#!/bin/sh

set -e

DB=francecrime.db
DEPT=department
REG=region
DEPT_CRIMES=department_crimes
rm -f "$DB"
sqlite-utils create-database "$DB"

# Region lookup table
curl -L https://www.insee.fr/fr/statistiques/fichier/6800675/v_region_2023.csv | \
	sqlite-utils insert "$DB" "$REG" - --csv --convert 'return {"code": row["REG"], "name": row["NCCENR"]}'
sqlite-utils transform "$DB" "$REG" --pk code

# Department lookup table
curl -L https://www.insee.fr/fr/statistiques/fichier/6800675/v_departement_2023.csv | \
	sqlite-utils insert "$DB" "$DEPT" - --csv --convert 'return {"code": row["DEP"], "name": row["NCCENR"]}'
sqlite-utils transform "$DB" "$DEPT" --pk code
# TODO Add a way to filter out island possessions

# Crimes by department
curl -L https://www.data.gouv.fr/fr/datasets/r/acc332f6-92be-42af-9721-f3609bea8cfc | gzip -d | \
	sqlite-utils insert "$DB" "$DEPT_CRIMES" - --csv --delimiter=";" --convert '
	def full_year(year):
		return "20" + year
	def comma_float(s):
		return float(s.replace(",", "."))
	return {
		"class": row["classe"],
		"year": full_year(row["annee"]),
		"department": row["Code.département"],
		"region": row["Code.région"],
		"cases": int(row["faits"]),
		"population": int(row["POP"]),
		"rate_per_thousand": comma_float(row["tauxpourmille"])}'
sqlite-utils extract "$DB" "$DEPT_CRIMES" class
sqlite-utils transform "$DB" "$DEPT_CRIMES" --pk class_id --pk year --pk department
sqlite-utils add-foreign-keys "$DB" \
	"$DEPT_CRIMES" department "$DEPT" code \
	"$DEPT_CRIMES" region "$REG" code
# TODO Rename class_id to class and show English text using crime_classes.csv
# TODO Add some charts or something
# TODO Add stats by commune so you can compare them to each other and to overall departmental rates; the latter might show an urban-rural divide.

datasette "$DB" -o