GeneralBots/generalbots
1
0
Fork 0
Code Activity Actions 5

remove: HTML tag balancing and buffering - Remove HTML buffer accumulation - Send LLM chunks immediately without processing - Remove tag sanitization and balance checks - HTML now streams raw from LLM to client

Browse source
This commit is contained in:
Rodrigo Rodriguez 2026-04-30 21:32:42 -03:00
parent 86d4bb4f26
commit b31e08004f
1 changed files with 22 additions and 63 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

85
botserver/src/core/bot/mod.rs
View file

@ -1346,73 +1346,32 @@ while let Some(chunk) = stream_rx.recv().await {
trace!("Cleared leftover in_analysis state");
}
if !in_analysis {
full_response.push_str(&chunk);
html_buffer.push_str(&chunk);
if !in_analysis {
full_response.push_str(&chunk);
// Check if we should flush the buffer:
// 1. HTML tag pair completed (e.g., </div>, </h1>, </p>, </ul>, </li>)
// 2. Buffer is large enough (> 500 chars)
// 3. This is the last chunk (is_complete will be true next iteration)
let should_flush = (html_buffer.len() > 1000
|| html_buffer.contains("</div>")
|| html_buffer.contains("</h1>")
|| html_buffer.contains("</h2>")
|| html_buffer.contains("</h3>")
|| html_buffer.contains("</h4>")
|| html_buffer.contains("</h5>")
|| html_buffer.contains("</h6>")
|| html_buffer.contains("</p>")
|| html_buffer.contains("</ul>")
|| html_buffer.contains("</ol>")
|| html_buffer.contains("</li>")
|| html_buffer.contains("</table>")
|| html_buffer.contains("</tr>")
|| html_buffer.contains("</td>")
|| html_buffer.contains("</th>")
|| html_buffer.contains("</section>")
|| html_buffer.contains("</header>")
|| html_buffer.contains("</footer>")
|| html_buffer.contains("</azul>")
|| html_buffer.contains("</ouro>")
|| html_buffer.contains("</hero>")
|| html_buffer.contains("</wrap>")
|| html_buffer.contains("</corpo>")
|| html_buffer.contains("</item>")
|| html_buffer.contains("</rodape>"))
&& !is_inside_html_tag(&html_buffer);
// Send immediately without buffering
let response = BotResponse {
bot_id: message.bot_id.clone(),
user_id: message.user_id.clone(),
session_id: message.session_id.clone(),
channel: message.channel.clone(),
content: chunk.clone(),
message_type: MessageType::BOT_RESPONSE,
stream_token: None,
is_complete: false,
suggestions: Vec::new(),
switchers: Vec::new(),
context_name: None,
context_length: 0,
context_max_length: 0,
};
if should_flush {
// Sanitize malformed tags (e.g. "< class" -> "<class", "</ div" -> "</div")
let content_to_send = html_buffer
.replace("< ", "<")
.replace("</ ", "</");
html_buffer.clear();
let response = BotResponse {
bot_id: message.bot_id.clone(),
user_id: message.user_id.clone(),
session_id: message.session_id.clone(),
channel: message.channel.clone(),
content: content_to_send,
message_type: MessageType::BOT_RESPONSE,
stream_token: None,
is_complete: false,
suggestions: Vec::new(),
switchers: Vec::new(),
context_name: None,
context_length: 0,
context_max_length: 0,
};
if response_tx.send(response).await.is_err() {
warn!("Response channel closed");
break;
}
}
if response_tx.send(response).await.is_err() {
warn!("Response channel closed");
break;
}
}
}
info!("llm_end: Streaming loop ended for session {}, chunk_count={}, full_response_len={}", session.id, chunk_count, full_response.len());