daveadmin/dobetternorge-tools
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix sub-question expansion timeout + engine routing for Bedrock advocate

Browse Source 
- expandQueries(): truncate seedDescription to 2000 chars (full uploads were
  48K+ tokens, exceeding the 35s timeout with Sonnet); switch to Haiku gateway
  when Bedrock is active (fast + adequate for sub-Q generation); timeout → 60s
- interpretSeed(): same Haiku + 60s fix for English non-advocate path
- synthesise(): add explicit azure_mini + Bedrock → Haiku branch so the fast
  engine actually uses Haiku (~20-40s) instead of falling through to Sonnet (~180s)
- advocate.php: relabel azure_mini as "Claude Haiku 4.5 (fast)" with accurate
  timing; relabel claude_sonnet as "(thorough)" to reflect the distinction

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
davegilligan
2026-05-25 20:50:37 +02:00
parent b78ab1e257
commit 883d813f1b
2 changed files with 25 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
includes/DeepResearchAgent.php
+22 -6
View File
@@ -501,8 +501,11 @@ PROMPT;
]);
$raw = (string)($resp['choices'][0]['message']['content'] ?? '');
} else {
$raw = $this->azure->chatText([$sysMsg, $userMsg],
['json' => true, 'temperature' => 0.1, 'max_tokens' => 500, 'timeout' => 30]);
$interpGateway = ($this->azure instanceof DbnBedrockGateway)
? $this->azure->withDeployment(DbnBedrockModelRouter::LITELLM_HAIKU)
: $this->azure;
$raw = $interpGateway->chatText([$sysMsg, $userMsg],
['json' => true, 'temperature' => 0.1, 'max_tokens' => 500, 'timeout' => 60]);
}
$json = $this->azure->decodeJsonObject($raw);
if (is_array($json) && !empty($json['brief'])) {
@@ -532,6 +535,10 @@ PROMPT;
? "\nKey retrieval anchors (incorporate these terms into your sub-questions where relevant):\n" . implode(', ', $keySignals) . "\n"
: '';
// Truncate seed to 2000 chars — $brief already captures the key context;
// the full upload text (up to 192K chars) would push past the 60s timeout.
$seedExcerpt = mb_strimwidth($seedDescription, 0, 2000, '…', 'UTF-8');
if ($advocateRole !== '') {
$prompt = <<<PROMPT
You are a Norwegian family-law research assistant building a case for: {$advocateRole}.
@@ -546,7 +553,7 @@ Research brief:
{$brief}
{$anchorsLine}
Raw input:
{$seedDescription}
{$seedExcerpt}
Return JSON only in {$locale}:
{
@@ -571,7 +578,7 @@ Research brief:
{$brief}
{$anchorsLine}
Raw input:
{$seedDescription}
{$seedExcerpt}
Return JSON only:
{
@@ -600,8 +607,11 @@ PROMPT;
]);
$raw = (string)($resp['choices'][0]['message']['content'] ?? '');
} else {
$raw = $this->azure->chatText([$sysMsg, $userMsg],
['json' => true, 'temperature' => 0.2, 'max_tokens' => 700, 'timeout' => 35]);
$expGateway = ($this->azure instanceof DbnBedrockGateway)
? $this->azure->withDeployment(DbnBedrockModelRouter::LITELLM_HAIKU)
: $this->azure;
$raw = $expGateway->chatText([$sysMsg, $userMsg],
['json' => true, 'temperature' => 0.2, 'max_tokens' => 700, 'timeout' => 60]);
}
$json = $this->azure->decodeJsonObject($raw);
$items = is_array($json['sub_questions'] ?? null) ? $json['sub_questions'] : [];
@@ -1142,6 +1152,12 @@ PROMPT;
} elseif ($engine === 'azure_full') {
$raw = $this->azure->withDeployment('gpt-4o')->chatText($messages, $opts);
$deployLabel = 'gpt-4o';
} elseif ($engine === 'azure_mini' && $this->azure instanceof DbnBedrockGateway) {
// When Bedrock enabled, azure_mini → Haiku (fast, ~20-40s synthesis)
$haiku = $this->azure->withDeployment(DbnBedrockModelRouter::LITELLM_HAIKU);
$raw = $haiku->chatText($messages, array_merge($opts, ['max_tokens' => 2500, 'timeout' => 90]));
$deployLabel = 'Claude Haiku 4.5 (AWS Bedrock)';
$thinkingTrace = null;
} elseif ($engine === 'claude_sonnet' || ($this->azure instanceof DbnBedrockGateway)) {
if (
$this->azure instanceof DbnBedrockGateway