From 375191c497cd3e870148083c725a7851756f0c07 Mon Sep 17 00:00:00 2001
From: Guillem Hernandez Sola <guillem.hernandez.sola@gmail.com>
Date: Thu, 30 Apr 2026 10:20:41 +0200
Subject: [PATCH] Add --bsky-langs es to Jenkins pipeline for specific language
 tagging

---
 jenkins/3catTw         |  3 ++-
 twitter2bsky_daemon.py | 40 ++++++++++++++++++++++++++++++++++++----
 2 files changed, 38 insertions(+), 5 deletions(-)

diff --git a/jenkins/3catTw b/jenkins/3catTw
index a38bbdc..7fe9926 100644
--- a/jenkins/3catTw
+++ b/jenkins/3catTw
@@ -82,7 +82,8 @@ pipeline {
                             --twitter-handle "$TWITTER_3CAT_HANDLE" \
                             --bsky-handle "$BSKY_3CAT_HANDLE" \
                             --bsky-password "$BSKY_3CAT_APP_PASSWORD" \
-                            --bsky-base-url https://eurosky.social
+                            --bsky-base-url https://eurosky.social \
+                            --bsky-langs ca
                     '''
                 }
             }
diff --git a/twitter2bsky_daemon.py b/twitter2bsky_daemon.py
index 72fb1f2..53411bf 100644
--- a/twitter2bsky_daemon.py
+++ b/twitter2bsky_daemon.py
@@ -97,6 +97,7 @@ class _RunCache:
         self.og_title: dict = {}
         self.url_resolution: dict = {}
         self.url_validity: dict = {}
+        self.locale: str = "en-US"   # ← ADDED locale cache here
 
     def clear(self):
         self.og_title.clear()
@@ -113,6 +114,30 @@ def grapheme_len(text):
     """Return the grapheme cluster count, matching Bluesky's character counting."""
     return grapheme.length(text)
 
+# BCP-47 language tag → sensible locale for Playwright
+_LANG_TO_LOCALE = {
+    "ca": "ca-ES",
+    "es": "es-ES",
+    "en": "en-US",
+    "fr": "fr-FR",
+    "de": "de-DE",
+    "pt": "pt-PT",
+    "it": "it-IT",
+    "nl": "nl-NL",
+    "eu": "eu-ES",
+    "gl": "gl-ES",
+}
+
+def bsky_langs_to_playwright_locale(bsky_langs):
+    """
+    Convert the first configured Bluesky language tag to a Playwright locale
+    string (e.g. ['ca'] → 'ca-ES'). Falls back to 'en-US' if unknown.
+    """
+    if not bsky_langs:
+        return "en-US"
+    primary = bsky_langs[0].strip().lower()
+    return _LANG_TO_LOCALE.get(primary, f"{primary}-{primary.upper()}")
+
 # --- Custom Classes ---
 class ScrapedMedia:
     def __init__(self, url, media_type="photo"):
@@ -562,7 +587,7 @@ def should_fetch_og_title(tweet):
     return False
 
 
-def fetch_tweet_og_title_text(tweet_url):
+def fetch_tweet_og_title_text(tweet_url, locale="en-US"):
     if not tweet_url:
         return None
 
@@ -589,6 +614,7 @@ def fetch_tweet_og_title_text(tweet_url):
                     "Chrome/145.0.7632.6 Safari/537.36"
                 ),
                 viewport={"width": 1280, "height": 900},
+                locale=_cache.locale,   # ← USE CACHE   
             )
             page = browser_context.new_page()
             page.goto(
@@ -665,7 +691,7 @@ def resolve_tco_with_httpx(url, http_client):
     return canonicalize_url(url)
 
 
-def resolve_tco_with_playwright(url):
+def resolve_tco_with_playwright(url, locale="en-US"):
     browser = None
     browser_context = None
     page = None
@@ -685,6 +711,7 @@ def resolve_tco_with_playwright(url):
                     "Chrome/145.0.7632.6 Safari/537.36"
                 ),
                 viewport={"width": 1280, "height": 900},
+                locale=locale,
             )
             page = browser_context.new_page()
 
@@ -2337,7 +2364,7 @@ def build_video_embed(video_blob, alt_text):
 
 
 # --- Twitter Scraping ---
-def scrape_tweets_via_playwright(username, password, email, target_handle):
+def scrape_tweets_via_playwright(username, password, email, target_handle, locale="en-US"):
     tweets = []
     state_file = "twitter_browser_state.json"
 
@@ -2372,6 +2399,7 @@ def scrape_tweets_via_playwright(username, password, email, target_handle):
                 user_agent=clean_ua,
                 viewport={"width": 1920, "height": 1080},
                 storage_state=state_file,
+                locale=locale,
             )
             session_check_page = browser_context.new_page()
             session_check_page.goto("https://x.com/home")
@@ -2406,6 +2434,7 @@ def scrape_tweets_via_playwright(username, password, email, target_handle):
             browser_context = browser.new_context(
                 user_agent=clean_ua,
                 viewport={"width": 1920, "height": 1080},
+                locale=locale,   # ✅ add this
             )
             login_page = browser_context.new_page()
 
@@ -2980,9 +3009,10 @@ def candidate_matches_existing_bsky(candidate, recent_bsky_posts):
 # --- Main Sync Logic ---
 def sync_feeds(args):
     logging.info("🔄 Starting sync cycle...")
-
     dry_run = getattr(args, "dry_run", False)
     bsky_langs = getattr(args, "bsky_langs", None) or DEFAULT_BSKY_LANGS
+    bot_locale = bsky_langs_to_playwright_locale(bsky_langs)  # ✅ now defined
+    _cache.locale = bot_locale     
 
     if dry_run:
         logging.info("🧪 DRY RUN MODE — no posts will be created on Bluesky.")
@@ -2996,6 +3026,7 @@ def sync_feeds(args):
             args.twitter_password,
             args.twitter_email,
             args.twitter_handle,
+            locale=bot_locale,  
         )
 
         if not tweets:
@@ -3213,6 +3244,7 @@ def sync_feeds(args):
                     "Chrome/145.0.7632.6 Safari/537.36"
                 ),
                 "viewport": {"width": 1920, "height": 1080},
+                "locale": bot_locale, 
             }
             if os.path.exists(browser_state_file):
                 context_kwargs["storage_state"] = browser_state_file