From 68367f04c37f57784941e87a4e7958783d0d9c49 Mon Sep 17 00:00:00 2001
From: mpabi <mpabi@fake.pl>
Date: Thu, 20 Jun 2024 20:19:38 +0000
Subject: [PATCH] sequences and api are ok

---
 app/models.py |  3 +--
 entrypoint.py | 36 +++++++++++++++++++++++-------------
 2 files changed, 24 insertions(+), 15 deletions(-)

diff --git a/app/models.py b/app/models.py
index b847657..6dd3152 100644
--- a/app/models.py
+++ b/app/models.py
@@ -1,7 +1,6 @@
-# app/models.py
 from sqlalchemy import Column, Integer, String, Text, ForeignKey
 from sqlalchemy.orm import relationship
-from .database import Base
+from app.database import Base
 
 class Sequence(Base):
     __tablename__ = "sequences"
diff --git a/entrypoint.py b/entrypoint.py
index f76d8a4..5cce905 100644
--- a/entrypoint.py
+++ b/entrypoint.py
@@ -1,5 +1,3 @@
-#%%
-# entrypoint.py
 from Bio import SeqIO
 from sqlalchemy.orm import Session
 from app.database import engine, SessionLocal, Base
@@ -9,7 +7,6 @@ import json
 # Tworzenie tabeli w bazie danych
 Base.metadata.create_all(bind=engine)
 
-#%%
 def get_db():
     db = SessionLocal()
     try:
@@ -25,25 +22,38 @@ def load_genbank_to_db(file_path: str):
         # Zapisz sekwencję do bazy danych
         sequence_data = Sequence(
             name=record.name,
-            description=record.description,
+            description=record.description if 'description' in record.annotations else 'No description',
             sequence=str(record.seq)
         )
         db.add(sequence_data)
         db.commit()
+        db.refresh(sequence_data)  # Odśwież sekwencję, aby uzyskać jej ID
 
         # Zapisz cechy (features) do bazy danych
         for feature in record.features:
-            feature_data = Feature(
-                type=feature.type,
-                location=str(feature.location),
-                sequence=str(record.seq[feature.location.start:feature.location.end]),
-                qualifiers=json.dumps(feature.qualifiers),
-                sequence_id=sequence_data.id
-            )
-            db.add(feature_data)
+            if feature.type in ["promoter", "RBS","CDS", "protein_bind", "misc_feature", "rep_origin", "terminator"]:
+                qualifiers = {}
+                if feature.type == "CDS":
+                    qualifiers['gene_name'] = feature.qualifiers.get('gene', ['Unknown gene'])[0]
+                    qualifiers['product'] = feature.qualifiers.get('product', ['Unknown product'])[0]
+                elif feature.type == "protein_bind":
+                    qualifiers['binding_site'] = feature.qualifiers.get('bound_moiety', ['Unknown binding site'])[0]
+                elif feature.type == "misc_feature":
+                    qualifiers['note'] = feature.qualifiers.get('note', ['No additional information'])[0]
+                elif feature.type == "rep_origin":
+                    qualifiers['note'] = "This is a replication origin."
+                
+                feature_data = Feature(
+                    type=feature.type,
+                    location=str(feature.location),
+                    sequence=str(record.seq[feature.location.start:feature.location.end]),
+                    qualifiers=json.dumps(qualifiers),
+                    sequence_id=sequence_data.id  # Ustawienie poprawnego ID
+                )
+                db.add(feature_data)
         db.commit()
     print(f"Loaded {file_path} to database.")
 
 if __name__ == "__main__":
-    load_genbank_to_db("data/pET-28+(a).gb")
+    load_genbank_to_db("data/plasmid.gb")