WIP

Signed-off-by: Mihai Todor <todormihai@gmail.com>

Author: mihaitodor

.vscode/launch.json

@@ -23,6 +23,7 @@
             "type": "debugpy",
             "request": "launch",
             "program": "${file}",
+            "cwd": "${fileDirname}",
             "console": "integratedTerminal",
             "justMyCode": false
         }

app/__init__.py

@@ -1,9 +1,10 @@
+import os
+
 import connexion
 import flask
+import hgvs
 from flask_cors import CORS
-import os
 
-import hgvs
 # Disable the hgvs LRU cache to avoid blowing up memory
 # TODO: Revisit this, since this caching might not use a ton of memory.
 hgvs.global_config.lru_cache.maxsize = 0

app/endpoints.py

@@ -1,7 +1,8 @@
 from collections import OrderedDict
+
 from flask import abort, jsonify
-from app import common
-from app import utilities_endpoints
+
+from app import common, utilities_endpoints
 
 
 def find_subject_variants(
@@ -180,11 +181,7 @@ def find_subject_specific_variants(
     subject = subject.strip()
     common.validate_subject(subject)
 
-    try:
-        variants = list(map(common.get_variant, variants))
-    except Exception as err:
-        print(f"Unexpected {err=}, {type(err)=}")
-        abort(422, 'Failed LiftOver')
+    variants = list(map(common.get_variant, variants))
 
     # Query
     query = {}
@@ -836,22 +833,13 @@ def find_subject_tx_implications(
         ranges = list(map(common.get_range, ranges))
         common.get_lift_over_range(ranges)
 
-        try:
-            variants = common.get_variants(ranges, query)
-        except Exception as err:
-            print(f"Unexpected {err=}, {type(err)=}")
-            abort(422, 'Failed LiftOver')
-
+        variants = common.get_variants(ranges, query)
         if not variants:
             return jsonify({"resourceType": "Parameters"})
         normalized_variants = [{variant["BUILD"]: variant["SPDI"]} for variant in variants]
 
     if variants and not ranges:
-        try:
-            normalized_variants = list(map(common.get_variant, variants))
-        except Exception as err:
-            print(f"Unexpected {err=}, {type(err)=}")
-            abort(422, 'Failed LiftOver')
+        normalized_variants = list(map(common.get_variant, variants))
 
     # Result Object
     result = OrderedDict()
@@ -1110,22 +1098,13 @@ def find_subject_dx_implications(
         ranges = list(map(common.get_range, ranges))
         common.get_lift_over_range(ranges)
 
-        try:
-            variants = common.get_variants(ranges, query)
-        except Exception as err:
-            print(f"Unexpected {err=}, {type(err)=}")
-            abort(422, 'Failed LiftOver')
-
+        variants = common.get_variants(ranges, query)
         if not variants:
             return jsonify({"resourceType": "Parameters"})
         normalized_variants = [{variant["BUILD"]: variant["SPDI"]} for variant in variants]
 
     if variants and not ranges:
-        try:
-            normalized_variants = list(map(common.get_variant, variants))
-        except Exception as err:
-            print(f"Unexpected {err=}, {type(err)=}")
-            abort(422, 'Failed LiftOver')
+        normalized_variants = list(map(common.get_variant, variants))
 
     # Result Object
     result = OrderedDict()
@@ -1488,11 +1467,7 @@ def find_population_specific_variants(
     # Parameters
     variants = list(map(lambda x: x.strip().split(","), variants))
     for i in range(len(variants)):
-        try:
-            variants[i] = list(map(common.get_variant, variants[i]))
-        except Exception as err:
-            print(f"Unexpected {err=}, {type(err)=}")
-            abort(422, 'Failed LiftOver')
+        variants[i] = list(map(common.get_variant, variants[i]))
 
     # Query
     query = {}
@@ -2001,11 +1976,7 @@ def find_population_tx_implications(
         return jsonify({"resourceType": "Parameters"})
 
     if variants:
-        try:
-            variants = list(map(common.get_variant, variants))
-        except Exception as err:
-            print(f"Unexpected {err=}, {type(err)=}")
-            abort(422, 'Failed LiftOver')
+        variants = list(map(common.get_variant, variants))
 
     condition_code_list = []
     if conditions:
@@ -2217,11 +2188,7 @@ def find_population_dx_implications(
         return jsonify({"resourceType": "Parameters"})
 
     if variants:
-        try:
-            variants = list(map(common.get_variant, variants))
-        except Exception as err:
-            print(f"Unexpected {err=}, {type(err)=}")
-            abort(422, 'Failed LiftOver')
+        variants = list(map(common.get_variant, variants))
 
     condition_code_list = []
     if conditions:

app/input_normalization.py

@@ -6,12 +6,12 @@
 
 from utilities.spdi_refseq import normalize as spdi_normalize
 
-from . import common
+from .common import get_build_and_chrom_by_ref_seq, lift_over
 
 # Set the HGVS_SEQREPO_URL env var so the hgvs library will use the local `utilities/seqfetcher` endpoint instead of
 # making NCBI API calls.
-port = os.environ('PORT', 5000)  # The localhost debugger starts the app on port 5000
-os.environ('HGVS_SEQREPO_URL') = f"http://127.0.0.1:{port}/utilities/seqfetcher"
+port = os.environ.get('PORT', 5000)  # The localhost debugger starts the app on port 5000
+os.environ['HGVS_SEQREPO_URL'] = f"http://127.0.0.1:{port}/utilities/seqfetcher"
 
 database_schema = os.environ['UTA_DATABASE_SCHEMA']
 # Use the biocommons UTA database if we don't specify a custom one.
@@ -113,8 +113,8 @@ def process_NC_HGVS(NC_HGVS):
         print(f"b37normalized: {b37SPDI}; b38normalized: {b38SPDI}")
 
     except Exception:
-        provided_genomic_build = common.get_build_and_chrom_by_ref_seq(parsed_variant.ac)
-        liftover = common.lift_over(parsed_variant.ac, str(parsed_variant.posedit.pos.start), str(parsed_variant.posedit.pos.end))
+        provided_genomic_build = get_build_and_chrom_by_ref_seq(parsed_variant.ac)
+        liftover = lift_over(parsed_variant.ac, str(parsed_variant.posedit.pos.start), str(parsed_variant.posedit.pos.end))
         iv = hgvs.location.Interval(start=liftover['start'], end=liftover['end'])
         posedit = hgvs.posedit.PosEdit(pos=iv, edit=parsed_variant.posedit.edit)
         lifted_variant = hgvs.sequencevariant.SequenceVariant(ac=liftover['refSeq'], type="g", posedit=posedit)

requirements.txt

@@ -9,25 +9,17 @@ pyfastx==2.0.1
 pymongo==4.2.0
 pyranges==0.0.120
 python_dateutil==2.8.2
-requests==2.28.1
+requests==2.32.5
 streamlit==1.19.0
-<<<<<<< HEAD
 streamlit-aggrid==1.0.5
 pytest==7.1.1
-=======
-pytest==7.1.1
-deepdiff==6.7.1
->>>>>>> 1b4630a (Add SPDI normalization with refseq files)
 python-dotenv==1.0.1
 toml==0.10.2
 pandas==2.2.2
 numpy==2.0.2
 pyliftover==0.4
 hgvs==1.5.4
 py-ard==1.5.3
-<<<<<<< HEAD
 openpyxl==3.1.5
 pyvcf3==1.0.3
 tqdm==4.65.1
-=======
->>>>>>> 1b4630a (Add SPDI normalization with refseq files)

utilities/common.py

@@ -11,7 +11,7 @@
 PROJECT_ROOT = BASE_DIR.parent
 load_dotenv(PROJECT_ROOT / "secrets.env")
 
-utilities_data_client_uri = f"mongodb+srv://readonly:{os.environ('MONGODB_READONLY_PASSWORD')}@cluster0.8ianr.mongodb.net/UtilitiesData"
+utilities_data_client_uri = f"mongodb+srv://readonly:{os.environ['MONGODB_READONLY_PASSWORD']}@cluster0.8ianr.mongodb.net/UtilitiesData"
 utilities_client = pymongo.MongoClient(utilities_data_client_uri)
 utilities_db = utilities_client.UtilitiesData
 transcript_data = utilities_db.Transcripts

utilities/vcf2json.py

@@ -7,9 +7,8 @@
 import common
 import pyfastx
 import vcf
-
-from .gene_ref_seq import get_ref_seq_by_chrom
-from .spdi import normalize
+from gene_ref_seq import get_ref_seq_by_chrom
+from spdi import normalize
 
 # Fasta file handles cache
 fasta_cache = {}
@@ -34,7 +33,7 @@ def get_fasta(file):
 def normalize_spdi(ref_seq, pos, ref, alt, build):
     fasta = get_fasta(BUILD37_FILE) if build == 'GRCh37' else get_fasta(BUILD38_FILE)
 
-    return normalize(fasta[ref_seq], pos, ref, alt)
+    return normalize(fasta[ref_seq], ref_seq, pos, ref, alt)
 
 
 def add_phase_records(record, phased_rec_map, sample_position):