geney 1.2.2__py2.py3-none-any.whl → 1.2.4__py2.py3-none-any.whl

geney/data_setup.py

@@ -7,7 +7,7 @@ from tqdm import tqdm
 import requests
 import argparse
 from sh import gunzip
-
+import shutil
 
 def download(external_url, local_path):
     print(f"Grabbing {external_url}")
@@ -94,6 +94,7 @@ def process_transcript(transcript_df, rev, chrm, cons_data):
         cds_start, cds_end = cds_start[0], cds_end[0]
         data.update({'TIS': cds_start, 'TTS': cds_end, 'protein_id': transcript.protein_id})
 
+    print(f"{transcript.transcript_id} in cons_data: {transcript.transcript_id in cons_data}")
     if transcript.transcript_id in cons_data:
         data.update({'cons_available': True, 'cons_vector': cons_data[transcript.transcript_id]['scores'], 'cons_seq': cons_data[transcript.transcript_id]['seq']})
 
@@ -104,6 +105,8 @@ def process_transcript(transcript_df, rev, chrm, cons_data):
 
 
 def retrieve_and_parse_ensembl_annotations(local_path, annotations_file, cons_data, gtex_file='', valid_biotypes=('protein_coding')):
+    print(cons_data.keys())
+
     if gtex_file:
         gtex_df = pd.read_csv(gtex_file, delimiter='\t', header=2)
         gtex_df.Name = gtex_df.apply(lambda row: row.Name.split('.')[0], axis=1)
@@ -225,21 +228,24 @@ def main():
         'TEMP': os.path.join(args.basepath, args.organism, 'temp')
     }
 
+    base_path = Path(args.basepath) / args.organism
+
     if config_file.exists():
         config_data = unload_json(config_file)
         overwrite = 'y'
         if args.organism in config_data:
             overwrite = input("Organism {args.organism} already configured... Overwrite? (y/n)")
+
         if overwrite == 'y':
             config_data[args.organism] = config_paths
             dump_json(config_file, config_data)
+            shutil.rmtree(base_path)
         else:
             raise SystemExit("Exiting configuration.")
     else:
         config_data = {args.organism: config_paths}
         dump_json(config_file, config_data)
 
-    base_path = Path(args.basepath) / args.organism
     if base_path.exists() and len(os.listdir(base_path)) > 0:
         raise FileExistsError(f"Directory {base_path} not empty.")
 

geney/oncosplice.py

@@ -271,7 +271,7 @@ class Gene:
 
         if tid not in self.transcripts:
             raise AttributeError(f"Transcript '{tid}' not found in gene '{self.gene_name}'.")
-        return Transcript(self.transcripts[tid])
+        return Transcript(self.transcripts[tid], organism=self.organism)
 
     def run_transcripts(self, primary_transcript=False, protein_coding=False):
         for tid, annotations in self.transcripts.items():

{geney-1.2.2.dist-info → geney-1.2.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: geney
-Version: 1.2.2
+Version: 1.2.4
 Summary: A Python package for gene expression modeling.
 Home-page: https://github.com/nicolaslynn/geney
 Author: Nicolas Lynn
@@ -17,7 +17,7 @@ Requires-Dist: numpy ==1.26.4
 Requires-Dist: pandas ==2.2.1
 Requires-Dist: networkx ==3.2.1
 Requires-Dist: viennarna ==2.6.4
-Requires-Dist: tqdm ==4.66.1
+Requires-Dist: tqdm >=4.66.1
 Requires-Dist: spliceai ==1.3.1
 Requires-Dist: scikit-learn ==1.0.2
 Requires-Dist: biopython ==1.81

{geney-1.2.2.dist-info → geney-1.2.4.dist-info}/RECORD

@@ -4,12 +4,12 @@ geney/__init__.py,sha256=r-Yvpo_Tc236DcsqsFyexT21iVoYCVl9zoJj5pFuWEE,407
 geney/benchmark_clinvar.py,sha256=LLl77e95Qbg9Kd-m2yL8ilmzubSz9SKogeARwssT4Ks,5532
 geney/compare_sets.py,sha256=TcgL57V7BUPxBoW9lv3xr8qK2Acmykn85Ev3avicQr8,2977
 geney/config_setup.py,sha256=VA6mhVGMRadwlpEx4m1wrssmDM8qpfKT21MAijIwjyQ,428
-geney/data_setup.py,sha256=mV_sSCMT8C41q_PD_G34MIBvsBKA0Czrpw17-DcYmT4,12052
+geney/data_setup.py,sha256=LTiJMYPgv9KnIgUNw-D57Fu4nxL4OojXMpmdhE8QSYU,12228
 geney/gtex.py,sha256=asL2lHyU5KsbWpV096vkf1Ka7hSo_RRfZqw7p5nERmE,1919
 geney/gtex_utils.py,sha256=asL2lHyU5KsbWpV096vkf1Ka7hSo_RRfZqw7p5nERmE,1919
 geney/immune_utils.py,sha256=ZRni5ttrhpYBnmNr0d0ZatIbNPYs4nmQuoUO00SpsS4,5271
 geney/netchop.py,sha256=AMiy9YsdTmX4B3k3Y5Yh7EmoGAojM1O3AzhPKOiB--g,3050
-geney/oncosplice.py,sha256=_uQ13nuOFNNVsl_7EzCmJtC9WLbzX9zEcqoA7PvkQWY,71786
+geney/oncosplice.py,sha256=xDzCLivFyurx-qlQo9cyrV-9KJ9VykYAb8lY9DDWl7Q,71810
 geney/oncosplice_mouse.py,sha256=LYLOukI9qI1IBkyl1qVRFR5d1NAw7Orlj8Zth-4xCW8,12962
 geney/oncosplice_pipeline.py,sha256=hpGqFHOdn8i8tvvs1-t3-G9Ko18zInwoDXBJbbrfbC4,68036
 geney/performance_utils.py,sha256=FQt7rA4r-Wuq3kceCxsSuMfj3wU1tMG8QnbL59aBohs,4700
@@ -45,7 +45,7 @@ geney/translation_initiation/resources/kozak_pssm.json,sha256=pcd0Olziutq-6H3mFW
 geney/translation_initiation/resources/tis_regressor_model.joblib,sha256=IXb4DUDhJ5rBDKcqMk9zE3ECTZZcdj7Jixz3KpoZ7OA,2592025
 geney/translation_termination/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 geney/translation_termination/tts_utils.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-geney-1.2.2.dist-info/METADATA,sha256=M-hFK8XG5ZbWtH0gsZvGgDBH18l6iGuDKT8xOCGYDSs,1198
-geney-1.2.2.dist-info/WHEEL,sha256=iYlv5fX357PQyRT2o6tw1bN-YcKFFHKqB_LwHO5wP-g,110
-geney-1.2.2.dist-info/top_level.txt,sha256=O-FuNUMb5fn9dhZ-dYCgF0aZtfi1EslMstnzhc5IIVo,6
-geney-1.2.2.dist-info/RECORD,,
+geney-1.2.4.dist-info/METADATA,sha256=0oE3AHzIGNnpkmPiPK0wXpDRuxZbA6ZcdvpOloz82bQ,1198
+geney-1.2.4.dist-info/WHEEL,sha256=iYlv5fX357PQyRT2o6tw1bN-YcKFFHKqB_LwHO5wP-g,110
+geney-1.2.4.dist-info/top_level.txt,sha256=O-FuNUMb5fn9dhZ-dYCgF0aZtfi1EslMstnzhc5IIVo,6
+geney-1.2.4.dist-info/RECORD,,