updates to DB schema

Rappsilber-Laboratory · Aug 15, 2024 · be492b5 · be492b5
1 parent dbdf5fb
commit be492b5
Show file tree

Hide file tree

Showing 4 changed files with 22 additions and 12 deletions.
diff --git a/models/dbsequence.py b/models/dbsequence.py
@@ -5,8 +5,8 @@
 
 class DBSequence(Base):
     __tablename__ = "dbsequence"
-    id: Mapped[str] = mapped_column(Text, primary_key=True, nullable=False)
     upload_id: Mapped[str] = mapped_column(Integer, ForeignKey("upload.id"), primary_key=True, nullable=False)
+    id: Mapped[str] = mapped_column(Text, primary_key=True, nullable=False)
     accession: Mapped[str] = mapped_column(Text, nullable=False)
     name: Mapped[str] = mapped_column(Text, nullable=True)
     description: Mapped[str] = mapped_column(Text, nullable=True)

diff --git a/models/match.py b/models/match.py
@@ -1,5 +1,5 @@
 from sqlalchemy.orm import Mapped, mapped_column
-from sqlalchemy import ForeignKey, Text, FLOAT, JSON, BOOLEAN, Integer, ForeignKeyConstraint, CHAR
+from sqlalchemy import ForeignKey, Text, FLOAT, JSON, BOOLEAN, Integer, ForeignKeyConstraint, CHAR, Index
 from models.base import Base
 from typing import Optional, Any
 
@@ -9,6 +9,7 @@ class Match(Base):
     id: Mapped[str] = mapped_column(Text, primary_key=True, nullable=False)
     upload_id: Mapped[int] = mapped_column(Integer, ForeignKey("upload.id"), index=True, primary_key=True,
                                            nullable=False)
+    id: Mapped[str] = mapped_column(Text, primary_key=True, nullable=False)
     spectrum_id: Mapped[str] = mapped_column(Text, nullable=True)
     spectra_data_id: Mapped[int] = mapped_column(Integer, nullable=True)  # nullable for csv data
     multiple_spectra_identification_id: Mapped[str] = mapped_column(Integer, nullable=True)
@@ -29,16 +30,17 @@ class Match(Base):
              "spectrumidentificationprotocol.upload_id"],
         ),
         ForeignKeyConstraint(
-            ["pep1_id", "upload_id"],
-            ["modifiedpeptide.id", "modifiedpeptide.upload_id"],
+            [ "upload_id", "pep1_id",],
+            ["modifiedpeptide.upload_id", "modifiedpeptide.id"],
         ),
         ForeignKeyConstraint(
-            ["pep2_id", "upload_id"],
-            ["modifiedpeptide.id", "modifiedpeptide.upload_id"],
+            ["upload_id", "pep2_id"],
+            ["modifiedpeptide.upload_id", "modifiedpeptide.id", ],
         ),
         # ForeignKeyConstraint(
         # ["spectrum_id", "spectra_data_id", "upload_id"],
         # ["spectrum.id", "spectrum.spectra_data_id", "spectrum.upload_id"],
         # ),
+        Index("ix_match_id", "id"),
     )
 
diff --git a/models/modifiedpeptide.py b/models/modifiedpeptide.py
@@ -1,14 +1,14 @@
 from sqlalchemy.orm import Mapped, mapped_column
-from sqlalchemy import ForeignKey, Text, Integer, JSON, FLOAT
+from sqlalchemy import ForeignKey, Text, Integer, JSON, FLOAT, Index
 from models.base import Base
 from typing import Optional, Any
 
 
 class ModifiedPeptide(Base):
     __tablename__ = "modifiedpeptide"
-    id: Mapped[int] = mapped_column(Integer, primary_key=True, nullable=False)
     upload_id: Mapped[str] = mapped_column(Integer, ForeignKey("upload.id"), index=True, primary_key=True,
                                            nullable=False)
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, nullable=False)
     base_sequence: Mapped[str] = mapped_column(Text, nullable=False)
     mod_accessions: Mapped[dict[str, Any]] = mapped_column(JSON, nullable=False)
     mod_avg_mass_deltas: Mapped[Optional[dict[str, Any]]] = mapped_column(JSON, nullable=True)
@@ -21,3 +21,11 @@ class ModifiedPeptide(Base):
     crosslinker_modmass: Mapped[float] = mapped_column(FLOAT, nullable=True)
     crosslinker_pair_id: Mapped[str] = mapped_column(Text, nullable=True)  # yes, it's a string
     crosslinker_accession: Mapped[str] = mapped_column(Text, nullable=True)
+# CREATE INDEX ix_modifiedpeptide_id ON public.modifiedpeptide USING btree (id);
+# CREATE INDEX modifiedpeptide_upload_id__id_seq_idx ON public.modifiedpeptide (upload_id,id, base_sequence);
+# CREATE INDEX modifiedpeptide_upload_id_linksite_idx ON public.modifiedpeptide (upload_id, link_site1);
+    __table_args__ = (
+        Index("ix_modifiedpeptide_id", "id"),
+        Index("modifiedpeptide_upload_id__id_seq_idx", "upload_id", "id", "base_sequence"),
+        Index("modifiedpeptide_upload_id_linksite_idx", "upload_id", "link_site1"),
+    )
diff --git a/models/peptideevidence.py b/models/peptideevidence.py
@@ -13,12 +13,12 @@ class PeptideEvidence(Base):
     is_decoy: Mapped[bool] = mapped_column(BOOLEAN, nullable=True)
     __table_args__ = (
         ForeignKeyConstraint(
-            ("dbsequence_id", "upload_id"),
-            ("dbsequence.id", "dbsequence.upload_id"),
+            ("upload_id", "dbsequence_id"),
+            ("dbsequence.upload_id", "dbsequence.id"),
         ),
         ForeignKeyConstraint(
-            ("peptide_id", "upload_id"),
-            ("modifiedpeptide.id", "modifiedpeptide.upload_id"),
+            ("upload_id", "peptide_id"),
+            ("modifiedpeptide.upload_id", "modifiedpeptide.id"),
         ),
         # add index on upload_id, peptide_id
         Index("peptideevidence_upload_id_peptide_id_idx", "upload_id", "peptide_id"),