cool-seq-tool 0.14.4__py3-none-any.whl → 0.14.5__py3-none-any.whl

cool_seq_tool/mappers/exon_genomic_coords.py

@@ -2,6 +2,7 @@
 
 import logging
 
+from ga4gh.core.models import Extension
 from ga4gh.vrs.models import SequenceLocation, SequenceReference
 from pydantic import ConfigDict, Field, StrictInt, StrictStr, model_validator
 
@@ -65,9 +66,6 @@ class TxSegment(BaseModelForbidExtra):
     genomic_location: SequenceLocation = Field(
         ..., description="The genomic position of a transcript segment."
     )
-    is_exonic: bool = Field(
-        default=True, description="If the position occurs on an exon"
-    )
 
     @model_validator(mode="before")
     def check_seg_pos(cls, values: dict) -> dict:  # noqa: N805
@@ -99,8 +97,8 @@ class TxSegment(BaseModelForbidExtra):
                         "refgetAccession": "SQ.Ya6Rs7DHhDeg7YaOSg1EoNi3U_nQ9SvO",
                     },
                     "end": 154192135,
+                    "extensions": [{"name": "is_exonic", "value": True}],
                 },
-                "is_exonic": True,
             }
         }
     )
@@ -157,8 +155,8 @@ class GenomicTxSeg(BaseModelForbidExtra):
                             "refgetAccession": "SQ.Ya6Rs7DHhDeg7YaOSg1EoNi3U_nQ9SvO",
                         },
                         "end": 154192135,
+                        "extensions": [{"name": "is_exonic", "value": True}],
                     },
-                    "is_exonic": True,
                 },
                 "errors": [],
             }
@@ -224,8 +222,8 @@ class GenomicTxSegService(BaseModelForbidExtra):
                             "refgetAccession": "SQ.Ya6Rs7DHhDeg7YaOSg1EoNi3U_nQ9SvO",
                         },
                         "end": 154192135,
+                        "extensions": [{"name": "is_exonic", "value": True}],
                     },
-                    "is_exonic": True,
                 },
                 "seg_end": {
                     "exon_ord": 7,
@@ -237,8 +235,8 @@ class GenomicTxSegService(BaseModelForbidExtra):
                             "refgetAccession": "SQ.Ya6Rs7DHhDeg7YaOSg1EoNi3U_nQ9SvO",
                         },
                         "start": 154170399,
+                        "extensions": [{"name": "is_exonic", "value": True}],
                     },
-                    "is_exonic": True,
                 },
             }
         }
@@ -730,7 +728,12 @@ class ExonGenomicCoordsMapper:
         ), None
 
     def _get_vrs_seq_loc(
-        self, genomic_ac: str, genomic_pos: int, is_seg_start: bool, strand: Strand
+        self,
+        genomic_ac: str,
+        genomic_pos: int,
+        is_seg_start: bool,
+        strand: Strand,
+        is_exonic: bool = True,
     ) -> tuple[SequenceLocation | None, str | None]:
         """Create VRS Sequence Location for genomic position where transcript segment
         occurs
@@ -740,6 +743,8 @@ class ExonGenomicCoordsMapper:
         :param is_seg_start: ``True`` if ``genomic_pos`` is where the transcript segment
             starts. ``False`` if ``genomic_pos`` is where the transcript segment ends.
         :param strand: Strand
+        :param is_exonic: A boolean indicating if the genomic breakpoint occurs
+            on an exon. By default, this is set to ``True``.
         :return: Tuple containing VRS location (if successful) and error message (if
             unable to get GA4GH identifier for ``genomic_ac``).
         """
@@ -759,6 +764,7 @@ class ExonGenomicCoordsMapper:
             ),
             start=genomic_pos if use_start else None,
             end=genomic_pos if not use_start else None,
+            extensions=[Extension(name="is_exonic", value=is_exonic)],
         ), None
 
     async def _genomic_to_tx_segment(
@@ -947,7 +953,7 @@ class ExonGenomicCoordsMapper:
         )
 
         genomic_location, err_msg = self._get_vrs_seq_loc(
-            genomic_ac, genomic_pos, is_seg_start, strand
+            genomic_ac, genomic_pos, is_seg_start, strand, is_exonic
         )
         if err_msg:
             return GenomicTxSeg(errors=[err_msg])
@@ -961,7 +967,6 @@ class ExonGenomicCoordsMapper:
                 exon_ord=exon_num,
                 offset=offset,
                 genomic_location=genomic_location,
-                is_exonic=is_exonic,
             ),
         )
 

cool_seq_tool/resources/status.py

@@ -3,6 +3,7 @@
 import logging
 from collections import namedtuple
 from pathlib import Path
+from urllib.parse import urlparse
 
 from agct._core import ChainfileError
 from asyncpg import InvalidCatalogNameError, UndefinedTableError
@@ -11,7 +12,7 @@ from biocommons.seqrepo import SeqRepo
 from cool_seq_tool.handlers.seqrepo_access import SEQREPO_ROOT_DIR, SeqRepoAccess
 from cool_seq_tool.mappers.liftover import LiftOver
 from cool_seq_tool.resources.data_files import DataFile, get_data_file
-from cool_seq_tool.sources.uta_database import UTA_DB_URL, UtaDatabase
+from cool_seq_tool.sources.uta_database import UTA_DB_URL, ParseResult, UtaDatabase
 
 _logger = logging.getLogger(__name__)
 
@@ -119,14 +120,20 @@ async def check_status(
     else:
         status["liftover"] = True
 
+    parsed_result = ParseResult(urlparse(db_url))
+    sanitized_url = parsed_result.sanitized_url
     try:
         await UtaDatabase.create(db_url)
+    except ValueError:
+        _logger.exception("Database URL is not valid")
     except (OSError, InvalidCatalogNameError, UndefinedTableError):
-        _logger.exception("Encountered error instantiating UTA at URI %s", UTA_DB_URL)
+        _logger.exception(
+            "Encountered error instantiating UTA at URI %s", sanitized_url
+        )
     except Exception as e:
         _logger.critical(
             "Encountered unexpected error instantiating UTA from URI %s: %s",
-            UTA_DB_URL,
+            sanitized_url,
             e,
         )
     else:

cool_seq_tool/sources/uta_database.py

@@ -5,7 +5,7 @@ import logging
 from os import environ
 from typing import Any, Literal, TypeVar
 from urllib.parse import ParseResult as UrlLibParseResult
-from urllib.parse import quote, unquote, urlparse
+from urllib.parse import unquote, urlparse, urlunparse
 
 import asyncpg
 import boto3
@@ -101,8 +101,7 @@ class UtaDatabase:
         """
         self.schema = None
         self._connection_pool = None
-        original_pwd = db_url.split("//")[-1].split("@")[0].split(":")[-1]
-        self.db_url = db_url.replace(original_pwd, quote(original_pwd))
+        self.db_url = db_url
         self.args = self._get_conn_args()
 
     def _get_conn_args(self) -> DbConnectionArgs:
@@ -954,3 +953,28 @@ class ParseResult(UrlLibParseResult):
         """Create schema property."""
         path_elems = self.path.split("/")
         return path_elems[2] if len(path_elems) > 2 else None
+
+    @property
+    def sanitized_url(self) -> str:
+        """Sanitized DB URL with the password masked"""
+        netloc = ""
+        if self.username:
+            netloc += self.username
+            if self.password is not None and self.password != "":
+                netloc += ":***"
+            netloc += "@"
+        if self.hostname:
+            netloc += f"{self.hostname}"
+        if self.port:
+            netloc += f":{self.port}"
+
+        return urlunparse(
+            (
+                self.scheme,
+                netloc,
+                self.path,
+                self.params,
+                self.query,
+                self.fragment,
+            )
+        )

{cool_seq_tool-0.14.4.dist-info → cool_seq_tool-0.14.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cool_seq_tool
-Version: 0.14.4
+Version: 0.14.5
 Summary: Common Operation on Lots of Sequences Tool
 Author: Kori Kuzma, James Stevenson, Katie Stahl, Alex Wagner
 License: MIT License

{cool_seq_tool-0.14.4.dist-info → cool_seq_tool-0.14.5.dist-info}/RECORD

@@ -6,20 +6,20 @@ cool_seq_tool/handlers/__init__.py,sha256=KalQ46vX1MO4SJz2SlspKoIRy1n3c3Vp1t4Y2p
 cool_seq_tool/handlers/seqrepo_access.py,sha256=lRzPc8V0eZJTlefbHuVKeZTEC8-KcyPzpqX7vx3amu8,9118
 cool_seq_tool/mappers/__init__.py,sha256=tavpwkNogg_nF1J_kb6Q9jk7ezqdRz063v7BMZ4koLM,390
 cool_seq_tool/mappers/alignment.py,sha256=kWgYssM8YL-Z13H9GdpL77P7simNcbxltAs9YDXHE54,9640
-cool_seq_tool/mappers/exon_genomic_coords.py,sha256=fV4LyrpHPLRrx6AtV15g93q5XCH3i-y3Wj9tl-Cg8mM,45845
+cool_seq_tool/mappers/exon_genomic_coords.py,sha256=N6Wi7D8hs6gZi-BC4ICuWQEGeqUbBysqxG18EtIIgSk,46187
 cool_seq_tool/mappers/feature_overlap.py,sha256=X5UFClaH6ixRsO2fDLxqjywp-Z0bvNx4uzgBICy394U,9758
 cool_seq_tool/mappers/liftover.py,sha256=lltx9zxfkrb5PHtJlKp3a39JCwPP4e0Zft-mQc1jXL8,3367
 cool_seq_tool/mappers/mane_transcript.py,sha256=IluiLBxPQoY-CxkpqpjEBcMlHvrNLa34wdKdQxtKgDY,54613
 cool_seq_tool/resources/__init__.py,sha256=VwUC8YaucTS6SmRirToulZTF6CuvuLQRSxFfSfAovCc,77
 cool_seq_tool/resources/data_files.py,sha256=6d1M5WjeFHdTQpzxqjQ78auQRZvIBVqH8QNCrmRRDXw,4205
-cool_seq_tool/resources/status.py,sha256=5UKx5FIQuyIY7FU4kSinDIM4MhLpr9_MiQDDBNt9kRo,5990
+cool_seq_tool/resources/status.py,sha256=iP-4NiSmqV-D--gypZyrSqVbOWQvyBZICKQb-VinTik,6241
 cool_seq_tool/resources/transcript_mapping.tsv,sha256=AO3luYQAbFiCoRgiiPXotakb5pAwx1jDCeXpvGdIuac,24138769
 cool_seq_tool/sources/__init__.py,sha256=51QiymeptF7AeVGgV-tW_9f4pIUr0xtYbyzpvHOCneM,304
 cool_seq_tool/sources/mane_transcript_mappings.py,sha256=C5puIA1xuEzBaSvs8VtSxVb2OIDGUg5no8v6Ma2QSdw,6597
 cool_seq_tool/sources/transcript_mappings.py,sha256=903RKTMBO2rbKh6iTQ1BEWnY4C7saBFMPw2_4ATuudg,10054
-cool_seq_tool/sources/uta_database.py,sha256=zzRzmYuybqzEg7zeuQjhK46SPK5GfbiWWNRGNJju8AI,36197
-cool_seq_tool-0.14.4.dist-info/licenses/LICENSE,sha256=IpqC9A-tZW7XXXvCS8c4AVINqkmpxiVA-34Qe3CZSjo,1072
-cool_seq_tool-0.14.4.dist-info/METADATA,sha256=gPz48irXCwNUecTcgpKrdrHiBhi8R_Is55S2UYs9Qtk,6535
-cool_seq_tool-0.14.4.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-cool_seq_tool-0.14.4.dist-info/top_level.txt,sha256=cGuxdN6p3y16jQf6hCwWhE4OptwUeZPm_PNJlPb3b0k,14
-cool_seq_tool-0.14.4.dist-info/RECORD,,
+cool_seq_tool/sources/uta_database.py,sha256=38CQ0QHHh0kA87tdgsJHJiHdJHQc06ylBYfemGFUlZc,36759
+cool_seq_tool-0.14.5.dist-info/licenses/LICENSE,sha256=IpqC9A-tZW7XXXvCS8c4AVINqkmpxiVA-34Qe3CZSjo,1072
+cool_seq_tool-0.14.5.dist-info/METADATA,sha256=tgWwlBS_2Z71zUJpmO7MjDE3cNFy6GYH7W8YyJrTo0I,6535
+cool_seq_tool-0.14.5.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+cool_seq_tool-0.14.5.dist-info/top_level.txt,sha256=cGuxdN6p3y16jQf6hCwWhE4OptwUeZPm_PNJlPb3b0k,14
+cool_seq_tool-0.14.5.dist-info/RECORD,,