web: basic Accept-Language HTTP negotiation

2 files changed, 38 insertions, 1 deletions

diff --git a/fatcat_scholar/hacks.py b/fatcat_scholar/hacks.py
index 710a25f..fc16564 100644
--- a/fatcat_scholar/hacks.py
+++ b/fatcat_scholar/hacks.py
@@ -51,3 +51,29 @@ class Jinja2Templates:
             media_type=media_type,
             background=background,
         )
+
+
+def parse_accept_lang(header: str, options: typing.List[str]) -> typing.Optional[str]:
+    """
+    Crude HTTP Accept-Language content negotiation.
+    Assumes that languages are specified in order of priority, etc.
+    """
+    if not header:
+        return None
+    chunks = [v.split(";")[0].split("-")[0] for v in header.split(",")]
+    for c in chunks:
+        if len(c) == 2 and c in options:
+            return c
+    return None
+
+
+def test_parse_accept_lang() -> None:
+    assert parse_accept_lang("", []) == None
+    assert parse_accept_lang("en,de", []) == None
+    assert parse_accept_lang("en,de", ["en"]) == "en"
+    assert parse_accept_lang("en-GB,de", ["en"]) == "en"
+    assert parse_accept_lang("en,de", ["de"]) == "de"
+    assert (
+        parse_accept_lang("en-ca,en;q=0.8,en-us;q=0.6,de-de;q=0.4,de;q=0.2", ["de"])
+        == "de"
+    )
diff --git a/fatcat_scholar/web.py b/fatcat_scholar/web.py
index 73f79b5..f266e08 100644
--- a/fatcat_scholar/web.py
+++ b/fatcat_scholar/web.py
@@ -15,7 +15,7 @@ import sentry_sdk
 from sentry_sdk.integrations.asgi import SentryAsgiMiddleware
 
 from fatcat_scholar.config import settings, GIT_REVISION
-from fatcat_scholar.hacks import Jinja2Templates
+from fatcat_scholar.hacks import Jinja2Templates, parse_accept_lang
 from fatcat_scholar.search import do_fulltext_search, FulltextQuery, FulltextHits
 
 
@@ -49,6 +49,17 @@ class LangPrefix:
     def __init__(self, request: Request):
         self.prefix: str = ""
         self.code: str = settings.I18N_LANG_DEFAULT
+        # first try to parse a language code from header
+        try:
+            accept_code = parse_accept_lang(
+                request.headers.get("accept-language", ""), I18N_LANG_TRANSLATIONS,
+            )
+            if accept_code:
+                self.code = accept_code
+        except Exception:
+            pass
+
+        # then override this with any language code in URL
         for lang_option in I18N_LANG_OPTIONS:
             if request.url.path.startswith(f"/{lang_option}/"):
                 self.prefix = f"/{lang_option}"