Fix linkedin parsing

amadejkastelic · Nov 29, 2024 · e390206 · e390206
1 parent ea01859
commit e390206
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 8 deletions.
diff --git a/bot/common/utils.py b/bot/common/utils.py
@@ -149,9 +149,6 @@ def date_to_human_format(date: datetime.datetime) -> str:
 
 def parse_relative_time(relative_time: str) -> datetime.timedelta:
     units = {
-        'y': 'years',
-        'mo': 'months',
-        'w': 'weeks',
         'd': 'days',
         'h': 'hours',
         'm': 'minutes',
@@ -165,7 +162,17 @@ def parse_relative_time(relative_time: str) -> datetime.timedelta:
     unit = ''.join([ch for ch in relative_time if ch.isalpha()])
 
     if unit not in units:
-        raise ValueError(f"Unsupported time unit: {unit}")
+        if unit in ['w', 'week', 'weeks']:
+            number *= 7
+            unit = 'd'
+        elif unit in ['mo', 'month', 'months']:
+            number *= 30
+            unit = 'd'
+        elif unit in ['y', 'year', 'years']:
+            number *= 365
+            unit = 'd'
+        else:
+            raise ValueError(f"Unsupported time unit: {unit}")
 
     # Return the appropriate timedelta
     return datetime.timedelta(**{units[unit]: number})

diff --git a/bot/integrations/linkedin/client.py b/bot/integrations/linkedin/client.py
@@ -45,11 +45,11 @@ async def get_post(self, url: str) -> domain.Post:
             page = await context.new_page()
             await page.goto(url)
 
-            author = await page.locator('[data-tracking-control-name="public_post_feed-actor-name"]').inner_text()
-            description = await page.locator('[data-test-id="main-feed-activity-card__commentary"]').inner_text()
-            likes_text = await page.locator('[data-test-id="social-actions__reaction-count"]').inner_text()
+            author = await page.locator('[data-tracking-control-name="public_post_feed-actor-name"]').first.inner_text()
+            description = await page.locator('[data-test-id="main-feed-activity-card__commentary"]').first.inner_text()
+            likes_text = await page.locator('[data-test-id="social-actions__reaction-count"]').first.inner_text()
             likes = int(likes_text.replace(',', '').replace('.', '') or 0)
-            relative_time = await page.locator('time').inner_text()
+            relative_time = await page.locator('time').first.inner_text()
 
             post = domain.Post(
                 url=url,