Merge pull request #33 from Clinical-Genomics/level_up

Move inner schug dir one level up

CHANGELOG.md

@@ -1,6 +1,9 @@
 [unreleased]
+### Changed
+- Move the `schug` directory up in root dir folder
 ### Fixed
 - Use a memory database as default database in demo instance
+- Issues flagged by SonarCloud
 
 [1.0.0]
 ### Added

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "schug"
-version = "0.1.0"
+version = "1.0.0"
 description = "Keep track of genes, transcripts and exons from different sources"
 authors = ["moonso <mans.magnusson@scilifelab.se>"]
 readme = "README.md"

schug/demo/__init__.py

@@ -0,0 +1,15 @@
+from pathlib import Path
+
+import importlib_resources
+
+# Paths
+EXONS_37_FILE_PATH: Path = Path(importlib_resources.files("schug"), "demo", "exons_37.tsv")
+EXONS_38_FILE_PATH: Path = Path(importlib_resources.files("schug"), "demo", "exons_38.tsv")
+GENES_37_FILE_PATH: Path = Path(importlib_resources.files("schug"), "demo", "genes_37.tsv")
+GENES_38_FILE_PATH: Path = Path(importlib_resources.files("schug"), "demo", "genes_38.tsv")
+TRANSCRIPTS_37_FILE_PATH: Path = Path(
+    importlib_resources.files("schug"), "demo", "transcripts_37.tsv"
+)
+TRANSCRIPTS_38_FILE_PATH: Path = Path(
+    importlib_resources.files("schug"), "demo", "transcripts_38.tsv"
+)

src/schug/load/fetch_resource.py → schug/load/fetch_resource.py

@@ -1,7 +1,7 @@
 import logging
 import urllib
 import zlib
-from typing import List
+from typing import AsyncGenerator, List
 from urllib.request import urlopen
 from urllib.response import addinfourl
 
@@ -61,7 +61,7 @@ def fetch_resource(url: str) -> List[str]:
     return data
 
 
-async def stream_resource(url: str) -> bytes:
+async def stream_resource(url: str) -> AsyncGenerator:
     """Stream a file from an external service"""
     async with httpx.AsyncClient(timeout=None) as client:
         async with client.stream("GET", url) as r:

src/schug/load/gene_resources.py → schug/load/gene_resources.py

@@ -27,11 +27,11 @@ def fetch_genes_to_hpo_to_disease() -> List[str]:
 
 def fetch_hgnc() -> List[str]:
     """Fetch the hgnc genes file from
-        ftp://ftp.ebi.ac.uk/pub/databases/genenames/new/tsv/hgnc_complete_set.txt
+        sftp://sftp.ebi.ac.uk/pub/databases/genenames/new/tsv/hgnc_complete_set.txt
     Returns:
         hgnc_gene_lines(list(str))
     """
-    url = "ftp://ftp.ebi.ac.uk/pub/databases/genenames/new/tsv/hgnc_complete_set.txt"
+    url = "sftp://sftp.ebi.ac.uk/pub/databases/genenames/new/tsv/hgnc_complete_set.txt"
     LOG.info("Fetching HGNC genes from %s", url)
 
     return fetch_resource(url)
@@ -41,7 +41,7 @@ def fetch_exac_constraint() -> List[str]:
     """Fetch the file with exac constraint scores"""
     file_name = "fordist_cleaned_exac_r03_march16_z_pli_rec_null_data.txt"
     url = (
-        "ftp://ftp.broadinstitute.org/pub/ExAC_release/release0.3/functional_gene_constraint" "/{0}"
+        "sftp://sftp.broadinstitute.org/pub/ExAC_release/release0.3/functional_gene_constraint/{0}"
     ).format(file_name)
 
     LOG.info("Fetching ExAC genes")

src/schug/main.py → schug/main.py

@@ -4,6 +4,7 @@
 from .endpoints import exons, genes, transcripts
 
 app = FastAPI()
+NOT_FOUND_MESSAGE = "Not found"
 
 ### REST API
 
@@ -17,20 +18,20 @@ async def root():
     genes.router,
     prefix="/genes",
     tags=["genes"],
-    responses={status.HTTP_404_NOT_FOUND: {"description": "Not found"}},
+    responses={status.HTTP_404_NOT_FOUND: {"description": NOT_FOUND_MESSAGE}},
 )
 
 
 app.include_router(
     transcripts.router,
     prefix="/transcripts",
     tags=["transcripts"],
-    responses={status.HTTP_404_NOT_FOUND: {"description": "Not found"}},
+    responses={status.HTTP_404_NOT_FOUND: {"description": NOT_FOUND_MESSAGE}},
 )
 
 app.include_router(
     exons.router,
     prefix="/exons",
     tags=["exons"],
-    responses={status.HTTP_404_NOT_FOUND: {"description": "Not found"}},
+    responses={status.HTTP_404_NOT_FOUND: {"description": NOT_FOUND_MESSAGE}},
 )

src/schug/models/gene.py → schug/models/gene.py

@@ -1,4 +1,4 @@
-from typing import TYPE_CHECKING, List, Optional, Literal
+from typing import TYPE_CHECKING, List, Literal, Optional
 
 from pydantic import BaseModel
 from pydantic import Field as PydanticField
@@ -50,10 +50,10 @@ def convert_to_none(cls, v):
             return None
         return v
 
-    @validator('hgnc_id', pre=True)
+    @validator("hgnc_id", pre=True)
     def modify_id(cls, v):
         if type(v) != int:
-            return v.replace('HGNC:', '')
+            return v.replace("HGNC:", "")
         return v
 
 
@@ -63,7 +63,7 @@ class HgncGene(BaseModel):
     entrez_id: int
     hgnc_symbol: str = PydanticField(None, alias="HGNC symbol")
     hgnc_id: str
-    ccds_id: str = None
+    ccds_id: Optional[str] = None
     short_description: str = PydanticField(None, alias="alias_name")
     alias_symbol: str
 

src/schug/models/transcript.py → schug/models/transcript.py

@@ -51,7 +51,7 @@ class EnsemblTranscript(BaseModel):
     refseq_mrna: str = PydanticField(None, alias="RefSeq mRNA ID")
     refseq_mrna_predicted: str = PydanticField(None, alias="RefSeq mRNA predicted ID")
     refseq_ncrna_predicted: str = PydanticField(None, alias="RefSeq ncRNA ID")
-    refseq_id: str = None
+    refseq_id: Optional[str] = None
 
     @validator("refseq_id", always=True)
     def set_refseq_id(cls, _, values: dict) -> Optional[str]: