3 files changed, 59 insertions, 13 deletions

diff --git a/repos/__init__.py b/repos/__init__.py
index 0f71dd1..1dea3a1 100644
--- a/repos/__init__.py
+++ b/repos/__init__.py
@@ -1,4 +1,4 @@
-from typing import Mapping, List, Any
+from typing import Mapping, List
 
 from . import alpine_linux
 from .base import Repository
diff --git a/repos/alpine_linux.py b/repos/alpine_linux.py
index 64f159a..7c646cc 100644
--- a/repos/alpine_linux.py
+++ b/repos/alpine_linux.py
@@ -1,9 +1,10 @@
 from io import TextIOWrapper
 from pathlib import Path
+import re
 import tarfile
 from typing import Mapping, TextIO
 
-from .base import Repository
+from .base import Repository, Version
 
 __all__ = [
     'stable_main_x86_64',
@@ -13,7 +14,9 @@ __all__ = [
     'edge_testing_x86_64',
 ]
 
-def parse_apkindex(apkindex: TextIO) -> Mapping[str, str]:
+PACKAGE_REVISION_INFO = re.compile(r'-r\d+$')
+
+def parse_apkindex(apkindex: TextIO) -> Mapping[str, Version]:
     result = dict()
     current_package = None
     current_version = None
@@ -37,14 +40,16 @@ def parse_apkindex(apkindex: TextIO) -> Mapping[str, str]:
         elif line_type == 'P':
             current_package = line_data
         elif line_type == 'V':
-            current_version = line_data
+            version = line_data
+            clean_version = PACKAGE_REVISION_INFO.sub('', version)
+            current_version = Version(version, clean_version)
         elif line_type in ignore_lines:
             pass
         else:
             raise ValueError('unknown line type: ' + line_type + ' in line ' + repr(line))
     return result
 
-def parse_cached(cached: Path) -> Mapping[str, str]:
+def parse_cached(cached: Path) -> Mapping[str, Version]:
     apkindex = tarfile.open(cached)
     for archive_member in apkindex.getmembers():
         if archive_member.name == 'APKINDEX':
@@ -52,7 +57,7 @@ def parse_cached(cached: Path) -> Mapping[str, str]:
             apkindex_file = TextIOWrapper(apkindex_file)
             return parse_apkindex(apkindex_file)
 
-def build_repo(name: str, url_path: str):
+def build_repo(name: str, url_path: str) -> Repository:
     url = f'http://dl-cdn.alpinelinux.org/alpine/{url_path}/APKINDEX.tar.gz'
     return Repository(
         family='Alpine Linux',
diff --git a/repos/base.py b/repos/base.py
index 66ecf2d..220a30f 100644
--- a/repos/base.py
+++ b/repos/base.py
@@ -1,26 +1,66 @@
-from dataclasses import dataclass
+from dataclasses import dataclass, asdict as dataclass_asdict
 import datetime
+from functools import total_ordering
 import gzip
 import json
 import os
 from pathlib import Path
 import re
-from typing import Callable, Mapping
+from typing import Any, Callable, Mapping
 
 import requests
+import semver
+
+__all__ = [
+    'Repository',
+    'Version',
+]
 
 HTTP_DATE = '%a, %d %b %Y %H:%M:%S GMT'
-SLUGIFY = re.compile('\W+')
+SLUGIFY = re.compile(r'\W+')
 
 def slug(text: str) -> str:
     return SLUGIFY.sub('-', text.lower()).strip('-')
 
+@total_ordering
+@dataclass()
+class Version:
+    original: str
+    clean: str
+
+    def __str__(self) -> str:
+        return self.original
+
+    def __lt__(self, other: Any):
+        if not isinstance(other, Version):
+            return NotImplemented
+        if semver.VersionInfo.isvalid(self.clean) and semver.VersionInfo.isvalid(other.clean):
+            return semver.compare(self.clean, other.clean) < 0
+        return self.original < other.original
+
+class JSONEncoder(json.JSONEncoder):
+    def default(self, o: Any) -> Any:
+        if isinstance(o, Version):
+            return dataclass_asdict(o)
+        return super().default(o)
+
+class JSONDecoder(json.JSONDecoder):
+    @staticmethod
+    def object_hook(o: dict) -> Any:
+        if o.keys() == {'original', 'clean'}:
+            return Version(**o)
+        return o
+
+    def __init__(self):
+        super().__init__(object_hook=self.object_hook)
+
+
 @dataclass()
 class Repository:
     family: str
     repo: str
     index_url: str
-    parse: Callable[[Path], Mapping[str, str]]
+    parse: Callable[[Path], Mapping[str, Version]]
 
     def _full_name(self):
         return f'{self.family} {self.repo}'
@@ -31,7 +71,7 @@ class Repository:
     def _cache_file(self, name: str) -> Path:
         return self._cache_dir() / name
 
-    def get_versions(self) -> Mapping[str, str]:
+    def get_versions(self) -> Mapping[str, Version]:
         self._cache_dir().mkdir(parents=True, exist_ok=True)
         downloaded_file = self._cache_file('downloaded')
         if downloaded_file.exists():
@@ -57,10 +97,11 @@ class Repository:
                 set_mtime = datetime.datetime.strptime(set_mtime, HTTP_DATE)
                 os.utime(downloaded_file, (datetime.datetime.now().timestamp(), set_mtime.timestamp()))
 
+        if response.status_code != requests.codes.not_modified or not parsed_file.exists():
             parsed_data = self.parse(downloaded_file)
             with gzip.open(parsed_file, 'wt') as f:
-                json.dump(parsed_data, f)
+                json.dump(parsed_data, f, cls=JSONEncoder)
             return parsed_data
         else:
             with gzip.open(parsed_file, 'rt') as f:
-                return json.load(f)
+                return json.load(f, cls=JSONDecoder)