Controlling snippet image on blobs

twitter2bsky_daemon.py

@@ -1,6 +1,7 @@
 import argparse
 import arrow
 import hashlib
+import io
 import json
 import logging
 import re
@@ -14,6 +15,7 @@ from atproto import Client, client_utils, models
 from playwright.sync_api import sync_playwright
 from moviepy import VideoFileClip
 from bs4 import BeautifulSoup
+from PIL import Image
 
 # --- Configuration ---
 LOG_PATH = "twitter2bsky.log"
@@ -26,6 +28,14 @@ BSKY_TEXT_MAX_LENGTH = 275
 VIDEO_MAX_DURATION_SECONDS = 179
 MAX_VIDEO_UPLOAD_SIZE_MB = 45
 
+# External-card thumbnail constraints:
+# The user's PDS returned:
+#   BlobTooLarge: 1.15MB > 976.56KB
+# So we conservatively target a slightly smaller max size for safety.
+EXTERNAL_THUMB_MAX_BYTES = 950 * 1024
+EXTERNAL_THUMB_MAX_DIMENSION = 1200
+EXTERNAL_THUMB_MIN_JPEG_QUALITY = 40
+
 BSKY_BLOB_UPLOAD_MAX_RETRIES = 5
 BSKY_BLOB_UPLOAD_BASE_DELAY = 10
 BSKY_BLOB_UPLOAD_MAX_DELAY = 300
@@ -99,11 +109,8 @@ def repair_broken_urls(text):
 
     original = text
 
-    # Join protocol line breaks: https://\nexample.com -> https://example.com
     text = re.sub(r"(https?://)\s*[\r\n]+\s*", r"\1", text, flags=re.IGNORECASE)
 
-    # Join URL-internal line breaks when the next chunk still looks like URL content.
-    # This is intentionally conservative but effective for wrapped article URLs.
     prev_text = None
     while prev_text != text:
         prev_text = text
@@ -114,7 +121,6 @@ def repair_broken_urls(text):
             flags=re.IGNORECASE
         )
 
-    # Also fix accidental spaces inserted inside URLs after the protocol.
     text = re.sub(
         r"((?:https?://|www\.)[^\s<>\"]*)\s+([A-Za-z0-9/\-._~%!$&'()*+,;=:@?#]+)",
         r"\1\2",
@@ -348,6 +354,116 @@ def get_blob_from_file(file_path, client):
         return None
 
 
+def compress_external_thumb_to_limit(image_bytes, max_bytes=EXTERNAL_THUMB_MAX_BYTES):
+    """
+    Compress/resize an image to fit external thumbnail blob size limits.
+    Returns JPEG bytes or None.
+    """
+    try:
+        with Image.open(io.BytesIO(image_bytes)) as img:
+            img = img.convert("RGB")
+
+            width, height = img.size
+            max_dim = max(width, height)
+
+            if max_dim > EXTERNAL_THUMB_MAX_DIMENSION:
+                scale = EXTERNAL_THUMB_MAX_DIMENSION / max_dim
+                new_size = (max(1, int(width * scale)), max(1, int(height * scale)))
+                img = img.resize(new_size, Image.LANCZOS)
+                logging.info(f"🖼️ Resized external thumb to {new_size[0]}x{new_size[1]}")
+
+            # Try progressively lower qualities.
+            for quality in [85, 75, 65, 55, 45, EXTERNAL_THUMB_MIN_JPEG_QUALITY]:
+                out = io.BytesIO()
+                img.save(out, format="JPEG", quality=quality, optimize=True, progressive=True)
+                data = out.getvalue()
+
+                logging.info(
+                    f"🖼️ External thumb candidate size at JPEG quality {quality}: "
+                    f"{len(data) / 1024:.2f} KB"
+                )
+
+                if len(data) <= max_bytes:
+                    return data
+
+            # If still too large, try a second resize pass.
+            for target_dim in [1000, 900, 800, 700, 600]:
+                resized = img.copy()
+                width, height = resized.size
+                max_dim = max(width, height)
+
+                if max_dim > target_dim:
+                    scale = target_dim / max_dim
+                    new_size = (max(1, int(width * scale)), max(1, int(height * scale)))
+                    resized = resized.resize(new_size, Image.LANCZOS)
+
+                for quality in [60, 50, 45, EXTERNAL_THUMB_MIN_JPEG_QUALITY]:
+                    out = io.BytesIO()
+                    resized.save(out, format="JPEG", quality=quality, optimize=True, progressive=True)
+                    data = out.getvalue()
+
+                    logging.info(
+                        f"🖼️ External thumb resized to <= {target_dim}px at quality {quality}: "
+                        f"{len(data) / 1024:.2f} KB"
+                    )
+
+                    if len(data) <= max_bytes:
+                        return data
+
+    except Exception as e:
+        logging.warning(f"Could not compress external thumbnail: {repr(e)}")
+
+    return None
+
+
+def get_external_thumb_blob_from_url(image_url, client, http_client):
+    """
+    Download, size-check, compress if needed, and upload an external-card thumbnail blob.
+    If the image cannot fit within the PDS blob limit, return None so the external card
+    can still be posted without a thumbnail.
+    """
+    try:
+        r = http_client.get(image_url, timeout=MEDIA_DOWNLOAD_TIMEOUT, follow_redirects=True)
+        if r.status_code != 200:
+            logging.warning(f"Could not fetch external thumb {image_url}: HTTP {r.status_code}")
+            return None
+
+        content = r.content
+        if not content:
+            logging.warning(f"Could not fetch external thumb {image_url}: empty body")
+            return None
+
+        original_size_kb = len(content) / 1024
+        logging.info(f"🖼️ Downloaded external thumb {image_url} ({original_size_kb:.2f} KB)")
+
+        upload_bytes = content
+        if len(upload_bytes) > EXTERNAL_THUMB_MAX_BYTES:
+            logging.info(
+                f"🖼️ External thumb exceeds safe limit "
+                f"({original_size_kb:.2f} KB > {EXTERNAL_THUMB_MAX_BYTES / 1024:.2f} KB). Compressing..."
+            )
+            compressed = compress_external_thumb_to_limit(upload_bytes, EXTERNAL_THUMB_MAX_BYTES)
+            if compressed:
+                upload_bytes = compressed
+                logging.info(f"✅ External thumb compressed to {len(upload_bytes) / 1024:.2f} KB")
+            else:
+                logging.warning("⚠️ Could not compress external thumb to fit limit. Will omit thumbnail.")
+                return None
+        else:
+            logging.info("✅ External thumb already within safe size limit.")
+
+        blob = upload_blob_with_retry(client, upload_bytes, media_label=f"external-thumb:{image_url}")
+        if blob:
+            return blob
+
+        logging.warning("⚠️ External thumb upload failed. Will omit thumbnail.")
+        return None
+
+    except Exception as e:
+        logging.warning(f"Could not fetch/upload external thumb {image_url}: {repr(e)}")
+        return None
+
+
 def fetch_link_metadata(url, http_client):
     try:
         r = http_client.get(url, timeout=LINK_METADATA_TIMEOUT, follow_redirects=True)
@@ -376,11 +492,19 @@ def fetch_link_metadata(url, http_client):
 
 
 def build_external_link_embed(url, client, http_client, fallback_title="Link"):
+    """
+    Build a Bluesky external embed from a URL.
+    If the thumbnail image is too large, omit the thumbnail but still return the link card.
+    """
     link_metadata = fetch_link_metadata(url, http_client)
 
     thumb_blob = None
     if link_metadata.get("image"):
-        thumb_blob = get_blob_from_url(link_metadata["image"], client, http_client)
+        thumb_blob = get_external_thumb_blob_from_url(link_metadata["image"], client, http_client)
+        if thumb_blob:
+            logging.info("✅ External link card thumbnail prepared successfully")
+        else:
+            logging.info("ℹ️ External link card will be posted without thumbnail")
 
     if link_metadata.get("title") or link_metadata.get("description") or thumb_blob:
         return models.AppBskyEmbedExternal.Main(
@@ -641,7 +765,7 @@ def get_recent_bsky_posts(client, handle, limit=30):
                 if item.reason is not None:
                     continue
 
-                record = item.post.record
+                    record = item.post.record
                 if getattr(record, "reply", None) is not None:
                     continue