Align custom scripts with generic scripts (#445)

Align custom scripts with generic scripts, making them configurable. This has a few advantages: - script code is more unified, - no more hard-coded method names in scripts, which is less error prone and make it easier to rename scripts, - no more hard coded product names in scripts, which is less error prone and make it easier to rename products, - less hard-coded URLs and regexes in scripts, which makes auto-configuration more expressive / updatable, Also added method `endoflife.list_configs_from_argv()` so that it is easier to manipulate scripts arguments.
2025-06-07 12:41:59 +02:00
parent 60a62e4696
commit f404274310
63 changed files with 704 additions and 830 deletions
--- a/src/aws-lambda.py
+++ b/src/aws-lambda.py
@@ -1,47 +1,48 @@
 import logging

 from bs4 import BeautifulSoup
-from common import dates, http, releasedata
+from common import dates, endoflife, http, releasedata

 """Fetches AWS lambda runtimes with their support / EOL dates from https://docs.aws.amazon.com."""

-with releasedata.ProductData("aws-lambda") as product_data:
-    response = http.fetch_url("https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtimes.html")
-    soup = BeautifulSoup(response.text, features="html5lib")
+for config in endoflife.list_configs_from_argv():
+    with releasedata.ProductData(config.product) as product_data:
+        response = http.fetch_url(config.url)
+        soup = BeautifulSoup(response.text, features="html5lib")

-    for i, table in enumerate(soup.find_all("table")):
-        headers = [th.get_text().strip().lower() for th in table.find("thead").find_all("tr")[0].find_all("th")]
-        if "identifier" not in headers or "deprecation date" not in headers or "block function update" not in headers:
-            logging.info(f"table with header '{headers}' does not contain all the expected headers")
-            continue
+        for i, table in enumerate(soup.find_all("table")):
+            headers = [th.get_text().strip().lower() for th in table.find("thead").find_all("tr")[0].find_all("th")]
+            if "identifier" not in headers or "deprecation date" not in headers or "block function update" not in headers:
+                logging.info(f"table with header '{headers}' does not contain all the expected headers")
+                continue

-        is_supported_table = i == 0  # first table is for supported runtimes, second for deprecated ones
-        identifier_index = headers.index("identifier")
-        deprecation_date_index = headers.index("deprecation date")
-        block_function_update_index = headers.index("block function update")
+            is_supported_table = i == 0  # first table is for supported runtimes, second for deprecated ones
+            identifier_index = headers.index("identifier")
+            deprecation_date_index = headers.index("deprecation date")
+            block_function_update_index = headers.index("block function update")

-        for row in table.find("tbody").find_all("tr"):
-            cells = row.find_all("td")
-            identifier = cells[identifier_index].get_text().strip()
+            for row in table.find("tbody").find_all("tr"):
+                cells = row.find_all("td")
+                identifier = cells[identifier_index].get_text().strip()

-            deprecation_date_str = cells[deprecation_date_index].get_text().strip()
-            try:
-                deprecation_date = dates.parse_date(deprecation_date_str)
-            except ValueError:
-                deprecation_date = None
+                deprecation_date_str = cells[deprecation_date_index].get_text().strip()
+                try:
+                    deprecation_date = dates.parse_date(deprecation_date_str)
+                except ValueError:
+                    deprecation_date = None

-            if identifier == "nodejs4.3-edge":
-                # there is a mistake in the data: block function update date cannot be before the deprecation date
-                block_function_update_str = "2020-04-30"
-            else:
-                block_function_update_str = cells[block_function_update_index].get_text().strip()
-            try:
-                block_function_update = dates.parse_date(block_function_update_str)
-            except ValueError:
-                block_function_update = None
+                if identifier == "nodejs4.3-edge":
+                    # there is a mistake in the data: block function update date cannot be before the deprecation date
+                    block_function_update_str = "2020-04-30"
+                else:
+                    block_function_update_str = cells[block_function_update_index].get_text().strip()
+                try:
+                    block_function_update = dates.parse_date(block_function_update_str)
+                except ValueError:
+                    block_function_update = None

-            release = product_data.get_release(identifier)
-            # if no date is available, use False for supported runtimes and True for deprecated ones
-            release.set_eoas(deprecation_date if deprecation_date else not is_supported_table)
-            # if no date is available, use False for supported runtimes and True for deprecated ones
-            release.set_eol(block_function_update if block_function_update else not is_supported_table)
+                release = product_data.get_release(identifier)
+                # if no date is available, use False for supported runtimes and True for deprecated ones
+                release.set_eoas(deprecation_date if deprecation_date else not is_supported_table)
+                # if no date is available, use False for supported runtimes and True for deprecated ones
+                release.set_eol(block_function_update if block_function_update else not is_supported_table)