[eks] Refactor script (#224)

Make the script more readable, mostly by:

- changing slightly the logic,
- using the Product and AutoConfig classes,
- removing the use of functions when unnecessary,
- a little bit of renaming and documentation.

Disabled web.archive.org has also been re-enabled because HTTP retry mechanism has been improved and should handle timeouts a lot better.

src/eks.py

@@ -1,45 +1,40 @@
-import re
 from bs4 import BeautifulSoup
 from common import http
 from common import dates
 from common import endoflife
 
-# Now that AWS no longer publishes docs on GitHub,
-# we use the Web Archive to still get the older versions
-# Keep older pages at top of the list
+"""Fetches EKS versions from AWS docs.
+Now that AWS no longer publishes docs on GitHub, we use the Web Archive to still get the older versions."""
+
 URLS = [
     # 1.19.eks.1
-    # Disabled, too much timed out.
-    # "https://web.archive.org/web/20221007150452/https://docs.aws.amazon.com/eks/latest/userguide/platform-versions.html",
+    "https://web.archive.org/web/20221007150452/https://docs.aws.amazon.com/eks/latest/userguide/platform-versions.html",
     # + 1.20
-    # Disabled, too much timed out.
-    # "https://web.archive.org/web/20230521061347/https://docs.aws.amazon.com/eks/latest/userguide/platform-versions.html",
+    "https://web.archive.org/web/20230521061347/https://docs.aws.amazon.com/eks/latest/userguide/platform-versions.html",
     # + latest
     "https://docs.aws.amazon.com/eks/latest/userguide/platform-versions.html",
 ]
 
+product = endoflife.Product("eks")
+print(f"::group::{product.name}")
 
-def parse_platforms_pages():
-    all_versions = {}
-    print("::group::eks")
-    for url in URLS:
-        response = http.fetch_url(url)
-        soup = BeautifulSoup(response.text, features="html5lib")
-        for tr in soup.select("#main-col-body")[0].findAll("tr"):
-            td = tr.find("td")
-            if td and re.match(endoflife.DEFAULT_VERSION_REGEX, td.text.strip()):
-                data = tr.findAll("td")
-                date = data[-1].text.strip()
-                if len(date) > 0:
-                    d = dates.parse_date(date).strftime("%Y-%m-%d")
-                    k8s_version = ".".join(data[0].text.strip().split(".")[:-1])
-                    eks_version = data[1].text.strip().replace(".", "-")
-                    version = f"{k8s_version}-{eks_version}"
-                    all_versions[version] = d
-                    print(f"{version}: {d}")
-    print("::endgroup::")
-    return all_versions
+for version_list in http.fetch_urls(URLS):
+    version_list_soup = BeautifulSoup(version_list.text, features="html5lib")
+    for tr in version_list_soup.select("#main-col-body")[0].findAll("tr"):
+        cells = tr.findAll("td")
+        if not cells:
+            continue
 
+        k8s_version = cells[0].text.strip()
+        eks_version = cells[1].text.strip()
+        date_str = cells[-1].text.strip()
 
-versions = parse_platforms_pages()
-endoflife.write_releases('eks', versions)
+        k8s_version_match = endoflife.DEFAULT_VERSION_PATTERN.match(k8s_version)
+        if k8s_version_match:
+            date = dates.parse_date(date_str)
+            # K8S patch version is not kept to match versions on https://github.com/aws/eks-distro/tags.
+            version = f"{k8s_version_match.group('major')}.{k8s_version_match.group('minor')}-{eks_version.replace('.', '-')}"
+            product.declare_version(version, date)
+
+product.write()
+print("::endgroup::")