build based on 7adec40

dev/.documenter-siteinfo.json

@@ -1 +1 @@
-{"documenter":{"julia_version":"1.10.2","generation_timestamp":"2024-04-04T18:44:36","documenter_version":"1.3.0"}}
+{"documenter":{"julia_version":"1.10.2","generation_timestamp":"2024-04-25T14:03:56","documenter_version":"1.4.0"}}

dev/citation/index.html

@@ -11,4 +11,4 @@
   title = {Mill.jl framework: a flexible library for (hierarchical) multi-instance learning},
   url = {https://github.com/CTUAvastLab/Mill.jl},
   version = {...},
-}</code></pre></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../references/">« References</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.3.0 on <span class="colophon-date" title="Thursday 4 April 2024 18:44">Thursday 4 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>
+}</code></pre></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../references/">« References</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.4.0 on <span class="colophon-date" title="Thursday 25 April 2024 14:03">Thursday 25 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>

dev/examples/dag/index.html

@@ -28,4 +28,4 @@
 end
 
 millneighbors!(cache, g::DagGraph, model::DagModel, i::Int) = millneighbors!(cache, g, model, inneighbors(g.g, i))
-ChainRulesCore.@non_differentiable inneighbors(g, i)</code></pre><p>Note that this recursive approach is not the most efficient way to implement this. It would be better to spent a little time with graphs to identify sets of vertices that can be processed in parallel and for which all ancestors are known. But this was a fun little exercise.</p></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../gnn/">« GNNs in 16 lines</a><a class="docs-footer-nextpage" href="../jsons/">Processing JSONs »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.3.0 on <span class="colophon-date" title="Thursday 4 April 2024 18:44">Thursday 4 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>
+ChainRulesCore.@non_differentiable inneighbors(g, i)</code></pre><p>Note that this recursive approach is not the most efficient way to implement this. It would be better to spent a little time with graphs to identify sets of vertices that can be processed in parallel and for which all ancestors are known. But this was a fun little exercise.</p></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../gnn/">« GNNs in 16 lines</a><a class="docs-footer-nextpage" href="../jsons/">Processing JSONs »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.4.0 on <span class="colophon-date" title="Thursday 25 April 2024 14:03">Thursday 25 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>

dev/examples/gnn/index.html

@@ -26,7 +26,7 @@
     m::R
 end
 
-Flux.@functor GNN
+Flux.@layer :ignore GNN
 
 function mpstep(m::GNN, U, bags, n)
     n == 0 && return(U)
@@ -47,4 +47,4 @@
   4.619313f-6
   7.2975877f-6
  -7.123695f-6
-  4.521831f-6</code><br/><code class="language-julia-repl hljs" style="display:block;">julia&gt; gradient(m -&gt; m(g, X, 5) |&gt; sum, gnn)</code><code class="nohighlight hljs ansi" style="display:block;">((lift = (m = (layers = ((weight = Float32[-6.2311537f-6 -2.1438216f-6 6.7310534f-6; 1.0451618f-5 7.1090876f-6 -1.3662788f-5; -1.6941667f-7 1.6369462f-6 4.177652f-7; -9.1106847f-7 -4.3300406f-6 2.0227392f-6], bias = Float32[-3.3515346f-6, 7.861684f-6, 1.3820858f-7, -2.710481f-6], σ = nothing), (weight = Float32[-4.184251f-7 3.314137f-7 2.367027f-6 -1.021465f-7; -9.440357f-7 3.5489126f-8 2.832196f-6 -3.7543262f-7; -1.6309683f-7 -8.3873f-6 -2.1535452f-5 -2.7408876f-6; 1.2563737f-6 2.3163536f-6 4.6873906f-6 8.4355537f-7], bias = Float32[4.614298f-7, -4.460827f-7, -9.743454f-6, 4.1144303f-6], σ = nothing)),),), mp = (im = (m = (layers = ((weight = Float32[-3.5749432f-5 5.658251f-5 -4.8668713f-5 -1.0405847f-6; -1.0225553f-7 -2.45716f-7 2.6246704f-7 3.5376317f-7; 1.7279434f-7 7.3747503f-7 -7.103967f-7 -1.2383186f-6; 4.7143752f-5 -6.0898008f-5 5.536263f-5 1.132893f-6], bias = Float32[0.08966108, 1.5704284f-6, -2.621086f-6, -0.24559176], σ = nothing), (weight = Float32[3.2528795f-5 -2.4235618f-7 -6.626251f-7 1.0615856f-5; -9.3738716f-5 -3.7263037f-9 1.5493274f-8 -2.2950268f-5; 2.6827327f-5 -3.4215883f-8 -2.8393325f-8 5.9925787f-6; 0.0001342858 5.4141553f-9 3.6650327f-9 3.307389f-5], bias = Float32[-0.030241212, -0.30532664, 0.11367432, 0.44302276], σ = nothing)),),), a = (fs = ((ψ = Float32[0.0, 0.0, 0.0, 0.0],), (ψ = Float32[0.0, 0.0, 0.0, 0.0],)),), bm = (layers = ((weight = Float32[-3.389882f-8 -7.7773414f-8 … -3.5532235f-8 -2.0217596f-8; 8.378886f-5 1.2094616f-5 … -1.2023729f-6 3.9675968f-5; 0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0], bias = Float32[-3.7005095f-6, 0.4548328, 0.0, 0.0], σ = nothing), (weight = Float32[-6.556422f-8 -4.7089998f-5 0.0 0.0; -1.09523164f-7 -6.5535f-5 0.0 0.0; 1.5229857f-7 -0.00016093586 0.0 0.0; -9.035933f-9 4.2959266f-5 0.0 0.0], bias = Float32[-1.2480756, -0.87882626, -0.98240983, -0.14727944], σ = nothing)),)), m = (layers = ((weight = Float32[0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0; -9.626626f-6 1.0236514f-5 … -6.4234155f-6 -7.49814f-7; 0.0 0.0 … 0.0 0.0], bias = Float32[0.0, -0.0, 0.97359836, 0.0], σ = nothing), (weight = Float32[0.0 0.0 9.567639f-6 0.0; 0.0 0.0 9.567639f-6 0.0; 0.0 0.0 9.567639f-6 0.0; 0.0 0.0 9.567639f-6 0.0], bias = Fill(1.0f0, 4), σ = nothing)),)),)</code></pre><p>The above implementation is surprisingly general, as it supports an arbitrarily rich description of vertices. For simplicity, we used only vectors in <code>X</code>, however, any <code>Mill</code> hierarchy is applicable.</p><p>To put different weights on edges, one can use <a href="../../manual/aggregation/#Weighted-aggregation">Weighted aggregation</a>.</p></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../musk/musk/">« Musk</a><a class="docs-footer-nextpage" href="../dag/">DAGs »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.3.0 on <span class="colophon-date" title="Thursday 4 April 2024 18:44">Thursday 4 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>
+  4.521831f-6</code><br/><code class="language-julia-repl hljs" style="display:block;">julia&gt; gradient(m -&gt; m(g, X, 5) |&gt; sum, gnn)</code><code class="nohighlight hljs ansi" style="display:block;">((lift = (m = (layers = ((weight = Float32[-6.2311537f-6 -2.1438216f-6 6.7310534f-6; 1.0451618f-5 7.1090876f-6 -1.3662788f-5; -1.6941667f-7 1.6369462f-6 4.177652f-7; -9.1106847f-7 -4.3300406f-6 2.0227392f-6], bias = Float32[-3.3515346f-6, 7.861684f-6, 1.3820858f-7, -2.710481f-6], σ = nothing), (weight = Float32[-4.184251f-7 3.314137f-7 2.367027f-6 -1.021465f-7; -9.440357f-7 3.5489126f-8 2.832196f-6 -3.7543262f-7; -1.6309683f-7 -8.3873f-6 -2.1535452f-5 -2.7408876f-6; 1.2563737f-6 2.3163536f-6 4.6873906f-6 8.4355537f-7], bias = Float32[4.614298f-7, -4.460827f-7, -9.743454f-6, 4.1144303f-6], σ = nothing)),),), mp = (im = (m = (layers = ((weight = Float32[-3.5749432f-5 5.658251f-5 -4.8668713f-5 -1.0405847f-6; -1.0225553f-7 -2.45716f-7 2.6246704f-7 3.5376317f-7; 1.7279434f-7 7.3747503f-7 -7.103967f-7 -1.2383186f-6; 4.7143752f-5 -6.0898008f-5 5.536263f-5 1.132893f-6], bias = Float32[0.08966108, 1.5704284f-6, -2.621086f-6, -0.24559176], σ = nothing), (weight = Float32[3.2528795f-5 -2.4235618f-7 -6.626251f-7 1.0615856f-5; -9.3738716f-5 -3.7263037f-9 1.5493274f-8 -2.2950268f-5; 2.6827327f-5 -3.4215883f-8 -2.8393325f-8 5.9925787f-6; 0.0001342858 5.4141553f-9 3.6650327f-9 3.307389f-5], bias = Float32[-0.030241212, -0.30532664, 0.11367432, 0.44302276], σ = nothing)),),), a = (fs = ((ψ = Float32[0.0, 0.0, 0.0, 0.0],), (ψ = Float32[0.0, 0.0, 0.0, 0.0],)),), bm = (layers = ((weight = Float32[-3.389882f-8 -7.7773414f-8 … -3.5532235f-8 -2.0217596f-8; 8.378886f-5 1.2094616f-5 … -1.2023729f-6 3.9675968f-5; 0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0], bias = Float32[-3.7005095f-6, 0.4548328, 0.0, 0.0], σ = nothing), (weight = Float32[-6.556422f-8 -4.7089998f-5 0.0 0.0; -1.09523164f-7 -6.5535f-5 0.0 0.0; 1.5229857f-7 -0.00016093586 0.0 0.0; -9.035933f-9 4.2959266f-5 0.0 0.0], bias = Float32[-1.2480756, -0.87882626, -0.98240983, -0.14727944], σ = nothing)),)), m = (layers = ((weight = Float32[0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0; -9.626626f-6 1.0236514f-5 … -6.4234155f-6 -7.49814f-7; 0.0 0.0 … 0.0 0.0], bias = Float32[0.0, -0.0, 0.97359836, 0.0], σ = nothing), (weight = Float32[0.0 0.0 9.567639f-6 0.0; 0.0 0.0 9.567639f-6 0.0; 0.0 0.0 9.567639f-6 0.0; 0.0 0.0 9.567639f-6 0.0], bias = Fill(1.0f0, 4), σ = nothing)),)),)</code></pre><p>The above implementation is surprisingly general, as it supports an arbitrarily rich description of vertices. For simplicity, we used only vectors in <code>X</code>, however, any <code>Mill</code> hierarchy is applicable.</p><p>To put different weights on edges, one can use <a href="../../manual/aggregation/#Weighted-aggregation">Weighted aggregation</a>.</p></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../musk/musk/">« Musk</a><a class="docs-footer-nextpage" href="../dag/">DAGs »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.4.0 on <span class="colophon-date" title="Thursday 25 April 2024 14:03">Thursday 25 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>

dev/examples/jsons/index.html

@@ -6,4 +6,4 @@
         <img class="display-dark-only"
             src="https://raw.githubusercontent.com/CTUAvastLab/JsonGrinder.jl/master/docs/src/assets/logo-dark.svg" alt="JsonGrinder.jl logo"/>
     </a>
-</p><h1 id="Processing-JSONs"><a class="docs-heading-anchor" href="#Processing-JSONs">Processing JSONs</a><a id="Processing-JSONs-1"></a><a class="docs-heading-anchor-permalink" href="#Processing-JSONs" title="Permalink"></a></h1><p>Processing JSONs is actually one of the main motivations for building <a href="https://github.com/CTUAvastLab/Mill.jl"><code>Mill.jl</code></a>. As a matter of fact, with <code>Mill</code> one is now able to process a set of valid JSON documents that follow the same meta schema. <a href="https://github.com/CTUAvastLab/JsonGrinder.jl"><code>JsonGrinder.jl</code></a> is a library that helps with infering the schema and other steps in the pipeline. For some examples, please refer to its <a href="https://CTUAvastLab.github.io/JsonGrinder.jl/stable">documentation</a>.</p></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../dag/">« DAGs</a><a class="docs-footer-nextpage" href="../../tools/hierarchical/">HierarchicalUtils.jl »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.3.0 on <span class="colophon-date" title="Thursday 4 April 2024 18:44">Thursday 4 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>
+</p><h1 id="Processing-JSONs"><a class="docs-heading-anchor" href="#Processing-JSONs">Processing JSONs</a><a id="Processing-JSONs-1"></a><a class="docs-heading-anchor-permalink" href="#Processing-JSONs" title="Permalink"></a></h1><p>Processing JSONs is actually one of the main motivations for building <a href="https://github.com/CTUAvastLab/Mill.jl"><code>Mill.jl</code></a>. As a matter of fact, with <code>Mill</code> one is now able to process a set of valid JSON documents that follow the same meta schema. <a href="https://github.com/CTUAvastLab/JsonGrinder.jl"><code>JsonGrinder.jl</code></a> is a library that helps with infering the schema and other steps in the pipeline. For some examples, please refer to its <a href="https://CTUAvastLab.github.io/JsonGrinder.jl/stable">documentation</a>.</p></article><nav class="docs-footer"><a class="docs-footer-prevpage" href="../dag/">« DAGs</a><a class="docs-footer-nextpage" href="../../tools/hierarchical/">HierarchicalUtils.jl »</a><div class="flexbox-break"></div><p class="footer-message">Powered by <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> and the <a href="https://julialang.org/">Julia Programming Language</a>.</p></nav></div><div class="modal" id="documenter-settings"><div class="modal-background"></div><div class="modal-card"><header class="modal-card-head"><p class="modal-card-title">Settings</p><button class="delete"></button></header><section class="modal-card-body"><p><label class="label">Theme</label><div class="select"><select id="documenter-themepicker"><option value="auto">Automatic (OS)</option><option value="documenter-light">documenter-light</option><option value="documenter-dark">documenter-dark</option></select></div></p><hr/><p>This document was generated with <a href="https://github.com/JuliaDocs/Documenter.jl">Documenter.jl</a> version 1.4.0 on <span class="colophon-date" title="Thursday 25 April 2024 14:03">Thursday 25 April 2024</span>. Using Julia version 1.10.2.</p></section><footer class="modal-card-foot"></footer></div></div></div></body></html>