Merge pull request #379 from VariantEffect/release-2025.0.0

Release 2025.0.0

Author: bencap

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "mavedb"
-version = "2024.4.3"
+version = "2025.0.0"
 description = "API for MaveDB, the database of Multiplexed Assays of Variant Effect."
 license = "AGPL-3.0-only"
 readme = "README.md"

src/mavedb/__init__.py

@@ -6,6 +6,6 @@
 logger = module_logging.getLogger(__name__)
 
 __project__ = "mavedb-api"
-__version__ = "2024.4.3"
+__version__ = "2025.0.0"
 
 logger.info(f"MaveDB {__version__}")

src/mavedb/lib/experiments.py

@@ -1,7 +1,7 @@
 import logging
 from typing import Optional
 
-from sqlalchemy import func, or_
+from sqlalchemy import func, or_, not_
 from sqlalchemy.orm import Session
 
 from mavedb.lib.logging.context import logging_context, save_to_logging_context
@@ -99,6 +99,19 @@ def search_experiments(
             )
         )
 
+    if search.meta_analysis is not None:
+        if not search.meta_analysis:
+            query = query.filter(
+                or_(
+                    # Keep experiments without any score sets
+                    not_(Experiment.score_sets.any()),
+                    # Keep experiments where score sets exist but have no meta_analyzes_score_sets
+                    Experiment.score_sets.any(not_(ScoreSet.meta_analyzes_score_sets.any()))
+                )
+            )
+        else:
+            query = query.filter(Experiment.score_sets.any(ScoreSet.meta_analyzes_score_sets.any()))
+
     items: list[Experiment] = query.order_by(Experiment.urn, Experiment.title).all()
     if not items:
         items = []

src/mavedb/routers/experiments.py

@@ -7,6 +7,7 @@
 from fastapi import APIRouter, Depends, HTTPException
 from fastapi.encoders import jsonable_encoder
 from sqlalchemy.orm import Session
+from sqlalchemy import or_
 
 from mavedb import deps
 from mavedb.lib.authentication import UserData, get_current_user
@@ -43,7 +44,7 @@
 )
 
 
-# TODO: Rewrite this function.
+# None of any part calls this function. Feel free to modify it if we need it in the future.
 @router.get(
     "/experiments/",
     status_code=200,
@@ -53,30 +54,26 @@
 def list_experiments(
     *,
     editable: Optional[bool] = None,
-    q: Optional[str] = None,
     db: Session = Depends(deps.get_db),
     user_data: Optional[UserData] = Depends(get_current_user),
 ) -> list[Experiment]:
     """
     List experiments.
     """
     query = db.query(Experiment)
-    if q is not None:
-        save_to_logging_context({"query_string": q})
 
+    if editable:
         if user_data is None or user_data.user is None:
             logger.debug(msg="User is anonymous; Cannot list their experiments.", extra=logging_context())
             return []
 
-        if len(q) > 0:
-            logger.debug(msg="Listing experiments for the current user.", extra=logging_context())
-            query = query.filter(
-                Experiment.created_by_id == user_data.user.id
-            )  # .filter(Experiment.published_date is None)
-        # else:
-        #     query = query.filter(Experiment.created_by_id == user.id).filter(Experiment.published_date is None)
-    else:
-        logger.debug(msg="No query string was provided; Listing all experiments.", extra=logging_context())
+        logger.debug(msg="Listing experiments for the current user.", extra=logging_context())
+        query = query.filter(
+            or_(
+                Experiment.created_by_id == user_data.user.id,
+                Experiment.contributors.any(Contributor.orcid_id == user_data.user.username)
+            )
+        )
 
     items = query.order_by(Experiment.urn).all()
     return items

src/mavedb/routers/hgvs.py

@@ -11,7 +11,7 @@
 
 router = APIRouter(
     prefix="/api/v1/hgvs",
-    tags=["hgvs"],
+    tags=["transcripts"],
     responses={404: {"description": "Not found"}},
 )
 
@@ -85,7 +85,7 @@ def gene_info(gene: str, hdp: RESTDataProvider = Depends(hgvs_data_provider)) ->
     return gene_info
 
 
-@router.get("/transcripts/gene/{gene}", status_code=200, response_model=list[str])
+@router.get("/gene/{gene}", status_code=200, response_model=list[str])
 def list_transcripts_for_gene(gene: str, hdp: RESTDataProvider = Depends(hgvs_data_provider)) -> list[str]:
     """
     List transcripts associated with a particular gene
@@ -98,7 +98,7 @@ def list_transcripts_for_gene(gene: str, hdp: RESTDataProvider = Depends(hgvs_da
     return list(transcripts)
 
 
-@router.get("/transcripts/{transcript}", status_code=200, response_model=dict[str, Any])
+@router.get("/{transcript}", status_code=200, response_model=dict[str, Any])
 def transcript_info(transcript: str, hdp: RESTDataProvider = Depends(hgvs_data_provider)) -> dict[str, Any]:
     """
     List transcript information for a particular transcript
@@ -111,7 +111,7 @@ def transcript_info(transcript: str, hdp: RESTDataProvider = Depends(hgvs_data_p
     return transcript_info
 
 
-@router.get("/transcripts/protein/{transcript}", status_code=200, response_model=str)
+@router.get("/protein/{transcript}", status_code=200, response_model=str)
 def convert_to_protein(transcript: str, hdp: RESTDataProvider = Depends(hgvs_data_provider)) -> str:
     """
     Convert a provided transcript from it's nucleotide accession identifier to its protein accession identifier

src/mavedb/view_models/search.py

@@ -11,6 +11,7 @@ class ExperimentsSearch(BaseModel):
     publication_identifiers: Optional[list[str]]
     keywords: Optional[list[str]]
     text: Optional[str]
+    meta_analysis: Optional[bool]
 
 
 class ScoreSetsSearch(BaseModel):

tests/routers/test_experiments.py

@@ -973,6 +973,58 @@ def test_search_my_experiments(session, client, setup_router_db):
     assert response.json()[0]["title"] == experiment["title"]
 
 
+def test_search_meta_analysis_experiment(session, data_provider, client, setup_router_db, data_files):
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set_with_variants(
+        client, session, data_provider, experiment["urn"], data_files / "scores.csv"
+    )
+
+    score_set = (client.post(f"/api/v1/score-sets/{score_set['urn']}/publish")).json()
+    meta_score_set = create_seq_score_set_with_variants(
+        client,
+        session,
+        data_provider,
+        None,
+        data_files / "scores.csv",
+        update={"title": "Test Meta Analysis", "metaAnalyzesScoreSetUrns": [score_set["urn"]]},
+    )
+
+    meta_score_set = (client.post(f"/api/v1/score-sets/{meta_score_set['urn']}/publish")).json()
+    score_set_refresh = (client.get(f"/api/v1/score-sets/{score_set['urn']}")).json()
+    search_payload = {"metaAnalysis": True}
+    response = client.post("/api/v1/me/experiments/search", json=search_payload)
+    assert response.status_code == 200
+    response_data = response.json()
+    assert any(item["urn"] == meta_score_set["experiment"]["urn"] for item in response_data)
+    assert all(item["urn"] != score_set_refresh["experiment"]["urn"] for item in response_data)
+
+
+def test_search_exclude_meta_analysis_experiment(session, data_provider, client, setup_router_db, data_files):
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set_with_variants(
+        client, session, data_provider, experiment["urn"], data_files / "scores.csv"
+    )
+
+    score_set = (client.post(f"/api/v1/score-sets/{score_set['urn']}/publish")).json()
+    meta_score_set = create_seq_score_set_with_variants(
+        client,
+        session,
+        data_provider,
+        None,
+        data_files / "scores.csv",
+        update={"title": "Test Meta Analysis", "metaAnalyzesScoreSetUrns": [score_set["urn"]]},
+    )
+
+    meta_score_set = (client.post(f"/api/v1/score-sets/{meta_score_set['urn']}/publish")).json()
+    score_set_refresh = (client.get(f"/api/v1/score-sets/{score_set['urn']}")).json()
+    search_payload = {"metaAnalysis": False}
+    response = client.post("/api/v1/me/experiments/search", json=search_payload)
+    assert response.status_code == 200
+    response_data = response.json()
+    assert any(item["urn"] == score_set_refresh["experiment"]["urn"] for item in response_data)
+    assert all(item["urn"] != meta_score_set["experiment"]["urn"] for item in response_data)
+
+
 def test_search_score_sets_for_experiments(session, client, setup_router_db, data_files, data_provider):
     experiment = create_experiment(client)
     score_set_pub = create_seq_score_set_with_variants(

tests/routers/test_hgvs.py

@@ -122,7 +122,7 @@ def test_hgvs_gene_transcript_valid(client, setup_router_db):
             json={"results": [{"hgnc": f"{VALID_GENE}", "tx_ac": VALID_TRANSCRIPT}]},
         )
 
-        response = client.get(f"/api/v1/hgvs/transcripts/gene/{VALID_GENE}")
+        response = client.get(f"/api/v1/hgvs/gene/{VALID_GENE}")
         assert response.status_code == 200
         assert VALID_TRANSCRIPT in response.json()
 
@@ -131,15 +131,15 @@ def test_hgvs_gene_transcript_invalid(client, setup_router_db):
     with requests_mock.mock() as m:
         m.get(f"https://cdot.cc/transcripts/gene/{INVALID_GENE}", status_code=404)
 
-        response = client.get(f"/api/v1/hgvs/transcripts/gene/{INVALID_GENE}")
+        response = client.get(f"/api/v1/hgvs/gene/{INVALID_GENE}")
 
         assert m.called
         assert response.status_code == 404
 
 
 def test_hgvs_transcript_valid(client, setup_router_db):
     with patch.object(cdot.hgvs.dataproviders.RESTDataProvider, "_get_transcript", return_value=TEST_CDOT_TRANSCRIPT):
-        response = client.get(f"/api/v1/hgvs/transcripts/{VALID_TRANSCRIPT}")
+        response = client.get(f"/api/v1/hgvs/{VALID_TRANSCRIPT}")
 
     assert response.status_code == 200
     assert response.json()["hgnc"] == VALID_GENE
@@ -149,7 +149,7 @@ def test_hgvs_transcript_invalid(client, setup_router_db):
     with requests_mock.mock() as m:
         m.get(f"https://cdot.cc/transcript/{INVALID_TRANSCRIPT}", status_code=404)
 
-        response = client.get(f"/api/v1/hgvs/transcripts/{INVALID_TRANSCRIPT}")
+        response = client.get(f"/api/v1/hgvs/{INVALID_TRANSCRIPT}")
 
         assert m.called
         assert response.status_code == 404
@@ -163,7 +163,7 @@ def test_hgvs_transcript_protein_valid(client, setup_router_db):
             json={"biotype": ["protein_coding"], "gene_name": "A2M", "gene_vesion": "2", "protein": "NP_000005.2"},
         )
 
-        response = client.get(f"/api/v1/hgvs/transcripts/protein/{HAS_PROTEIN_ACCESSION}")
+        response = client.get(f"/api/v1/hgvs/protein/{HAS_PROTEIN_ACCESSION}")
 
         assert m.called
 
@@ -175,7 +175,7 @@ def test_hgvs_transcript_protein_no_protein(client, setup_router_db):
     with requests_mock.mock() as m:
         m.get(f"https://cdot.cc/transcript/{SMALL_ACCESSION}", status_code=404)
 
-        response = client.get(f"/api/v1/hgvs/transcripts/protein/{SMALL_ACCESSION}")
+        response = client.get(f"/api/v1/hgvs/protein/{SMALL_ACCESSION}")
 
         assert m.called
         assert response.status_code == 404
@@ -185,7 +185,7 @@ def test_hgvs_transcript_protein_invalid(client, setup_router_db):
     with requests_mock.mock() as m:
         m.get(f"https://cdot.cc/transcript/{INVALID_ACCESSION}", status_code=404)
 
-        response = client.get(f"/api/v1/hgvs/transcripts/protein/{INVALID_ACCESSION}")
+        response = client.get(f"/api/v1/hgvs/protein/{INVALID_ACCESSION}")
 
         assert m.called
         assert response.status_code == 404