Datasets:

Modalities:
Image
Text
Formats:
parquet
Languages:
English
Libraries:
Datasets
Dask
License:
openaccess / add_images.py
Brett Renfer
Fixes for broken rows; full dataset now
5d110fb
import pandas as pd
import gzip
import csv
import requests
from requests.adapters import HTTPAdapter, Retry
import urllib3
import urllib.parse
from io import StringIO
# NOTE: this is not a good idea; this is solely a fix for Met networks
do_verify = False
if not do_verify:
urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)
# Setup HTTPAdaptor & requests session to add retry pattern
s = requests.Session()
retries = Retry(total=3,
backoff_factor=0.1,
status_forcelist=[ 500, 502, 503, 504 ])
s.mount('https://', HTTPAdapter(max_retries=retries))
# Function to load and clean the CSV data & add images
def load_and_clean_csv(file_path):
valid_rows = []
invalid_rows = []
index = 0
# Read the gzip file line by line
with gzip.open(file_path, 'rt', newline='\r\n', encoding='utf-8') as f:
reader = csv.reader(f)
header = next(reader) # Read the header separately
header.append("primaryImageSmall") # Add the new column to the header
valid_rows.append(header)
expected_columns = len(header) - 1 # Exclude the new column
for line in f:
try:
# Try to parse the line
row = next(csv.reader([line]))
index = index + 1
# print(len(row)+":"+expected_columns)
if len(row) == expected_columns:
# Fetch primaryImageSmall from the API
object_id = row[4]
image_url = fetch_primary_image_small(object_id)
image_url = image_url.replace(" ","%20")
image_url = image_url.replace(u'\u2013',"–")
row.append(image_url)
valid_rows.append(row)
if index % 100 == 0:
print("Fetched " + str(index) +" image URLs")
else:
print("Invalid: "+object_id)
print(row)
invalid_rows.append(line)
except Exception as e:
print(e)
print("Invalid + error: "+object_id)
invalid_rows.append(line)
print(f"Found {len(invalid_rows)} invalid rows")
return valid_rows, invalid_rows
# Function to load and clean the CSV data & add images
def test_csv(file_path):
valid_rows = []
invalid_rows = []
index = 0
# Read the gzip file line by line
with gzip.open(file_path, 'rt', newline='\r\n', encoding='utf-8') as f:
reader = csv.reader(f)
header = next(reader) # Read the header separately
valid_rows.append(header)
expected_columns = len(header)
for line in f:
try:
# Try to parse the line
row = next(csv.reader([line]))
index = index + 1
if len(row) == expected_columns:
object_id = row[4]
print(object_id)
valid_rows.append(row)
else:
print("Invalid: "+object_id)
print(len(row), expected_columns)
print(row)
invalid_rows.append(line)
except Exception as e:
print(e)
print("Invalid + error: "+object_id)
invalid_rows.append(line)
print(f"Found {len(invalid_rows)} invalid rows")
return valid_rows, invalid_rows
# Function to fetch the primaryImageSmall URL from the MET Museum API
def fetch_primary_image_small(object_id):
url = f"https://collectionapi.metmuseum.org/public/collection/v1/objects/{object_id}"
try:
response = s.get(url, verify=do_verify)
response.raise_for_status() # Raise an error for bad status codes
data = response.json()
# print (data.get("primaryImageSmall", ""))
return data.get("primaryImageSmall", "")
except Exception as e:
print(f"Error fetching image for object ID {object_id}: {e}")
return ""
# Function to save the cleaned data to a new gzip CSV file
def save_cleaned_csv(valid_rows, output_path):
with gzip.open(output_path, 'wt', newline='') as f:
writer = csv.writer(f)
writer.writerows(valid_rows)
print(f"Cleaned data saved to {output_path}")
def main():
input_file = 'metadata.csv.gz'
output_file = 'metadata_images.csv.gz'
# Test
# test_csv(input_file)
# Load and clean the data
valid_rows, invalid_rows = load_and_clean_csv(input_file)
# Save the cleaned data
save_cleaned_csv(valid_rows, output_file)
if __name__ == "__main__":
main()