From 6a76814ec56902d437c65bd3e53e66063cf43a52 Mon Sep 17 00:00:00 2001
From: JustAnotherArchivist <JustAnotherArchivist@users.noreply.github.com>
Date: Tue, 20 Dec 2022 06:59:27 +0000
Subject: [PATCH] Add crude in-progress upload listing

---
 ia-upload-stream | 45 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 43 insertions(+), 2 deletions(-)

diff --git a/ia-upload-stream b/ia-upload-stream
index 01f5137..818606e 100755
--- a/ia-upload-stream
+++ b/ia-upload-stream
@@ -276,6 +276,40 @@ def upload(item, filename, metadata, *, iaConfigFile = None, partSize = 100*1024
 	logger.info('Done!')
 
 
+def list_uploads(item, *, tries = 3):
+	# No auth needed
+	url = f'https://s3.us.archive.org/{item}/?uploads'
+
+	# This endpoint redirects to the server storing the item under ia######.s3dns.us.archive.org, but those servers present an invalid TLS certificate for *.us.archive.org.
+	class IAS3CertificateFixHTTPAdapter(requests.adapters.HTTPAdapter):
+		def init_poolmanager(self, *args, **kwargs):
+			kwargs['assert_hostname'] = 's3.us.archive.org'
+			return super().init_poolmanager(*args, **kwargs)
+
+	for attempt in range(1, tries + 1):
+		r = requests.get(url, allow_redirects = False)
+		if r.status_code == 307 and '.s3dns.us.archive.org' in r.headers['Location']:
+			s3dnsUrl = r.headers['Location']
+			s3dnsUrl = s3dnsUrl.replace('http://', 'https://')
+			s3dnsUrl = s3dnsUrl.replace('.s3dns.us.archive.org:80/', '.s3dns.us.archive.org/')
+			domain = s3dnsUrl[8:s3dnsUrl.find('/', 9)]
+			s = requests.Session()
+			s.mount(f'https://{domain}/', IAS3CertificateFixHTTPAdapter())
+			r = s.get(s3dnsUrl)
+			if r.status_code == 200:
+				print(f'In-progress uploads for {item} (initiation datetime, upload ID, filename):')
+				for upload in re.findall(r'<Upload>.*?</Upload>', r.text):
+					uploadId = re.search(r'<UploadId>(.*?)</UploadId>', upload).group(1)
+					filename = re.search(r'<Key>(.*?)</Key>', upload).group(1)
+					date = re.search(r'<Initiated>(.*?)</Initiated>', upload).group(1)
+					print(f'{date} {uploadId} {filename}')
+				break
+		retrying = f', retrying' if attempt < tries else ''
+		logger.error(f'Could not list uploads; got status {r.status_code} from IA S3{retrying}')
+		if attempt == tries:
+			raise UploadError(f'Could not list uploads; got status {r.status_code} from IA S3', r = r)
+
+
 def abort(item, filename, uploadId, *, iaConfigFile = None, tries = 3):
 	# Read `ia` config
 	access, secret = get_ia_access_secret(iaConfigFile)
@@ -338,18 +372,23 @@ def main():
 	parser.add_argument('--upload-id', dest = 'uploadId', help = 'upload ID when resuming or aborting an upload')
 	parser.add_argument('--parts', type = parts, help = 'previous parts data for resumption; can only be used with --upload-id')
 	parser.add_argument('--abort', action = 'store_true', help = 'aborts an upload; can only be used with --upload-id; most other options are ignored when this is used')
+	parser.add_argument('--list', action = 'store_true', help = 'list in-progress uploads for item; most other options are ignored when this is used')
 	parser.add_argument('item', help = 'identifier of the target item')
-	parser.add_argument('filename', help = 'filename to store the data to')
+	parser.add_argument('filename', nargs = '?', help = 'filename to store the data to')
 	parser.add_argument('metadata', nargs = '*', type = metadata, help = "metadata for the item in the form 'key:value'; only has an effect if the item doesn't exist yet")
 	args = parser.parse_args()
 	if (args.parts or args.abort) and not args.uploadId:
 		parser.error('--parts and --abort can only be used together with --upload-id')
 	if args.uploadId and (args.parts is not None) == bool(args.abort):
 		parser.error('--upload-id requires exactly one of --parts and --abort')
+	if args.abort and args.list:
+		parser.error('--abort and --list cannot be used together')
+	if not args.list and not args.filename:
+		parser.error('filename is required when not using --list')
 
 	logging.basicConfig(level = logging.INFO, format = '{asctime}.{msecs:03.0f}  {levelname}  {name}  {message}', datefmt = '%Y-%m-%d %H:%M:%S', style = '{')
 	try:
-		if not args.abort:
+		if not args.abort and not args.list:
 			upload(
 				args.item,
 				args.filename,
@@ -365,6 +404,8 @@ def main():
 				parts = args.parts,
 				progress = args.progress,
 			)
+		elif args.list:
+			list_uploads(args.item, tries = args.tries)
 		else:
 			abort(
 				args.item,