From 9cdae30f7ef9c66b2b3a49738a737010d69d9b35 Mon Sep 17 00:00:00 2001
From: Marcel Schwarz <marcel@icaotix.de>
Date: Tue, 22 Dec 2020 02:25:56 +0100
Subject: [PATCH] Fix stupid accident dataset (remove duplicates?!?!)

---
 projects/project-3/backend/db_init.py           | 16 ++++++++++------
 projects/project-3/backend/routers/accidents.py |  9 ++++++++-
 2 files changed, 18 insertions(+), 7 deletions(-)

diff --git a/projects/project-3/backend/db_init.py b/projects/project-3/backend/db_init.py
index 51eae02..0bc2688 100644
--- a/projects/project-3/backend/db_init.py
+++ b/projects/project-3/backend/db_init.py
@@ -85,7 +85,8 @@ def init_database():
             lon REAL,
             location TEXT,
             date TEXT,
-            severity TEXT
+            severity TEXT,
+            UNIQUE (lat, lon, date)
         )""")
     conn.commit()
     conn.close()
@@ -143,7 +144,7 @@ def import_bikepoints():
 
 
 def import_accidents(year):
-    LOG.info("Importing accidents")
+    LOG.info(f"Importing accidents for year {year}")
     conn = sqlite3.connect(DB_NAME, timeout=300)
 
     def filter_pedal_cycles(accident):
@@ -155,13 +156,15 @@ def import_accidents(year):
     accidents = requests.get(f"https://api.tfl.gov.uk/AccidentStats/{year}").text
     accidents = json.loads(accidents)
     accidents = list(filter(filter_pedal_cycles, accidents))
-    accidents = list(map(lambda a: (a['id'], a['lat'], a['lon'], a['location'], a['date'], a['severity']), accidents))
+    accidents = list(map(lambda a: (a['lat'], a['lon'], a['location'], a['date'], a['severity']), accidents))
 
     LOG.info(f"Writing {len(accidents)} bike accidents to DB")
-    conn.executemany("INSERT OR IGNORE INTO accidents VALUES (?, ?, ?, ?, ?, ?)", accidents)
+    conn.executemany("""INSERT OR IGNORE INTO 
+                        accidents(lat, lon, location, date, severity) 
+                        VALUES (?, ?, ?, ?, ?)""", accidents)
     conn.commit()
     conn.close()
-    LOG.info("Accidents importet")
+    LOG.info(f"Accidents imported for year {year}")
 
 
 def import_usage_stats_file(export_file: ApiExportFile):
@@ -232,7 +235,8 @@ def main():
     # Import Bikepoints
     import_bikepoints()
     # Import bike accidents
-    import_accidents(2019)
+    for year in range(2005, 2020):
+        import_accidents(year)
 
     if count_after - count_pre > 0:
         create_dashboard_table()
diff --git a/projects/project-3/backend/routers/accidents.py b/projects/project-3/backend/routers/accidents.py
index e9fe942..1706823 100644
--- a/projects/project-3/backend/routers/accidents.py
+++ b/projects/project-3/backend/routers/accidents.py
@@ -1,4 +1,5 @@
 import logging
+from enum import Enum
 from typing import List
 
 from fastapi import APIRouter
@@ -10,10 +11,16 @@ router = APIRouter(prefix="/accidents", tags=["accidents", "local"])
 LOG = logging.getLogger()
 
 
+class Severity(str, Enum):
+    slight = "Slight"
+    serious = "Serious"
+    fatal = "Fatal"
+
+
 class Accident(BaseModel):
     lat: float
     lon: float
-    severity: str
+    severity: Severity
 
 
 @router.get(