Embedding dedup hash uses only first 60 chars - updated content goes stale silently

[jiya2401]

• embed_and_store() computes the dedup key as sha256(f"{url}|{i}|{chunk[:60]}") - only the first 60 characters of each chunk are hashed, not the full text.
• If a scraped blog post is edited but the first 60 characters of a chunk stay the same, the new content is wrongly treated as a duplicate and skipped.
• Result: the stale embedding stays in ChromaDB indefinitely, and your RAG pipeline analyzes outdated competitor data without you knowing.