Merge pull request #4 from MultiQC/buffer-visits

Summarise visits by minute; buffer in memory and in CSV before DB

.gitignore

@@ -138,3 +138,5 @@ dmypy.json
 .idea/
 
 .DS_Store
+
+visits.csv

app/__init__.py

@@ -1,3 +1,7 @@
 """api.multiqc.info: Providing run-time information about available updates."""
 
 __version__ = "0.1.0.dev0"
+
+from dotenv import load_dotenv
+
+load_dotenv()

app/db.py

@@ -1,52 +1,51 @@
 """Functions to interact with the database."""
+import os
 
 import datetime
-from os import getenv
 
 from sqlmodel import create_engine, Field, select, Session, SQLModel
 
-sql_url = getenv("DATABASE_URL")
+sql_url = os.getenv("DATABASE_URL")
+assert sql_url is not None, sql_url
 engine = create_engine(sql_url)
 
 
-class Visit(SQLModel, table=True):  # type: ignore # mypy doesn't like this, not sure why
-    """Table to record raw individual visits to the version endpoint."""
+class Visits(SQLModel, table=True):  # type: ignore # mypy doesn't like this, not sure why
+    """
+    Table to record per-minute visit summaries.
+    
+    Start is a primary key, and start and end are both indexed.
+    """
 
-    id: int | None = Field(default=None, primary_key=True)
-    version_multiqc: str | None = None
-    version_python: str | None = None
-    operating_system: str | None = None
-    installation_method: str | None = None
-    ci_environment: str | None = None
-    called_at: datetime.datetime | None = Field(default_factory=datetime.datetime.utcnow, nullable=False)
+    start: datetime.datetime = Field(primary_key=True)
+    end: datetime.datetime = Field(primary_key=True)
+    count: int
+    version_multiqc: str = Field(index=True)
+    version_python: str = Field(default=None, index=True)
+    operating_system: str = Field(default=None, index=True)
+    installation_method: str = Field(default=None, index=True)
+    ci_environment: str = Field(default=None, index=True)
 
 
 def create_db_and_tables() -> None:
     """Create the database and tables if they don't exist."""
     SQLModel.metadata.create_all(engine)
 
 
-def add_visit(visit: Visit) -> None:
-    """Add a visit to the database."""
-    with Session(engine) as session:
-        session.add(visit)
-        session.commit()
-
-
 def get_visits(
     start: datetime.datetime | None = None,
     end: datetime.datetime | None = None,
     limit: int | None = None,
-) -> list[Visit]:
-    """Return list of raw visits from the DB."""
+) -> list[Visits]:
+    """Return list of per-minute visit summary from the DB."""
     with Session(engine) as session:
-        statement = select(Visit)
+        statement = select(Visits)
         if start:
             # Ignore type because Visit.called_at can be None for default value
-            statement.where(Visit.called_at > start)  # type: ignore
+            statement.where(Visits.start >= start)  # type: ignore
         if end:
-            statement.where(Visit.called_at < end)  # type: ignore
+            statement.where(Visits.end <= end)  # type: ignore
         if limit:
             statement.limit(limit)
-        statement.order_by(Visit.called_at.desc())  # type: ignore
+        statement.order_by(Visits.start.desc())  # type: ignore
         return session.exec(statement).all()