Upload folder using huggingface_hub

.gitattributes

@@ -53,3 +53,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.jpg filter=lfs diff=lfs merge=lfs -text
 *.jpeg filter=lfs diff=lfs merge=lfs -text
 *.webp filter=lfs diff=lfs merge=lfs -text
+metadata.jsonl filter=lfs diff=lfs merge=lfs -text

collect-hkm.py

@@ -0,0 +1,68 @@
+#!/usr/bin/env python
+
+from finna_client import FinnaClient, IMAGE_BASE
+import requests
+import json
+import os
+import os.path
+import logging
+
+finna = FinnaClient()
+logging.basicConfig(format='%(asctime)s - %(message)s', level=logging.INFO)
+
+MAX_YEAR = 1917  # retrieve only images older than this (inclusive)
+IMAGE_DIR = 'images'
+
+# Facet filters that define the set of images we want to 
+FILTERS = [
+    '~format_ext_str_mv:"1/Image/Photo/"',	# Photographs
+    '~building:"0/HKM/"',	# from HKM (Helsinki City Museum)
+    '~usage_rights_ext_str_mv:"0/B BY/"',	# CC By license
+    f'search_daterange_mv:"overlap|[* TO {MAX_YEAR}]"', # date up to MAX_YEAR
+]
+
+# Fields to include in the metadata records
+FIELDS = [
+    'formats',
+    'id',
+    'imageRights',
+    'images',
+    'languages',
+    'nonPresenterAuthors',
+    'onlineUrls',
+    'presenters',
+    'rating',
+    'series',
+    'subjects',
+    'title',
+    'year',
+    'rawData',
+]
+
+def download_url_to_file(url, filename):
+    if os.path.exists(filename) and os.stat(filename).st_size > 0:
+        return  # file already exists, no need to re-download
+    with requests.get(url, stream=True) as r:
+        r.raise_for_status()
+        with open(filename, 'wb') as f:
+            for chunk in r.iter_content(chunk_size=8192): 
+                f.write(chunk)
+
+with open('metadata.jsonl', 'w') as outfile:
+    page = 1
+
+    while True:
+        logging.info(f'Loading result page {page}')
+        result = finna.search(filters=FILTERS, fields=FIELDS, page=page)
+        if 'records' not in result:
+            logging.info('No more records, stopping.')
+            break
+        logging.debug(f'Got {len(result["records"])} records')
+        for record in result['records']:
+            image_url = IMAGE_BASE + record['images'][0]
+            image_file_out = os.path.join(IMAGE_DIR, record['id'] + '.jpg')
+            download_url_to_file(image_url, image_file_out)
+            record['file_name'] = image_file_out
+            json.dump(record, outfile)
+            print("", file=outfile)
+        page += 1