqMats.py,sha256=kpv5v_6VgUc72Qip3fLYZV_7Jb6dIoQLIrECL5BCBl4,18207
 geney/utils/TranscriptLibrary.py,sha256=ma_ZVPgglxXDDneEvdqxxeqxG8eSFL-zgLUXyC6BqY8,2070
 geney/utils/__init__.py,sha256=-nJ-DMx1JzP-ZCe_QuQCeM0ZYIT_16jxoXDhUaO_4Oc,714
 geney/utils/mutation_utils.py,sha256=r-pHr56gEa5kh_DPX8MjFY3ZfYaOtyo4CUfJ5ZHlXPw,3243
@@ -45,7 +45,7 @@ geney/utils/pangolin_utils.py,sha256=JQSPbWxdzqGFYfWQktkfLMaMSGR28eGQhNzO7MLMe5M
 geney/utils/spliceai_utils.py,sha256=VtrIbjyQxk_3lw86eWjftRYyal9OzxArJ0GV5u_ymTg,2721
 geney/utils/splicing_utils.py,sha256=vPCGnCPR1ooEZEHR79yFHLmRQXEJHXEQjjxpBR-YWOs,20635
 geney/utils/utils.py,sha256=m51Vd0cEbrcIHo6_8BAuI9YSPcKRs22e5LfVd2Qj6Is,2181
-geney-1.4.7.dist-info/METADATA,sha256=w5BvzaWbcngGPb-CJ-5pybBIEouDvh9fLJtTgSeO1fg,989
-geney-1.4.7.dist-info/WHEEL,sha256=AHX6tWk3qWuce7vKLrj7lnulVHEdWoltgauo8bgCXgU,109
-geney-1.4.7.dist-info/top_level.txt,sha256=O-FuNUMb5fn9dhZ-dYCgF0aZtfi1EslMstnzhc5IIVo,6
-geney-1.4.7.dist-info/RECORD,,
+geney-1.4.8.dist-info/METADATA,sha256=VBMjJtnqbqt_cumsy9JEjo-ZHk-_KneCBR5-315suQ8,989
+geney-1.4.8.dist-info/WHEEL,sha256=AHX6tWk3qWuce7vKLrj7lnulVHEdWoltgauo8bgCXgU,109
+geney-1.4.8.dist-info/top_level.txt,sha256=O-FuNUMb5fn9dhZ-dYCgF0aZtfi1EslMstnzhc5IIVo,6
+geney-1.4.8.dist-info/RECORD,,