Merge pull request #2 from wvanhed/develop

v0.5.5
wvanhed · Mar 5, 2024 · b76e91b · b76e91b
2 parents 82a5264 + 93b4ea2
commit b76e91b
Show file tree

Hide file tree

Showing 5 changed files with 33 additions and 4 deletions.
diff --git a/README.md b/README.md
@@ -58,12 +58,14 @@ Tenslotte, via de commandline kan je de module ook als volgt aanroepen:
 ## Opmerkingen
 
 - **Authenticatie**. Inloggen bij de bibliotheek.be website gebeurt standaard
-  via een webformulier. Het is ook mogelijk om de snellere `oauth` manier te
-  gebruiken; dit is nog experimenteel.
+  via een webformulier. Het is ook mogelijk om de `oauth` manier te gebruiken;
+  maar dit is nog experimenteel.
 
         mb = MijnBibliotheek(username, password, login_by="oauth")
         accounts = mb.get_accounts()
 
+  ! Opmerking: De oauth flow is sinds maart 2024 broken, en vereist nog een aanpassing.
+
 - **Foutafhandeling**. Afhankelijk van de toepassing, kan het aangeraden zijn om
   foutafhandeling te voorzien. Het bestand `errors.py` bevat de lijst van
   Mijnbib-specifieke exceptions. De docstrings van de publieke methods bevatten

diff --git a/changelog.md b/changelog.md
@@ -2,6 +2,12 @@
 
 new: new feature /  impr: improvement /  fix: bug fix
 
+## v0.5.5 - 2024-03-05
+
+- fix: broken login (form) because of change at site. 
+  Note: alternative oauth login still broken.
+- impr: raise TemporarySiteError at oauth login when 5xx (part 3)
+
 ## v0.5.4 - 2024-02-21
 
 - impr: improve extensibility of oauth login handler

diff --git a/pyproject.toml b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mijnbib"
-version = "0.5.4"
+version = "0.5.5"
 description = "Python API voor de website mijn.bibliotheek.be"
 readme = "README.md"
 authors = [{ name = "Ward Van Heddeghem", email = "[email protected]" }]

diff --git a/src/mijnbib/login_handlers.py b/src/mijnbib/login_handlers.py
@@ -45,6 +45,12 @@ def _log_in(self):
             _log.debug("Opening login page ... ")
             response = self._br.open(self._url, timeout=TIMEOUT)
             html_string_start_page = response.read().decode("utf-8")  # type:ignore
+            # Workaround for mechanize.BrowserStateError: not viewing HTML
+            # because suddenly (March 2024) Content-Type header is "application/octet-stream;charset=UTF-8"
+            # which is not recognized as html by mechanize
+            # Alternative is to configure the browser instance with
+            #     self._br.set_header("Accept", "text/html")
+            self._br._factory.is_html = True
             self._br.select_form(nr=0)
             self._br["email"] = self._username
             self._br["password"] = self._pwd
@@ -100,6 +106,7 @@ def _log_in(self):
         # GET https://gent.bibliotheek.be/mijn-bibliotheek/aanmelden
         # example response:
         # header Location: https://mijn.bibliotheek.be/openbibid/rest/auth/authorize?hint=login&oauth_callback=https://gent.bibliotheek.be/my-library/login/callback&oauth_token=5abee3c0f5c04beead64d8e625ead0e7&uilang=nl
+        _log.debug("----")
         response = self._s.get(self._url, allow_redirects=False)
         _log.debug(f"login (1) status code       : {response.status_code}")
         _log.debug(f"login (1) headers           : {response.headers}")
@@ -125,9 +132,10 @@ def _log_in(self):
             )
         if "/mijn-bibliotheek/overzicht" in oauth_location_url:
             _log.info("Already authenticated. No need to log in again.")
-            return
+            return response  # better for extensibility (i.e. sOlid)
 
         # (2) Authorize based on Location url (get session id)
+        _log.debug("----")
         response = self._s.get(oauth_location_url, allow_redirects=False)
         _log.debug(f"login (2) status code       : {response.status_code}")
         _log.debug(f"login (2) headers           : {response.headers}")
@@ -149,6 +157,7 @@ def _log_in(self):
             "email": self._username,
             "password": self._pwd,
         }
+        _log.debug("----")
         response = self._s.post(url, data=data, allow_redirects=False)
         _log.debug(f"login (3) status code       : {response.status_code}")
         _log.debug(f"login (3) headers           : {response.headers}")
@@ -164,13 +173,18 @@ def _log_in(self):
         _log.debug(f"login (3) hint              : {hint}")
         if response.status_code == 200:
             raise AuthenticationError("Login not accepted. Correct credentials?")
+        if response.status_code >= 500:  # we've observed 500
+            raise TemporarySiteError(
+                f"Expected status code 303 during log in. Got '{response.status_code}'"
+            )
         if response.status_code != 303:
             raise IncompatibleSourceError(
                 f"Expected status code 303 during log in. Got '{response.status_code}'",
                 response.text,
             )
 
         # (4) Call login callback based on Location url
+        _log.debug("----")
         response = self._s.get(login_location_url, allow_redirects=False)
         _log.debug(f"login (4) status code       : {response.status_code}")
         _log.debug(f"login (4) headers           : {response.headers}")

diff --git a/tests/test_mijnbibliotheek.py b/tests/test_mijnbibliotheek.py
@@ -19,9 +19,16 @@ def creds_config(scope="module"):
     yield dict(**config.defaults())
 
 
+class X(str):
+    pass
+
+
 class FakeMechanizeBrowser:
     def __init__(self, form_response: str) -> None:
         self._form_response = form_response.encode("utf8")
+        # trick for nested prop, from https://stackoverflow.com/a/35190607/50899
+        self._factory = X("_factory")
+        self._factory.is_html = None  # can be whatever
 
     def __setitem__(self, key, value):
         pass