local-deep-research 0.6.0__tar.gz → 0.6.1__tar.gz

{local_deep_research-0.6.0 → local_deep_research-0.6.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: local-deep-research
-Version: 0.6.0
+Version: 0.6.1
 Summary: AI-powered research assistant with deep, iterative analysis using LLMs and web searches
 Author-Email: LearningCircuit <185559241+LearningCircuit@users.noreply.github.com>, HashedViking <6432677+HashedViking@users.noreply.github.com>, djpetti <djpetti@gmail.com>
 License: MIT License
@@ -302,7 +302,7 @@ Early experiments on small SimpleQA dataset samples:
 | Configuration | Accuracy | Notes |
 |--------------|----------|--------|
 | gpt-4.1-mini + SearXNG + focused_iteration | 90-95% | Limited sample size |
-| gpt-4.1-mini + Tavily | Up to 95% | Limited sample size |
+| gpt-4.1-mini + Tavily + focused_iteration | 90-95% | Limited sample size |
 | gemini-2.0-flash-001 + SearXNG | 82% | Single test run |
 
 Note: These are preliminary results from initial testing. Performance varies significantly based on query types, model versions, and configurations. [Run your own benchmarks →](docs/BENCHMARKING.md)

{local_deep_research-0.6.0 → local_deep_research-0.6.1}/README.md

@@ -220,7 +220,7 @@ Early experiments on small SimpleQA dataset samples:
 | Configuration | Accuracy | Notes |
 |--------------|----------|--------|
 | gpt-4.1-mini + SearXNG + focused_iteration | 90-95% | Limited sample size |
-| gpt-4.1-mini + Tavily | Up to 95% | Limited sample size |
+| gpt-4.1-mini + Tavily + focused_iteration | 90-95% | Limited sample size |
 | gemini-2.0-flash-001 + SearXNG | 82% | Single test run |
 
 Note: These are preliminary results from initial testing. Performance varies significantly based on query types, model versions, and configurations. [Run your own benchmarks →](docs/BENCHMARKING.md)

{local_deep_research-0.6.0 → local_deep_research-0.6.1}/pyproject.toml

@@ -69,7 +69,7 @@ dependencies = [
     "aiohttp>=3.9.0",
     "tenacity>=8.0.0",
 ]
-version = "0.6.0"
+version = "0.6.1"
 
 [project.license]
 file = "LICENSE"

local_deep_research-0.6.1/src/local_deep_research/__version__.py

@@ -0,0 +1 @@
+__version__ = "0.6.1"

{local_deep_research-0.6.0 → local_deep_research-0.6.1}/src/local_deep_research/web/templates/pages/benchmark_results.html

@@ -6,6 +6,7 @@
 
 {% block extra_head %}
 <meta name="csrf-token" content="{{ csrf_token() }}">
+<meta name="app-version" content="{{ version }}">
 <style>
 .benchmark-results-card {
     width: 100%;
@@ -769,14 +770,21 @@ function createRunCard(run) {
                     <div class="summary-label">Avg Time/Question</div>
                 </div>
                 <div class="summary-item">
-                    ${status === 'in_progress' ?
-                        `<button class="btn btn-outline btn-sm delete-btn" onclick="event.stopPropagation(); cancelAndDeleteBenchmarkRun(${run.id})" style="background: #3a1e1e !important; border-color: #f44336 !important; color: #f44336 !important;">
-                            <i class="fas fa-stop"></i> Cancel & Delete
-                        </button>` :
-                        `<button class="btn btn-outline btn-sm delete-btn" onclick="event.stopPropagation(); deleteBenchmarkRun(${run.id})">
-                            <i class="fas fa-trash"></i> Delete
-                        </button>`
-                    }
+                    <div style="display: flex; gap: 8px;">
+                        ${status === 'completed' ?
+                            `<button class="btn btn-outline btn-sm" onclick="event.stopPropagation(); downloadBenchmarkYAML(${run.id})" style="background: #1e3a1e !important; border-color: #4caf50 !important; color: #4caf50 !important;">
+                                <i class="fas fa-download"></i> YAML
+                            </button>` : ''
+                        }
+                        ${status === 'in_progress' ?
+                            `<button class="btn btn-outline btn-sm delete-btn" onclick="event.stopPropagation(); cancelAndDeleteBenchmarkRun(${run.id})" style="background: #3a1e1e !important; border-color: #f44336 !important; color: #f44336 !important;">
+                                <i class="fas fa-stop"></i> Cancel & Delete
+                            </button>` :
+                            `<button class="btn btn-outline btn-sm delete-btn" onclick="event.stopPropagation(); deleteBenchmarkRun(${run.id})">
+                                <i class="fas fa-trash"></i> Delete
+                            </button>`
+                        }
+                    </div>
                     <div class="summary-label">Actions</div>
                 </div>
             </div>
@@ -1061,6 +1069,136 @@ async function cancelAndDeleteBenchmarkRun(runId) {
     }
 }
 
+async function downloadBenchmarkYAML(runId) {
+    try {
+        // Find the run in our local data
+        const run = benchmarkRuns.find(r => r.id === runId);
+        if (!run) {
+            showAlert('Benchmark run not found', 'error');
+            return;
+        }
+
+        // Get current date for filename
+        const date = new Date().toISOString().split('T')[0];
+
+        // Get app version from meta tag
+        const appVersion = document.querySelector('meta[name="app-version"]')?.content || 'Could not fetch version';
+
+        // Extract model name and clean it for filename
+        const modelName = run.search_config?.model_name || 'unknown-model';
+        const cleanModelName = modelName.replace(/[^a-zA-Z0-9.-]/g, '-').toLowerCase();
+
+        // Get all relevant settings from database
+        let localContextWindow = 'Could not fetch';
+        let maxTokens = 'Could not fetch';
+        let contextWindowUnrestricted = 'Could not fetch';
+        let contextWindowSize = 'Could not fetch';
+        let supportsMaxTokens = 'Could not fetch';
+
+        try {
+            const settingsResponse = await fetch('/settings/api');
+            if (settingsResponse.ok) {
+                const data = await settingsResponse.json();
+                if (data.status === 'success' && data.settings) {
+                    const settings = data.settings;
+                    // LLM settings - extract the 'value' property from each setting object
+                    localContextWindow = settings['llm.local_context_window_size']?.value || 'Could not fetch';
+                    maxTokens = settings['llm.max_tokens']?.value || 'Could not fetch';
+                    contextWindowUnrestricted = settings['llm.context_window_unrestricted']?.value !== undefined ?
+                        (settings['llm.context_window_unrestricted'].value ? 'Yes' : 'No') : 'Could not fetch';
+                    contextWindowSize = settings['llm.context_window_size']?.value || 'Could not fetch';
+                    supportsMaxTokens = settings['llm.supports_max_tokens']?.value !== undefined ?
+                        (settings['llm.supports_max_tokens'].value ? 'Yes' : 'No') : 'Could not fetch';
+                }
+            }
+        } catch (e) {
+            console.error('Could not fetch current settings:', e);
+        }
+
+        // Calculate average search results if available
+        const avgSearchResults = formatAvgSearchResults(run).replace(' results', '');
+        const searchResultsNum = avgSearchResults !== 'N/A' ? avgSearchResults : '# Please fill in';
+
+        // Generate YAML content
+        const yamlContent = `# Benchmark Result
+# Generated from Local Deep Research v${appVersion}
+# Date: ${date}
+
+# Model Information
+model: ${modelName}
+model_provider: ${run.search_config?.provider || 'unknown'}
+quantization: # Please fill in if applicable
+
+# Search Engine (critical for benchmark reproducibility)
+search_engine: ${run.search_config?.search_tool || 'unknown'}
+search_provider_version: # if known, e.g., "latest", "2024.1.0"
+average_results_per_query: ${searchResultsNum}
+
+# Hardware
+hardware:
+  gpu: # Please fill in
+  ram: # Please fill in
+  cpu: # Please fill in
+
+# Benchmark Results
+results:
+  dataset: SimpleQA
+  total_questions: ${run.total_examples}
+
+  ${run.search_config?.search_strategy === 'focused_iteration' ? 'focused_iteration' : 'source_based'}:
+    accuracy: ${run.overall_accuracy ? run.overall_accuracy.toFixed(1) : 0}% (${Math.round(run.overall_accuracy * run.total_examples / 100)}/${run.total_examples})
+    iterations: ${run.search_config?.iterations || 'N/A'}
+    questions_per_iteration: ${run.search_config?.questions_per_iteration || 'N/A'}
+    avg_time_per_question: ${formatAvgProcessingTime(run)}
+    total_tokens_used: # if available
+
+# Configuration
+configuration:
+  context_window: ${localContextWindow} # Current setting at download time - may differ from benchmark run
+  temperature: ${run.search_config?.temperature || 'N/A'}
+  max_tokens: ${maxTokens} # Current setting at download time
+  local_provider_context_window_size: ${localContextWindow} # Current setting at download time
+  context_window_unrestricted: ${contextWindowUnrestricted} # Current setting at download time
+
+# Versions
+versions:
+  ldr_version: ${appVersion}
+  ollama_version: # if applicable
+
+# Test Details
+test_details:
+  date_tested: ${date}
+  rate_limiting_issues: # yes/no
+  search_failures: # number of failed searches, if any
+
+# Notes
+notes: |
+  # Add any observations, errors, or insights here
+  # Search strategy: ${run.search_config?.search_strategy || 'unknown'}
+  # Provider: ${run.search_config?.provider || 'unknown'}
+  # Note: Configuration values are from current settings at download time,
+  # not necessarily the values used during the benchmark run
+`;
+
+        // Create blob and download
+        const blob = new Blob([yamlContent], { type: 'text/yaml' });
+        const url = window.URL.createObjectURL(blob);
+        const a = document.createElement('a');
+        a.style.display = 'none';
+        a.href = url;
+        a.download = `${cleanModelName}_${date}.yaml`;
+        document.body.appendChild(a);
+        a.click();
+        window.URL.revokeObjectURL(url);
+        document.body.removeChild(a);
+
+        showAlert('Benchmark YAML downloaded! Hardware details are optional but helpful for performance context.', 'success');
+    } catch (error) {
+        console.error('Error downloading YAML:', error);
+        showAlert('Error downloading YAML: ' + error.message, 'error');
+    }
+}
+
 async function deleteBenchmarkRun(runId) {
     try {
         const response = await fetch(`/benchmark/api/delete/${runId}`, {

local_deep_research-0.6.1/tests/package.json

@@ -0,0 +1,5 @@
+{
+  "dependencies": {
+    "puppeteer": "^24.9.0"
+  }
+}

local_deep_research-0.6.1/tests/ui_tests/test_simple_cost.js

@@ -0,0 +1,155 @@
+const puppeteer = require('puppeteer');
+
+(async () => {
+    const browser = await puppeteer.launch({
+        headless: false,
+        args: ['--no-sandbox'],
+        defaultViewport: { width: 1920, height: 1080 }
+    });
+
+    const page = await browser.newPage();
+
+    // Monitor console
+    page.on('console', msg => {
+        console.log(`CONSOLE ${msg.type()}: ${msg.text()}`);
+    });
+
+    // Monitor network
+    page.on('response', async (response) => {
+        if (response.url().includes('/metrics/api/cost-analytics')) {
+            console.log(`API Response: ${response.status()}`);
+            try {
+                const text = await response.text();
+                console.log('API Data:', text.substring(0, 200));
+            } catch (e) {}
+        }
+    });
+
+    console.log('Loading page...');
+    // Clear cache and reload
+    await page.setCacheEnabled(false);
+    await page.goto('http://localhost:5000/metrics/costs?' + Date.now(), {
+        waitUntil: 'domcontentloaded',
+        timeout: 15000
+    });
+
+    // Wait for initial load
+    await new Promise(resolve => setTimeout(resolve, 3000));
+
+    // Check loading state
+    const states = await page.evaluate(() => {
+        const loading = document.getElementById('loading');
+        const error = document.getElementById('error');
+        const content = document.getElementById('cost-content');
+        const noData = document.getElementById('no-data');
+
+        return {
+            loading: loading ? loading.style.display : 'not found',
+            error: error ? error.style.display : 'not found',
+            content: content ? content.style.display : 'not found',
+            noData: noData ? noData.style.display : 'not found'
+        };
+    });
+
+    console.log('Element states:', states);
+
+    // Try manual API call and simulate the loadCostData function
+    const apiTest = await page.evaluate(async () => {
+        try {
+            console.log('Starting manual API test...');
+            const response = await fetch('/metrics/api/cost-analytics?period=7d');
+            const data = await response.json();
+            console.log('API Response received:', data.status);
+            console.log('Total calls:', data.overview?.total_calls);
+            console.log('Total cost:', data.overview?.total_cost);
+
+            // Now simulate what the loadCostData function should do
+            if (data.status === 'success') {
+                console.log('Status is success');
+                if (data.overview.total_calls > 0) {
+                    console.log('Total calls > 0, should show content');
+
+                    // Try to manually call the show functions
+                    function showContent() {
+                        document.getElementById('loading').style.display = 'none';
+                        document.getElementById('cost-content').style.display = 'block';
+                        document.getElementById('error').style.display = 'none';
+                        document.getElementById('no-data').style.display = 'none';
+                    }
+
+                    showContent();
+                    console.log('Called showContent() manually');
+                } else {
+                    console.log('Total calls is 0, should show no data');
+                }
+            } else {
+                console.log('Status is not success:', data.status);
+            }
+
+            return { success: true, data: data };
+        } catch (error) {
+            console.error('Manual API test failed:', error);
+            return { success: false, error: error.message };
+        }
+    });
+
+    console.log('Manual API test:', apiTest.success ? 'SUCCESS' : 'FAILED');
+    if (apiTest.success) {
+        console.log('API returned cost:', apiTest.data.overview?.total_cost);
+        console.log('API returned calls:', apiTest.data.overview?.total_calls);
+    }
+
+    // Test the actual logic from the page
+    const testLogic = await page.evaluate(() => {
+        // Simulate what the loadCostData function does
+        const testData = {
+            status: 'success',
+            overview: {
+                total_cost: 0.0,
+                total_calls: 58
+            }
+        };
+
+        console.log('Testing logic: total_calls =', testData.overview.total_calls);
+        console.log('Testing logic: total_cost =', testData.overview.total_cost);
+
+        if (testData.status === 'success') {
+            if (testData.overview.total_calls > 0) {
+                console.log('Should show content!');
+                return 'should_show_content';
+            } else {
+                console.log('Should show no data');
+                return 'should_show_no_data';
+            }
+        }
+        return 'unknown';
+    });
+
+    console.log('Logic test result:', testLogic);
+
+    // Wait a bit more
+    await new Promise(resolve => setTimeout(resolve, 3000));
+
+    // Check final states after manual showContent
+    const finalStates = await page.evaluate(() => {
+        const loading = document.getElementById('loading');
+        const error = document.getElementById('error');
+        const content = document.getElementById('cost-content');
+        const noData = document.getElementById('no-data');
+
+        return {
+            loading: loading ? loading.style.display : 'not found',
+            error: error ? error.style.display : 'not found',
+            content: content ? content.style.display : 'not found',
+            noData: noData ? noData.style.display : 'not found'
+        };
+    });
+
+    console.log('Final element states:', finalStates);
+
+    console.log('Taking screenshot...');
+    await page.screenshot({ path: './cost-debug.png', fullPage: true });
+
+    console.log('Done - check cost-debug.png');
+    await browser.close();
+})();

local_deep_research-0.6.0/src/local_deep_research/__version__.py

@@ -1 +0,0 @@
-__version__ = "0.6.0"