shortener v2

Author: ramalho

links/shorten.py

@@ -30,8 +30,8 @@
 type TargetMap = dict[Url, ShortCode]
 
 class ShortPair(NamedTuple):
-    code = ShortCode
-    url = Url
+    code: ShortCode
+    url: Url
 
 def load_redirects() -> tuple[RedirMap, TargetMap]:
     redirects:RedirMap = {}

links/shortener.py

@@ -0,0 +1,34 @@
+from collections.abc import Iterable, Iterator
+
+
+def parse_htaccess(text: str) -> Iterator[tuple[str, str]]:
+    for line in text.splitlines():
+        fields = line.split()
+        if len(fields) < 3 or fields[0] != 'RedirectTemp':
+            continue
+        key = fields[1]
+        assert key[0] == '/'
+        key = key[1:]
+        assert len(key) > 0
+        yield (key, fields[2])
+
+
+def choose(a: str, b: str) -> str:
+    def key(k):
+        parts = k.split('-')
+        if len(parts) > 1:
+            parts = ['z' + p.rjust(8, '0') if p.isnumeric() else p for p in parts]
+        return (len(k), '-' in k, parts)
+    return min(a, b, key=key)
+
+
+def load_redirects(pairs: Iterable[tuple[str, str]]) -> tuple[dict, dict]:
+    redirects = {}
+    for key, url in pairs:
+        existing_url = redirects.get(key)
+        if existing_url:
+            if len(existing_url) < len(url):
+                continue
+            elif len(existing_url) == len(url):
+                url = choose(existing_url, url)
+            

links/test_shortener.py

@@ -0,0 +1,49 @@
+from pytest import mark
+
+from shortener import parse_htaccess, choose
+
+
+HTACCESS_1 = """
+ErrorDocument 404 /404.html
+
+# main resources
+RedirectTemp /book	https://www.oreilly.com/.../9781492056348/
+RedirectTemp /home	https://www.fluentpython.com/  # extra content site
+
+# duplicate targets
+RedirectTemp /1-20	https://www.fluentpython.com/
+RedirectTemp /ora	https://www.oreilly.com/.../9781492056348/
+
+"""
+
+PARSED_HTACCESS_1 = [
+        ('book', 'https://www.oreilly.com/.../9781492056348/'),
+        ('home', 'https://www.fluentpython.com/'),
+        ('1-20', 'https://www.fluentpython.com/'),
+        ('ora', 'https://www.oreilly.com/.../9781492056348/'),
+    ]
+
+def test_parse_htaccess():
+    res = list(parse_htaccess(HTACCESS_1))
+    assert res == PARSED_HTACCESS_1
+
+@mark.parametrize('a,b,expected', [
+    ('a', 'b', 'a'),
+    ('b', 'a', 'a'),
+    ('aa', 'a', 'a'),
+    ('a-a', 'aaa', 'aaa'),
+    ('2-10', '10-2', '2-10'),
+    ('p-1', '1-1', 'p-1'),
+])
+def test_choose(a, b, expected):
+    res = choose(a, b)
+    assert res == expected
+
+
+# def test_load_redirects():
+#     expected = {
+#         'home': 'https://www.fluentpython.com/',
+#         'ora': 'https://www.oreilly.com/.../9781492056348/'
+#     }
+#     redirects, _ = load_redirects(PARSED_HTACCESS_1)
+#     assert redirects == expected

ruff.toml

@@ -0,0 +1,4 @@
+line-length = 100
+[format]
+# Use single quotes for strings, like Python's repr()
+quote-style = "single"