From 00585f72278274d04b6979536861ad901d7112d4 Mon Sep 17 00:00:00 2001
From: "Antoine R. Dumont (@ardumont)" <ardumont@softwareheritage.org>
Date: Mon, 10 Jul 2023 10:04:10 +0200
Subject: [PATCH] gitweb: Parse the last update interval as a last update
Refs. swh/devel/swh-lister#1800
---
mypy.ini | 3 +++
requirements.txt | 1 +
swh/lister/gitweb/lister.py | 18 ++++++++++++------
swh/lister/gitweb/tests/test_lister.py | 3 +--
swh/lister/utils.py | 5 +++++
5 files changed, 22 insertions(+), 8 deletions(-)
diff --git a/mypy.ini b/mypy.ini
index 7f9436b9..76468c2a 100644
--- a/mypy.ini
+++ b/mypy.ini
@@ -43,6 +43,9 @@ ignore_missing_imports = True
[mypy-dulwich.*]
ignore_missing_imports = True
+[mypy-dateparser.*]
+ignore_missing_imports = True
+
[mypy-testing.postgresql.*]
ignore_missing_imports = True
diff --git a/requirements.txt b/requirements.txt
index 2614f0a5..0e588063 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -6,6 +6,7 @@ beautifulsoup4
launchpadlib
tenacity >= 6.2
lxml
+dateparser
dulwich
testing.postgresql
psycopg2
diff --git a/swh/lister/gitweb/lister.py b/swh/lister/gitweb/lister.py
index 57810a9e..3c902e4f 100644
--- a/swh/lister/gitweb/lister.py
+++ b/swh/lister/gitweb/lister.py
@@ -8,9 +8,11 @@ from typing import Any, Dict, Iterator, List, Optional
from urllib.parse import urljoin, urlparse
from bs4 import BeautifulSoup
+from dateparser import parse
from requests.exceptions import HTTPError
from swh.lister.pattern import CredentialsType, StatelessLister
+from swh.lister.utils import now
from swh.scheduler.interface import SchedulerInterface
from swh.scheduler.model import ListedOrigin
@@ -60,6 +62,7 @@ class GitwebLister(StatelessLister[Repositories]):
)
self.session.headers.update({"Accept": "application/html"})
+ self.listing_date = now()
def _get_and_parse(self, url: str) -> BeautifulSoup:
"""Get the given url and parse the retrieved HTML using BeautifulSoup"""
@@ -87,13 +90,12 @@ class GitwebLister(StatelessLister[Repositories]):
if repo_url.endswith("?o=descr"):
continue
- # FIXME: Add parsing step from date interval like '9 years ago' to
+ # This retrieves the date interval in natural language (e.g. '9 years ago') to
# actual python datetime interval so we can derive last update
- # span = tr.find("td", {"class": re.compile("age.*")})
- # last_updated_date = span.get("title") if span else None
- # last_updated_date = None
-
- page_results.append({"url": repo_url})
+ span = tr.find("td", {"class": re.compile("age.*")})
+ page_results.append(
+ {"url": repo_url, "last_update_interval": span.text if span else None}
+ )
yield page_results
@@ -108,10 +110,14 @@ class GitwebLister(StatelessLister[Repositories]):
if origin_url is None:
continue
+ last_update_interval = repo.get("last_update_interval")
+ last_update = parse(last_update_interval) if last_update_interval else None
+
yield ListedOrigin(
lister_id=self.lister_obj.id,
url=origin_url,
visit_type="git",
+ last_update=last_update,
)
def _get_origin_from_repository_url(self, repository_url: str) -> Optional[str]:
diff --git a/swh/lister/gitweb/tests/test_lister.py b/swh/lister/gitweb/tests/test_lister.py
index 1f784b1c..e52546d7 100644
--- a/swh/lister/gitweb/tests/test_lister.py
+++ b/swh/lister/gitweb/tests/test_lister.py
@@ -70,8 +70,7 @@ def test_lister_gitweb_run(requests_mock_datadir, swh_scheduler):
for listed_origin in scheduler_origins:
assert listed_origin.visit_type == "git"
assert listed_origin.url.startswith(url)
- # Not parsed
- assert listed_origin.last_update is None
+ assert listed_origin.last_update is not None
# test user agent content
for request in requests_mock_datadir.request_history:
diff --git a/swh/lister/utils.py b/swh/lister/utils.py
index 60cfc933..5d7a6e75 100644
--- a/swh/lister/utils.py
+++ b/swh/lister/utils.py
@@ -2,6 +2,7 @@
# License: GNU General Public License version 3, or any later version
# See top-level LICENSE file for more information
+from datetime import datetime, timezone
from typing import Iterator, Optional, Tuple
import urllib.parse
@@ -75,3 +76,7 @@ def is_valid_origin_url(url: Optional[str]) -> bool:
return False
return True
+
+
+def now() -> datetime:
+ return datetime.now(tz=timezone.utc)
--
GitLab