search_result ブロックで自前ナレッジに根拠を引かせる — Claude Citations 実装メモ

import anthropic
 
client = anthropic.Anthropic()
 
def to_search_results(hits):
    blocks = []
    for h in hits:
        blocks.append({
            "type": "search_result",
            "source": h["url"],
            "title": h["title"],
            "content": [{"type": "text", "text": p} for p in h["paragraphs"]],
            "citations": {"enabled": True},
        })
    return blocks
 
hits = my_search("プロンプトキャッシュの TTL はどう使い分けるか", top_k=5)
 
resp = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": [
            *to_search_results(hits),
            {"type": "text", "text": "上の資料だけを根拠に、TTLの使い分けを3点でまとめてください。"},
        ],
    }],
)

# Before: 本文をひと続きの文字列に連結して渡す
context = ""
for i, h in enumerate(hits):
    context += f"## 資料{i+1}: {h['title']}\n{h['full_text']}\n\n"
 
prompt = f"""次の資料を参考に質問へ答え、末尾に [資料N] 形式で出典を書いてください。
{context}
質問: {question}"""
# → 出典は自然文。突合は正規表現頼みで、言い換えに弱い

# After: 構造化された検索結果として渡す
resp = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": [*to_search_results(hits),
                    {"type": "text", "text": question}],
    }],
)
 
# 引用は構造化データとして取れる
for block in resp.content:
    if block.type != "text":
        continue
    for c in getattr(block, "citations", None) or []:
        print(c.source, "→", c.cited_text[:40])

def render_with_footnotes(content_blocks):
    html, notes = [], []
    for block in content_blocks:
        if block.type != "text":
            continue
        cites = getattr(block, "citations", None) or []
        if not cites:
            html.append(escape(block.text))
            continue
        marks = []
        for c in cites:
            idx = len(notes) + 1
            notes.append((idx, c.source, c.title))
            marks.append(f'<sup><a href="#fn{idx}">[{idx}]</a></sup>')
        html.append(escape(block.text) + "".join(marks))
    foot = "".join(
        f'<li id="fn{i}"><a href="{escape(src)}">{escape(t)}</a></li>'
        for i, src, t in notes
    )
    return "".join(html) + f"<ol>{foot}</ol>"

# tool_use への応答として search_result を詰める
tool_result = {
    "type": "tool_result",
    "tool_use_id": tu.id,
    "content": to_search_results(my_search(tu.input["query"], top_k=5)),
}
messages.append({"role": "user", "content": [tool_result]})