cool-seq-tool 0.4.0.dev0__py3-none-any.whl → 0.4.0.dev2__py3-none-any.whl

cool_seq_tool/data/__init__.py

@@ -1,2 +1,2 @@
 """Module for data"""
-from .data_downloads import DataDownload  # noqa: F401
+from .data_downloads import DataDownload

cool_seq_tool/data/data_downloads.py

@@ -4,7 +4,6 @@ import gzip
 import logging
 import shutil
 from ftplib import FTP
-from os import remove
 from pathlib import Path
 
 from dateutil import parser
@@ -38,18 +37,20 @@ class DataDownload:
             files = ftp.nlst()
             mane_summary_file = [f for f in files if f.endswith(".summary.txt.gz")]
             if not mane_summary_file:
-                raise Exception("Unable to download MANE summary data")
+                msg = "Unable to download MANE summary data"
+                raise Exception(msg)
             mane_summary_file = mane_summary_file[0]
             self._mane_summary_path = self._data_dir / mane_summary_file[:-3]
             mane_data_path = self._data_dir / mane_summary_file
             if not self._mane_summary_path.exists():
                 logger.info("Downloading MANE summary file from NCBI.")
-                with open(mane_data_path, "wb") as fp:
+                with mane_data_path.open("wb") as fp:
                     ftp.retrbinary(f"RETR {mane_summary_file}", fp.write)
-                with gzip.open(mane_data_path, "rb") as f_in:
-                    with open(self._mane_summary_path, "wb") as f_out:
-                        shutil.copyfileobj(f_in, f_out)
-                remove(mane_data_path)
+                with gzip.open(
+                    mane_data_path, "rb"
+                ) as f_in, self._mane_summary_path.open("wb") as f_out:
+                    shutil.copyfileobj(f_in, f_out)
+                mane_data_path.unlink()
                 logger.info("MANE summary file download complete.")
         return self._mane_summary_path
 
@@ -66,18 +67,23 @@ class DataDownload:
             ftp_file_path = f"{ftp_dir_path}{lrg_refseqgene_file}"
             timestamp = ftp.voidcmd(f"MDTM {ftp_file_path}")[4:].strip()
             date = str(parser.parse(timestamp)).split()[0]
-            version = datetime.datetime.strptime(date, "%Y-%m-%d").strftime("%Y%m%d")
+            version = (
+                datetime.datetime.strptime(date, "%Y-%m-%d")
+                .astimezone(tz=datetime.timezone.utc)
+                .strftime("%Y%m%d")
+            )
             fn_versioned = f"{lrg_refseqgene_file}_{version}"
             lrg_refseqgene_path = self._data_dir / lrg_refseqgene_file
             self._lrg_refseqgene_path = self._data_dir / fn_versioned
             if not self._lrg_refseqgene_path.exists():
                 logger.info("Downloading LRG RefSeq data from NCBI.")
                 ftp.cwd(ftp_dir_path)
-                with open(lrg_refseqgene_path, "wb") as fp:
+                with lrg_refseqgene_path.open("wb") as fp:
                     ftp.retrbinary(f"RETR {lrg_refseqgene_file}", fp.write)
-                with open(lrg_refseqgene_path, "rb") as f_in:
-                    with open(self._lrg_refseqgene_path, "wb") as f_out:
-                        shutil.copyfileobj(f_in, f_out)
-                remove(lrg_refseqgene_path)
+                with lrg_refseqgene_path.open(
+                    "rb"
+                ) as f_in, self._lrg_refseqgene_path.open("wb") as f_out:
+                    shutil.copyfileobj(f_in, f_out)
+                lrg_refseqgene_path.unlink()
                 logger.info("LRG RefSeq data download complete.")
         return self._lrg_refseqgene_path