declarative relay deployment on hetzner
relay-eval.waow.tech
atproto
relay
add collectiondir-diff script for comparing relay coverage
compares listReposByCollection results between two endpoints,
reports DIDs unique to each side. supports --resolve to look up
handles and PDS hosts via plc.directory.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
+154
+154
scripts/collectiondir-diff
+154
scripts/collectiondir-diff
···
1
+
#!/usr/bin/env -S PYTHONUNBUFFERED=1 uv run --script --quiet
2
+
# /// script
3
+
# requires-python = ">=3.12"
4
+
# dependencies = []
5
+
# ///
6
+
"""
7
+
compare listReposByCollection results between two relay endpoints.
8
+
9
+
fetches all DIDs for a collection from both endpoints and reports the
10
+
difference — which DIDs are unique to each side.
11
+
12
+
usage:
13
+
./scripts/collectiondir-diff --collection io.atcr.sailor.profile
14
+
./scripts/collectiondir-diff --collection app.bsky.feed.post --limit 10000
15
+
./scripts/collectiondir-diff --collection io.atcr.sailor.profile \
16
+
--a https://relay.waow.tech --b https://relay1.us-east.bsky.network
17
+
"""
18
+
19
+
import argparse
20
+
import json
21
+
import sys
22
+
import urllib.request
23
+
import urllib.error
24
+
25
+
26
+
def fetch_dids(base_url: str, collection: str, limit: int) -> set[str]:
27
+
"""paginate listReposByCollection and return all DIDs up to limit."""
28
+
endpoint = f"{base_url.rstrip('/')}/xrpc/com.atproto.sync.listReposByCollection"
29
+
dids: set[str] = set()
30
+
cursor = None
31
+
page_size = min(limit, 2000)
32
+
33
+
while len(dids) < limit:
34
+
params = f"collection={collection}&limit={page_size}"
35
+
if cursor:
36
+
params += f"&cursor={urllib.request.quote(cursor)}"
37
+
38
+
url = f"{endpoint}?{params}"
39
+
try:
40
+
req = urllib.request.Request(url)
41
+
with urllib.request.urlopen(req, timeout=30) as resp:
42
+
data = json.loads(resp.read())
43
+
except urllib.error.HTTPError as e:
44
+
print(f" error fetching {base_url}: {e.code} {e.reason}", file=sys.stderr)
45
+
break
46
+
except Exception as e:
47
+
print(f" error fetching {base_url}: {e}", file=sys.stderr)
48
+
break
49
+
50
+
repos = data.get("repos", [])
51
+
if not repos:
52
+
break
53
+
54
+
for repo in repos:
55
+
dids.add(repo["did"])
56
+
57
+
cursor = data.get("cursor")
58
+
if not cursor:
59
+
break
60
+
61
+
sys.stdout.write(f"\r {base_url}: {len(dids)} DIDs fetched...")
62
+
sys.stdout.flush()
63
+
64
+
return dids
65
+
66
+
67
+
def resolve_did(did: str) -> dict:
68
+
"""resolve a DID via plc.directory. returns handle + PDS host."""
69
+
try:
70
+
url = f"https://plc.directory/{did}"
71
+
with urllib.request.urlopen(url, timeout=10) as resp:
72
+
data = json.loads(resp.read())
73
+
handle = "?"
74
+
aka = data.get("alsoKnownAs", [])
75
+
if aka:
76
+
handle = aka[0].replace("at://", "")
77
+
pds = "?"
78
+
services = data.get("service", [])
79
+
if services:
80
+
pds = services[0].get("serviceEndpoint", "?")
81
+
return {"handle": handle, "pds": pds}
82
+
except Exception:
83
+
return {"handle": "?", "pds": "?"}
84
+
85
+
86
+
def main():
87
+
parser = argparse.ArgumentParser(
88
+
description="compare listReposByCollection between two relay endpoints"
89
+
)
90
+
parser.add_argument("--collection", required=True, help="collection NSID to compare")
91
+
parser.add_argument(
92
+
"--a",
93
+
default="https://relay.waow.tech",
94
+
help="first endpoint (default: https://relay.waow.tech)",
95
+
)
96
+
parser.add_argument(
97
+
"--b",
98
+
default="https://bsky.network",
99
+
help="second endpoint (default: https://bsky.network)",
100
+
)
101
+
parser.add_argument(
102
+
"--limit",
103
+
type=int,
104
+
default=100_000,
105
+
help="max DIDs to fetch per endpoint (default: 100000)",
106
+
)
107
+
parser.add_argument(
108
+
"--resolve",
109
+
action="store_true",
110
+
help="resolve DIDs to handles via plc.directory (slower)",
111
+
)
112
+
args = parser.parse_args()
113
+
114
+
print(f"collection: {args.collection}")
115
+
print(f" A: {args.a}")
116
+
print(f" B: {args.b}")
117
+
print(f" limit: {args.limit:,}")
118
+
print()
119
+
120
+
dids_a = fetch_dids(args.a, args.collection, args.limit)
121
+
print(f"\r A: {len(dids_a):,} DIDs" + " " * 40)
122
+
123
+
dids_b = fetch_dids(args.b, args.collection, args.limit)
124
+
print(f"\r B: {len(dids_b):,} DIDs" + " " * 40)
125
+
print()
126
+
127
+
only_a = dids_a - dids_b
128
+
only_b = dids_b - dids_a
129
+
common = dids_a & dids_b
130
+
131
+
print(f"common: {len(common):,}")
132
+
print(f"only in A ({args.a}): {len(only_a):,}")
133
+
print(f"only in B ({args.b}): {len(only_b):,}")
134
+
135
+
def print_dids(label: str, dids: set[str]):
136
+
if not dids:
137
+
return
138
+
print(f"\n{label}:")
139
+
for did in sorted(dids):
140
+
if args.resolve:
141
+
info = resolve_did(did)
142
+
print(f" {did} @{info['handle']} ({info['pds']})")
143
+
else:
144
+
print(f" {did}")
145
+
146
+
print_dids(f"only in A", only_a)
147
+
print_dids(f"only in B", only_b)
148
+
149
+
if not only_a and not only_b:
150
+
print("\nidentical.")
151
+
152
+
153
+
if __name__ == "__main__":
154
+
main()