Files
rgl-photo-backup/test_rgl_photo_backup.py
T
bobandClaude Fable 5.1 f05837bd5e Add a retrying HTTP helper shared by every remote call
One place decides what is worth another attempt: 429, 5xx and network
errors, with jittered exponential backoff and Retry-After honoured. Other
4xx responses fail at once. Downloads stream to disk under the same policy,
refuse a body over the size cap, and only flickr calls and photo downloads
are paced; the api key is redacted from debug output.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-09-10 12:07:44 +01:00

306 lines
12 KiB
Python

import contextlib
import importlib.machinery
import importlib.util
import io
import json
import os
import re
import sys
import tempfile
import unittest
import urllib.error
import urllib.parse
_here = os.path.dirname(os.path.abspath(__file__))
_path = os.path.join(_here, "rgl-photo-backup")
_spec = importlib.util.spec_from_file_location(
"rgl_photo_backup", _path,
loader=importlib.machinery.SourceFileLoader("rgl_photo_backup", _path))
rb = importlib.util.module_from_spec(_spec)
_spec.loader.exec_module(rb)
class FakeResponse:
"""What urlopen would return: a body, a status and headers."""
def __init__(self, body=b"", status=200, headers=None):
self.body = body if isinstance(body, bytes) else json.dumps(body).encode()
self.status = status
self.headers = headers or {}
self._pos = 0
def read(self, size=-1):
if size is None or size < 0:
data, self._pos = self.body[self._pos:], len(self.body)
else:
data = self.body[self._pos:self._pos + size]
self._pos += len(data)
return data
def close(self):
pass
def __enter__(self):
return self
def __exit__(self, *exc):
pass
def http_error(code, body=b"", headers=None):
return urllib.error.HTTPError("http://test", code, "reason %d" % code,
headers or {}, io.BytesIO(body))
class FakeOpen:
"""Replaces rb.OPEN. Serves scripted responses in order and records
every request so a test can assert on method, url, headers and body."""
def __init__(self, *responses):
self.responses = list(responses)
self.requests = []
def __call__(self, request, timeout=None):
self.requests.append(request)
if not self.responses:
raise AssertionError("unexpected request: %s %s"
% (request.get_method(), request.full_url))
answer = self.responses.pop(0)
if isinstance(answer, Exception):
raise answer
return answer
class TestCase(unittest.TestCase):
"""Silences output and restores the module globals tests move around."""
GLOBALS = ("OPEN", "SLEEP", "ATTEMPTS", "BACKOFF_BASE", "PACE", "QUIET",
"DEBUG", "HTTP_TIMEOUT", "USER_AGENT", "STATS", "DOWNLOAD_CAP")
def setUp(self):
self._output = contextlib.ExitStack()
if not os.environ.get("RGL_PHOTO_BACKUP_TEST_VERBOSE"):
self.stdout = self._output.enter_context(
contextlib.redirect_stdout(io.StringIO()))
self.stderr = self._output.enter_context(
contextlib.redirect_stderr(io.StringIO()))
self.addCleanup(self._output.close)
self._saved = {k: getattr(rb, k) for k in self.GLOBALS}
rb.SLEEP = lambda seconds: None
rb.PACE = 0.0
rb.STATS = dict.fromkeys(rb.STATS, 0)
self.tmp = tempfile.TemporaryDirectory()
self.addCleanup(self.tmp.cleanup)
self.dir = self.tmp.name
def tearDown(self):
for k, v in self._saved.items():
setattr(rb, k, v)
def write(self, name, content):
path = os.path.join(self.dir, name)
mode = "wb" if isinstance(content, bytes) else "w"
with open(path, mode) as fh:
fh.write(content)
return path
class Output(TestCase):
def test_say_prints_unless_quiet(self):
rb.QUIET = False
rb.say("hello")
self.assertEqual(self.stdout.getvalue(), "hello\n")
rb.QUIET = True
rb.say("again")
self.assertEqual(self.stdout.getvalue(), "hello\n")
def test_warn_always_reaches_stderr(self):
rb.QUIET = True
rb.warn("careful")
self.assertEqual(self.stderr.getvalue(), "careful\n")
def test_debug_only_when_enabled(self):
rb.DEBUG = False
rb.debug("x")
self.assertEqual(self.stderr.getvalue(), "")
rb.DEBUG = True
rb.debug("y")
self.assertEqual(self.stderr.getvalue(), "debug: y\n")
def test_errors_share_a_base(self):
self.assertTrue(issubclass(rb.PhotoGone, rb.BackupError))
e = rb.HTTPFailure("ctx: HTTP 404", code=404)
self.assertEqual(e.code, 404)
self.assertIsInstance(e, rb.BackupError)
class Config(TestCase):
def test_readyaml_reads_flat_pairs_and_ignores_comments(self):
path = self.write("a.yml", "# note\nkey: 'abc'\nurl: https://immich.example/api \n")
self.assertEqual(rb.readyaml(path), {"key": "abc", "url": "https://immich.example/api"})
def test_readyaml_missing_file_is_empty(self):
self.assertEqual(rb.readyaml(os.path.join(self.dir, "none.yml")), {})
def test_readyaml_unreadable_path_names_it(self):
with self.assertRaises(rb.BackupError) as cm:
rb.readyaml(self.dir) # a directory, not a file
self.assertIn(self.dir, str(cm.exception))
def test_loadflickr_needs_key(self):
path = self.write("f.yml", "secret: s\n")
with self.assertRaises(rb.BackupError):
rb.loadflickr(path)
path = self.write("f2.yml", "key: k\nsecret: s\n")
self.assertEqual(rb.loadflickr(path), "k")
def test_loadimmich_needs_url_and_key_and_strips_slash(self):
path = self.write("i.yml", "url: https://immich.example/api/\nkey: abc\n")
self.assertEqual(rb.loadimmich(path), ("https://immich.example/api", "abc"))
with self.assertRaises(rb.BackupError):
rb.loadimmich(self.write("i2.yml", "url: x\n"))
class Settings(TestCase):
def test_flag_beats_env_beats_file_beats_default(self):
path = self.write("s.yml", "directory: /from/file\npace: 3\nquiet: true\n")
args = rb.parseargs(["--settings", path])
conf = rb.settings(args, environ={})
self.assertEqual(conf["directory"], "/from/file")
self.assertEqual(conf["pace"], 3.0)
self.assertTrue(conf["quiet"])
conf = rb.settings(args, environ={rb.ENV + "PACE": "7"})
self.assertEqual(conf["pace"], 7.0)
args = rb.parseargs(["--settings", path, "--pace", "9", "/cli"])
conf = rb.settings(args, environ={rb.ENV + "PACE": "7"})
self.assertEqual(conf["pace"], 9.0)
self.assertEqual(conf["directory"], "/cli")
def test_defaults(self):
args = rb.parseargs(["--settings", os.path.join(self.dir, "none"), "/d"])
conf = rb.settings(args, environ={})
self.assertEqual(conf["index_url"], rb.INDEX_URL)
self.assertTrue(conf["fetch"] and conf["sync"])
self.assertFalse(conf["dry_run"])
self.assertIsNone(conf["limit"])
self.assertEqual(conf["attempts"], 6)
self.assertEqual(conf["flickr_auth"], os.path.expanduser(rb.FLICKR_CONFIG))
def test_defaults_are_literals_not_moving_constants(self):
rb.PACE, rb.ATTEMPTS, rb.USER_AGENT = 99.0, 99, "mangled"
args = rb.parseargs(["--settings", os.path.join(self.dir, "none"), "/d"])
conf = rb.settings(args, environ={})
self.assertEqual(conf["pace"], rb.DEFAULTS["pace"])
self.assertEqual(conf["attempts"], rb.DEFAULTS["attempts"])
self.assertEqual(conf["user_agent"], rb.DEFAULTS["user_agent"])
self.assertEqual((rb.DEFAULTS["pace"], rb.DEFAULTS["attempts"]), (1.0, 6))
def test_debug_is_a_setting(self):
path = self.write("s.yml", "directory: /d\ndebug: true\n")
self.assertEqual(self.stderr.getvalue(), "") # no "unknown setting"
args = rb.parseargs(["--settings", path])
self.assertTrue(rb.settings(args, environ={})["debug"])
args = rb.parseargs(["--settings", os.path.join(self.dir, "none"), "/d"])
self.assertFalse(rb.settings(args, environ={})["debug"])
self.assertTrue(rb.settings(args, environ={rb.ENV + "DEBUG": "yes"})["debug"])
args = rb.parseargs(["--settings", os.path.join(self.dir, "none"), "--debug", "/d"])
self.assertTrue(rb.settings(args, environ={})["debug"])
def test_a_bad_environment_value_is_a_configuration_error(self):
args = rb.parseargs(["--settings", os.path.join(self.dir, "none"), "/d"])
with self.assertRaises(rb.BackupError) as cm:
rb.settings(args, environ={rb.ENV + "ATTEMPTS": "lots"})
self.assertIn("ATTEMPTS", str(cm.exception))
def test_phase_flags(self):
base = ["--settings", os.path.join(self.dir, "none"), "/d"]
self.assertFalse(rb.settings(rb.parseargs(base + ["--sync-only"]), environ={})["fetch"])
self.assertFalse(rb.settings(rb.parseargs(base + ["--fetch-only"]), environ={})["sync"])
with self.assertRaises(rb.BackupError):
rb.settings(rb.parseargs(base + ["--fetch-only", "--sync-only"]), environ={})
def test_missing_directory_is_an_error(self):
args = rb.parseargs(["--settings", os.path.join(self.dir, "none")])
with self.assertRaises(rb.BackupError):
rb.settings(args, environ={})
def test_unknown_key_warns_and_wrong_type_fails(self):
path = self.write("s.yml", "directory: /d\nbogus: 1\n")
rb.loadsettings(path)
self.assertIn("bogus", self.stderr.getvalue())
path = self.write("s2.yml", "directory: /d\nlimit: many\n")
with self.assertRaises(rb.BackupError) as cm:
rb.loadsettings(path)
self.assertIn("limit", str(cm.exception))
def test_tilde_expanded_in_paths(self):
path = self.write("s.yml", "directory: ~/photos\n")
self.assertEqual(rb.loadsettings(path)["directory"], os.path.expanduser("~/photos"))
class Retrying(TestCase):
def test_success_returns_body(self):
rb.OPEN = FakeOpen(FakeResponse(b"ok"))
self.assertEqual(rb.fetch("http://t/x", "ctx"), b"ok")
self.assertEqual(rb.OPEN.requests[0].get_header("User-agent"), rb.USER_AGENT)
def test_404_is_not_retried(self):
rb.OPEN = FakeOpen(http_error(404, b"gone"))
with self.assertRaises(rb.HTTPFailure) as cm:
rb.fetch("http://t/x", "ctx")
self.assertEqual(cm.exception.code, 404)
self.assertIn("gone", str(cm.exception))
self.assertEqual(len(rb.OPEN.requests), 1)
def test_5xx_and_429_retry_then_succeed(self):
rb.OPEN = FakeOpen(http_error(503), http_error(429), FakeResponse(b"ok"))
self.assertEqual(rb.fetch("http://t/x", "ctx"), b"ok")
self.assertEqual(len(rb.OPEN.requests), 3)
self.assertEqual(rb.STATS["retries"], 2)
def test_connection_errors_retry(self):
rb.OPEN = FakeOpen(OSError("reset"), FakeResponse(b"ok"))
self.assertEqual(rb.fetch("http://t/x", "ctx"), b"ok")
def test_gives_up_after_attempts(self):
rb.ATTEMPTS = 3
rb.OPEN = FakeOpen(http_error(500), http_error(500), http_error(500))
with self.assertRaises(rb.HTTPFailure) as cm:
rb.fetch("http://t/x", "ctx")
self.assertIn("3 attempts", str(cm.exception))
def test_retry_after_header_is_honoured(self):
waits = []
rb.SLEEP = waits.append
rb.OPEN = FakeOpen(http_error(429, headers={"Retry-After": "12"}), FakeResponse(b"ok"))
rb.fetch("http://t/x", "ctx")
self.assertIn(12.0, waits)
def test_backoff_grows_and_caps(self):
rb.BACKOFF_BASE = 5
self.assertLessEqual(rb.backoff(1), 7.5)
self.assertGreaterEqual(rb.backoff(1), 2.5)
self.assertLessEqual(rb.backoff(20), rb.BACKOFF_CAP * 1.5)
def test_post_sends_body_and_headers(self):
rb.OPEN = FakeOpen(FakeResponse(b"{}"))
rb.fetch("http://t/x", "ctx", headers={"x-api-key": "k"}, data=b"body", method="POST")
req = rb.OPEN.requests[0]
self.assertEqual(req.get_method(), "POST")
self.assertEqual(req.data, b"body")
self.assertEqual(req.get_header("X-api-key"), "k")
def test_download_streams_to_path(self):
rb.OPEN = FakeOpen(FakeResponse(b"x" * 1000, headers={"Content-Type": "image/jpeg"}))
path = os.path.join(self.dir, "p.part")
self.assertEqual(rb.download("http://t/p.jpg", "ctx", path), ("image/jpeg", 1000))
self.assertEqual(os.path.getsize(path), 1000)
def test_download_refuses_an_oversized_body(self):
rb.DOWNLOAD_CAP = 10
rb.OPEN = FakeOpen(FakeResponse(b"x" * 100))
path = os.path.join(self.dir, "p.part")
with self.assertRaises(rb.BackupError) as cm:
rb.download("http://t/p.jpg", "ctx", path)
self.assertIn("10", str(cm.exception))
self.assertEqual(len(rb.OPEN.requests), 1) # not a retryable failure