pypi: support extracting project names from PURL references

For example "pkg:pypi/django@1.11.1" -> django

This means we can get rid of the non-purl "pypi" mapping in PKGBUILDS

app/fetch/pypi.py

@@ -5,6 +5,7 @@ import datetime
 import gzip
 import json
 import re
+from urllib.parse import unquote
 
 from ..appconfig import PYPI_URLS, REQUEST_TIMEOUT
 from ..appstate import ExtId, ExtInfo, state
@@ -18,6 +19,37 @@ def normalize(name: str) -> str:
     return re.sub(r"[-_.]+", "-", name).lower()
 
 
+def extract_pypi_project_from_purl(purl: str) -> str | None:
+    """Extract the project name from a PyPI PURL.
+    If not a proper PyPI PURL, return None.
+    """
+
+    if not purl.startswith("pkg:pypi/"):
+        return None
+    path_and_rest = purl[len("pkg:pypi/"):]
+    path_part = path_and_rest.split("@", 1)[0].split("?", 1)[0].split("#", 1)[0]
+    parts = path_part.rsplit("/", 1)
+    if not parts or not parts[-1]:
+        return None
+    return unquote(parts[-1])
+
+
+def extract_pypi_project_from_references(references: dict[str, list[str | None]]) -> str | None:
+    if "pypi" in references:
+        for entry in references["pypi"]:
+            if entry is not None:
+                return entry
+
+    for purl in references.get("purl", []):
+        if purl is None:
+            continue
+        project = extract_pypi_project_from_purl(purl)
+        if project is not None:
+            return project
+
+    return None
+
+
 async def update_pypi_versions(pkgextra: PkgExtra) -> None:
     urls = PYPI_URLS
     if not await check_needs_update(urls):
@@ -32,9 +64,9 @@ async def update_pypi_versions(pkgextra: PkgExtra) -> None:
 
     pypi_versions = {}
     for entry in pkgextra.packages.values():
-        if "pypi" not in entry.references:
+        pypi_name = extract_pypi_project_from_references(entry.references)
+        if pypi_name is None:
             continue
-        pypi_name = entry.references["pypi"][0]
         assert isinstance(pypi_name, str)
         normalized_name = normalize(pypi_name)
         if normalized_name in projects:

tests/test_main.py

@@ -8,6 +8,7 @@ import pytest
 from app import app
 from app.appstate import SrcInfoPackage, parse_packager
 from app.fetch.cygwin import parse_cygwin_versions
+from app.fetch.pypi import extract_pypi_project_from_purl
 from app.utils import split_optdepends, strip_vcs, vercmp
 from app.pkgextra import extra_to_pkgextra_entry
 from fastapi.testclient import TestClient
@@ -218,3 +219,10 @@ def test_extra_to_pkgextra_entry():
     assert extra_to_pkgextra_entry(
         {"changelog_url": "foo"}
     ).changelog_url == "foo"
+
+
+def test_extract_pypi_project_from_purl():
+    assert extract_pypi_project_from_purl("pkg:pypi/foo") == "foo"
+    assert extract_pypi_project_from_purl("pkg:pypi/django@1.11.1") == "django"
+    assert extract_pypi_project_from_purl("pkg:pypi/django?filename=Django-1.11.1.tar.gz") == "django"
+    assert extract_pypi_project_from_purl("pkg:cargo/rand@0.7.2") is None