tudor/school_compare
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix migration script to handle percentage signs in CSV data
All checks were successful
Build and Push Docker Image / build-and-push (push) Successful in 55s
Details

Browse Source 
- Updated parse_numeric() to strip percentage signs before parsing
- This fixes the issue where all percentage metrics were showing as NULL/empty
- School cards will now display actual performance data after re-running migration
This commit is contained in:
Tudor Sitaru
2026-01-06 22:22:33 +00:00
parent 1a8ec670b9
commit e2b2ddfb66
1 changed files with 198 additions and 141 deletions
Download Patch File Download Diff File Expand all files Collapse all files

245
scripts/migrate_csv_to_db.py
View File

@@ -10,23 +10,28 @@ Options:
--geocode Geocode postcodes (requires network access)
"""
import sys
import os
import sys
from pathlib import Path
# Add parent directory to path for imports
sys.path.insert(0, str(Path(__file__).parent.parent))
import argparse
import pandas as pd
import numpy as np
import re
from typing import Optional, Dict
from typing import Dict, Optional
import numpy as np
import pandas as pd
import requests
from backend.config import settings
from backend.database import engine, Base, get_db_session
from backend.models import School, SchoolResult, SCHOOL_FIELD_MAPPING, RESULT_FIELD_MAPPING
from backend.models import (
School,
SchoolResult,
SCHOOL_FIELD_MAPPING,
RESULT_FIELD_MAPPING,
)
from backend.schemas import (
COLUMN_MAPPINGS,
NUMERIC_COLUMNS,
@@ -43,8 +48,10 @@ def parse_numeric(value) -> Optional[float]:
if isinstance(value, (int, float)):
return float(value) if not np.isnan(value) else None
str_val = str(value).strip().upper()
if str_val in NULL_VALUES or str_val == '':
if str_val in NULL_VALUES or str_val == "":
return None
# Remove percentage signs if present
str_val = str_val.replace("%", "")
try:
return float(str_val)
except ValueError:
@@ -53,10 +60,10 @@ def parse_numeric(value) -> Optional[float]:
def extract_year_from_folder(folder_name: str) -> Optional[int]:
"""Extract year from folder name like '2023-2024'."""
match = re.search(r'(\d{4})-(\d{4})', folder_name)
match = re.search(r"(\d{4})-(\d{4})", folder_name)
if match:
return int(match.group(2))
match = re.search(r'(\d{4})', folder_name)
match = re.search(r"(\d{4})", folder_name)
if match:
return int(match.group(1))
return None
@@ -68,7 +75,11 @@ def geocode_postcodes_bulk(postcodes: list) -> Dict[str, tuple]:
Returns dict of postcode -> (latitude, longitude).
"""
results = {}
valid_postcodes = [p.strip().upper() for p in postcodes if p and isinstance(p, str) and len(p.strip()) >= 5]
valid_postcodes = [
p.strip().upper()
for p in postcodes
if p and isinstance(p, str) and len(p.strip()) >= 5
]
valid_postcodes = list(set(valid_postcodes))
if not valid_postcodes:
@@ -79,21 +90,23 @@ def geocode_postcodes_bulk(postcodes: list) -> Dict[str, tuple]:
for i, batch_start in enumerate(range(0, len(valid_postcodes), batch_size)):
batch = valid_postcodes[batch_start : batch_start + batch_size]
print(f" Geocoding batch {i+1}/{total_batches} ({len(batch)} postcodes)...")
print(
f" Geocoding batch {i + 1}/{total_batches} ({len(batch)} postcodes)..."
)
try:
response = requests.post(
'https://api.postcodes.io/postcodes',
json={'postcodes': batch},
timeout=30
"https://api.postcodes.io/postcodes",
json={"postcodes": batch},
timeout=30,
)
if response.status_code == 200:
data = response.json()
for item in data.get('result', []):
if item and item.get('result'):
pc = item['query'].upper()
lat = item['result'].get('latitude')
lon = item['result'].get('longitude')
for item in data.get("result", []):
if item and item.get("result"):
pc = item["query"].upper()
lat = item["result"].get("latitude")
lon = item["result"].get("longitude")
if lat and lon:
results[pc] = (lat, lon)
except Exception as e:
@@ -123,42 +136,59 @@ def load_csv_data(data_dir: Path) -> pd.DataFrame:
print(f" Loading {csv_file.name} (year {year})...")
try:
df = pd.read_csv(csv_file, encoding='latin-1', low_memory=False)
df = pd.read_csv(csv_file, encoding="latin-1", low_memory=False)
except Exception as e:
print(f" Error loading {csv_file}: {e}")
continue
# Rename columns
df.rename(columns=COLUMN_MAPPINGS, inplace=True)
df['year'] = year
df["year"] = year
# Handle local authority name
la_name_cols = ['LANAME', 'LA (name)', 'LA_NAME', 'LA NAME']
la_name_cols = ["LANAME", "LA (name)", "LA_NAME", "LA NAME"]
la_name_col = next((c for c in la_name_cols if c in df.columns), None)
if la_name_col and la_name_col != 'local_authority':
df['local_authority'] = df[la_name_col]
elif 'LEA' in df.columns:
df['local_authority_code'] = pd.to_numeric(df['LEA'], errors='coerce')
df['local_authority'] = df['local_authority_code'].map(LA_CODE_TO_NAME).fillna(df['LEA'].astype(str))
if la_name_col and la_name_col != "local_authority":
df["local_authority"] = df[la_name_col]
elif "LEA" in df.columns:
df["local_authority_code"] = pd.to_numeric(df["LEA"], errors="coerce")
df["local_authority"] = (
df["local_authority_code"]
.map(LA_CODE_TO_NAME)
.fillna(df["LEA"].astype(str))
)
# Store LEA code
if 'LEA' in df.columns:
df['local_authority_code'] = pd.to_numeric(df['LEA'], errors='coerce')
if "LEA" in df.columns:
df["local_authority_code"] = pd.to_numeric(df["LEA"], errors="coerce")
# Map school type
if 'school_type_code' in df.columns:
df['school_type'] = df['school_type_code'].map(SCHOOL_TYPE_MAP).fillna(df['school_type_code'])
if "school_type_code" in df.columns:
df["school_type"] = (
df["school_type_code"]
.map(SCHOOL_TYPE_MAP)
.fillna(df["school_type_code"])
)
# Create combined address
addr_parts = ['address1', 'address2', 'town', 'postcode']
addr_parts = ["address1", "address2", "town", "postcode"]
for col in addr_parts:
if col not in df.columns:
df[col] = None
df['address'] = df.apply(
lambda r: ', '.join(str(v) for v in [r.get('address1'), r.get('address2'), r.get('town'), r.get('postcode')] if pd.notna(v) and str(v).strip()),
axis=1
df["address"] = df.apply(
lambda r: ", ".join(
str(v)
for v in [
r.get("address1"),
r.get("address2"),
r.get("town"),
r.get("postcode"),
]
if pd.notna(v) and str(v).strip()
),
axis=1,
)
all_data.append(df)
@@ -179,19 +209,21 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
# Clean URN column - convert to integer, drop invalid values
df = df.copy()
df['urn'] = pd.to_numeric(df['urn'], errors='coerce')
df = df.dropna(subset=['urn'])
df['urn'] = df['urn'].astype(int)
df["urn"] = pd.to_numeric(df["urn"], errors="coerce")
df = df.dropna(subset=["urn"])
df["urn"] = df["urn"].astype(int)
# Group by URN to get unique schools (use latest year's data)
school_data = df.sort_values('year', ascending=False).groupby('urn').first().reset_index()
school_data = (
df.sort_values("year", ascending=False).groupby("urn").first().reset_index()
)
print(f"\nMigrating {len(school_data)} unique schools...")
# Geocode if requested
geocoded = {}
if geocode and 'postcode' in df.columns:
if geocode and "postcode" in df.columns:
print("\nGeocoding postcodes...")
postcodes = df['postcode'].dropna().unique().tolist()
postcodes = df["postcode"].dropna().unique().tolist()
geocoded = geocode_postcodes_bulk(postcodes)
print(f" Successfully geocoded {len(geocoded)} postcodes")
@@ -202,7 +234,7 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
for _, row in school_data.iterrows():
# Safely parse URN - handle None, NaN, whitespace, and invalid values
urn_val = row.get('urn')
urn_val = row.get("urn")
urn = None
if pd.notna(urn_val):
try:
@@ -219,7 +251,7 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
continue
# Get geocoding data
postcode = row.get('postcode')
postcode = row.get("postcode")
lat, lon = None, None
if postcode and pd.notna(postcode):
coords = geocoded.get(str(postcode).strip().upper())
@@ -228,7 +260,7 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
# Safely parse local_authority_code
la_code = None
la_code_val = row.get('local_authority_code')
la_code_val = row.get("local_authority_code")
if pd.notna(la_code_val):
try:
la_code_str = str(la_code_val).strip()
@@ -239,17 +271,29 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
school = School(
urn=urn,
school_name=row.get('school_name') if pd.notna(row.get('school_name')) else 'Unknown',
local_authority=row.get('local_authority') if pd.notna(row.get('local_authority')) else None,
school_name=row.get("school_name")
if pd.notna(row.get("school_name"))
else "Unknown",
local_authority=row.get("local_authority")
if pd.notna(row.get("local_authority"))
else None,
local_authority_code=la_code,
school_type=row.get('school_type') if pd.notna(row.get('school_type')) else None,
school_type_code=row.get('school_type_code') if pd.notna(row.get('school_type_code')) else None,
religious_denomination=row.get('religious_denomination') if pd.notna(row.get('religious_denomination')) else None,
age_range=row.get('age_range') if pd.notna(row.get('age_range')) else None,
address1=row.get('address1') if pd.notna(row.get('address1')) else None,
address2=row.get('address2') if pd.notna(row.get('address2')) else None,
town=row.get('town') if pd.notna(row.get('town')) else None,
postcode=row.get('postcode') if pd.notna(row.get('postcode')) else None,
school_type=row.get("school_type")
if pd.notna(row.get("school_type"))
else None,
school_type_code=row.get("school_type_code")
if pd.notna(row.get("school_type_code"))
else None,
religious_denomination=row.get("religious_denomination")
if pd.notna(row.get("religious_denomination"))
else None,
age_range=row.get("age_range")
if pd.notna(row.get("age_range"))
else None,
address1=row.get("address1") if pd.notna(row.get("address1")) else None,
address2=row.get("address2") if pd.notna(row.get("address2")) else None,
town=row.get("town") if pd.notna(row.get("town")) else None,
postcode=row.get("postcode") if pd.notna(row.get("postcode")) else None,
latitude=lat,
longitude=lon,
)
@@ -269,7 +313,7 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
for _, row in df.iterrows():
# Safely parse URN
urn_val = row.get('urn')
urn_val = row.get("urn")
urn = None
if pd.notna(urn_val):
try:
@@ -284,7 +328,7 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
school_id = urn_to_school_id[urn]
# Safely parse year
year_val = row.get('year')
year_val = row.get("year")
year = None
if pd.notna(year_val):
try:
@@ -297,48 +341,52 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
result = SchoolResult(
school_id=school_id,
year=year,
total_pupils=parse_numeric(row.get('total_pupils')),
eligible_pupils=parse_numeric(row.get('eligible_pupils')),
total_pupils=parse_numeric(row.get("total_pupils")),
eligible_pupils=parse_numeric(row.get("eligible_pupils")),
# Expected Standard
rwm_expected_pct=parse_numeric(row.get('rwm_expected_pct')),
reading_expected_pct=parse_numeric(row.get('reading_expected_pct')),
writing_expected_pct=parse_numeric(row.get('writing_expected_pct')),
maths_expected_pct=parse_numeric(row.get('maths_expected_pct')),
gps_expected_pct=parse_numeric(row.get('gps_expected_pct')),
science_expected_pct=parse_numeric(row.get('science_expected_pct')),
rwm_expected_pct=parse_numeric(row.get("rwm_expected_pct")),
reading_expected_pct=parse_numeric(row.get("reading_expected_pct")),
writing_expected_pct=parse_numeric(row.get("writing_expected_pct")),
maths_expected_pct=parse_numeric(row.get("maths_expected_pct")),
gps_expected_pct=parse_numeric(row.get("gps_expected_pct")),
science_expected_pct=parse_numeric(row.get("science_expected_pct")),
# Higher Standard
rwm_high_pct=parse_numeric(row.get('rwm_high_pct')),
reading_high_pct=parse_numeric(row.get('reading_high_pct')),
writing_high_pct=parse_numeric(row.get('writing_high_pct')),
maths_high_pct=parse_numeric(row.get('maths_high_pct')),
gps_high_pct=parse_numeric(row.get('gps_high_pct')),
rwm_high_pct=parse_numeric(row.get("rwm_high_pct")),
reading_high_pct=parse_numeric(row.get("reading_high_pct")),
writing_high_pct=parse_numeric(row.get("writing_high_pct")),
maths_high_pct=parse_numeric(row.get("maths_high_pct")),
gps_high_pct=parse_numeric(row.get("gps_high_pct")),
# Progress
reading_progress=parse_numeric(row.get('reading_progress')),
writing_progress=parse_numeric(row.get('writing_progress')),
maths_progress=parse_numeric(row.get('maths_progress')),
reading_progress=parse_numeric(row.get("reading_progress")),
writing_progress=parse_numeric(row.get("writing_progress")),
maths_progress=parse_numeric(row.get("maths_progress")),
# Averages
reading_avg_score=parse_numeric(row.get('reading_avg_score')),
maths_avg_score=parse_numeric(row.get('maths_avg_score')),
gps_avg_score=parse_numeric(row.get('gps_avg_score')),
reading_avg_score=parse_numeric(row.get("reading_avg_score")),
maths_avg_score=parse_numeric(row.get("maths_avg_score")),
gps_avg_score=parse_numeric(row.get("gps_avg_score")),
# Context
disadvantaged_pct=parse_numeric(row.get('disadvantaged_pct')),
eal_pct=parse_numeric(row.get('eal_pct')),
sen_support_pct=parse_numeric(row.get('sen_support_pct')),
sen_ehcp_pct=parse_numeric(row.get('sen_ehcp_pct')),
stability_pct=parse_numeric(row.get('stability_pct')),
disadvantaged_pct=parse_numeric(row.get("disadvantaged_pct")),
eal_pct=parse_numeric(row.get("eal_pct")),
sen_support_pct=parse_numeric(row.get("sen_support_pct")),
sen_ehcp_pct=parse_numeric(row.get("sen_ehcp_pct")),
stability_pct=parse_numeric(row.get("stability_pct")),
# Gender
rwm_expected_boys_pct=parse_numeric(row.get('rwm_expected_boys_pct')),
rwm_expected_girls_pct=parse_numeric(row.get('rwm_expected_girls_pct')),
rwm_high_boys_pct=parse_numeric(row.get('rwm_high_boys_pct')),
rwm_high_girls_pct=parse_numeric(row.get('rwm_high_girls_pct')),
rwm_expected_boys_pct=parse_numeric(row.get("rwm_expected_boys_pct")),
rwm_expected_girls_pct=parse_numeric(row.get("rwm_expected_girls_pct")),
rwm_high_boys_pct=parse_numeric(row.get("rwm_high_boys_pct")),
rwm_high_girls_pct=parse_numeric(row.get("rwm_high_girls_pct")),
# Disadvantaged
rwm_expected_disadvantaged_pct=parse_numeric(row.get('rwm_expected_disadvantaged_pct')),
rwm_expected_non_disadvantaged_pct=parse_numeric(row.get('rwm_expected_non_disadvantaged_pct')),
disadvantaged_gap=parse_numeric(row.get('disadvantaged_gap')),
rwm_expected_disadvantaged_pct=parse_numeric(
row.get("rwm_expected_disadvantaged_pct")
),
rwm_expected_non_disadvantaged_pct=parse_numeric(
row.get("rwm_expected_non_disadvantaged_pct")
),
disadvantaged_gap=parse_numeric(row.get("disadvantaged_gap")),
# 3-Year
rwm_expected_3yr_pct=parse_numeric(row.get('rwm_expected_3yr_pct')),
reading_avg_3yr=parse_numeric(row.get('reading_avg_3yr')),
maths_avg_3yr=parse_numeric(row.get('maths_avg_3yr')),
rwm_expected_3yr_pct=parse_numeric(row.get("rwm_expected_3yr_pct")),
reading_avg_3yr=parse_numeric(row.get("reading_avg_3yr")),
maths_avg_3yr=parse_numeric(row.get("maths_avg_3yr")),
)
db.add(result)
results_created += 1
@@ -355,9 +403,13 @@ def migrate_data(df: pd.DataFrame, geocode: bool = False):
def main():
parser = argparse.ArgumentParser(description='Migrate CSV data to PostgreSQL database')
parser.add_argument('--drop', action='store_true', help='Drop existing tables before migration')
parser.add_argument('--geocode', action='store_true', help='Geocode postcodes')
parser = argparse.ArgumentParser(
description="Migrate CSV data to PostgreSQL database"
)
parser.add_argument(
"--drop", action="store_true", help="Drop existing tables before migration"
)
parser.add_argument("--geocode", action="store_true", help="Geocode postcodes")
args = parser.parse_args()
print("=" * 60)
@@ -385,6 +437,11 @@ def main():
return 0
if __name__ == "__main__":
migrate_data(df, geocode=args.geocode)
return 0
if __name__ == '__main__':
sys.exit(main())