summaryrefslogtreecommitdiff
path: root/bin
diff options
context:
space:
mode:
authoremkael <emkael@tlen.pl>2024-11-25 23:24:26 +0100
committeremkael <emkael@tlen.pl>2024-11-25 23:24:26 +0100
commit07dc74d9e73585d840854c642523f0fc05075df8 (patch)
treea76bf82f6faee22a57bd6834595894161c98700c /bin
parent373141ac8765c76f06c496a51b320852447c20ba (diff)
Fuck off, Zuck
Diffstat (limited to 'bin')
-rw-r--r--bin/fb-scrape/get-fb-content.py7
-rw-r--r--bin/fb-scrape/requirements.txt2
-rwxr-xr-xbin/refresh-fb-cache.php8
3 files changed, 12 insertions, 5 deletions
diff --git a/bin/fb-scrape/get-fb-content.py b/bin/fb-scrape/get-fb-content.py
index 5b22f70..6d6b7ba 100644
--- a/bin/fb-scrape/get-fb-content.py
+++ b/bin/fb-scrape/get-fb-content.py
@@ -8,6 +8,7 @@ from pytz_deprecation_shim import PytzUsageWarning
from random import randint
from facebook_scraper import get_posts, enable_logging
+from facebook_scraper.exceptions import TemporarilyBanned
from requests.exceptions import RequestException
debug = len(sys.argv) > 2 and sys.argv[2] == 'debug'
@@ -21,6 +22,10 @@ warnings.filterwarnings(
)
warnings.filterwarnings(
action='ignore',
+ message=r"Facebook says 'Unsupported Browser'"
+)
+warnings.filterwarnings(
+ action='ignore',
category=PytzUsageWarning
)
@@ -36,7 +41,7 @@ try:
'images': post['images']
})
time.sleep(randint(10, 15))
-except RequestException:
+except (RequestException, TemporarilyBanned):
pass
print(json.dumps(posts))
diff --git a/bin/fb-scrape/requirements.txt b/bin/fb-scrape/requirements.txt
new file mode 100644
index 0000000..6632c5e
--- /dev/null
+++ b/bin/fb-scrape/requirements.txt
@@ -0,0 +1,2 @@
+facebook-scraper
+pytz-deprecation-shim
diff --git a/bin/refresh-fb-cache.php b/bin/refresh-fb-cache.php
index ca50146..6a841de 100755
--- a/bin/refresh-fb-cache.php
+++ b/bin/refresh-fb-cache.php
@@ -13,7 +13,7 @@ $cacheFiles = glob('../cache/facebook.*');
$fileToFetch = NULL;
$oldestCache = PHP_INT_MAX;
-foreach ($cacheFiles as $file) {
+/*foreach ($cacheFiles as $file) {
if (!preg_match('/\.Piwoteka$/', $file)) {
$cacheTime = filemtime($file);
if ($cacheTime < $oldestCache) {
@@ -23,12 +23,12 @@ foreach ($cacheFiles as $file) {
$fileToFetch = implode('.', $fileToFetch);
}
}
-}
+ }*/
-$filesToFetch = ['Piwoteka', 'fermentlodz', $fileToFetch];
+$filesToFetch = ['Piwoteka', 'EtreGourmet']; //, $fileToFetch];
foreach ($filesToFetch as $feed) {
- sleep(rand(60, 90));
+ sleep(rand(90, 120));
$provider = new \Providers\Facebook($feed, ['force' => TRUE]);
if (!$provider->get()) {
throw new \Exception(sprintf('Feed %s failed to fetch!', $feed));