diff --git a/diagrams/diagrams/pipeline-phase2.svg b/diagrams/diagrams/pipeline-phase2.svg
deleted file mode 100644
index a1bd146..0000000
--- a/diagrams/diagrams/pipeline-phase2.svg
+++ /dev/null
@@ -1,97 +0,0 @@
-<?xml version="1.0" encoding="UTF-8" standalone="no"?><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" contentScriptType="application/ecmascript" contentStyleType="text/css" height="668px" preserveAspectRatio="none" style="width:858px;height:668px;" version="1.1" viewBox="0 0 858 668" width="858px" zoomAndPan="magnify"><defs><filter height="300%" id="f8n6pzrpnkj0p" width="300%" x="-1" y="-1"><feGaussianBlur result="blurOut" stdDeviation="2.0"/><feColorMatrix in="blurOut" result="blurOut2" type="matrix" values="0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .4 0"/><feOffset dx="4.0" dy="4.0" in="blurOut2" result="blurOut3"/><feBlend in="SourceGraphic" in2="blurOut3" mode="normal"/></filter></defs><g><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="258" x="299.5" y="12.0669">Phase 2 -- Chroma Vector Store Pipeline</text><!--MD5=[d43075e66b9401f5f099c90dff66d346]
-cluster Ingestion (ingest.py)--><polygon fill="#FFFFFF" filter="url(#f8n6pzrpnkj0p)" points="22,39.6328,182,39.6328,189,60.7656,192,60.7656,192,497.1328,22,497.1328,22,39.6328" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="22" x2="189" y1="60.7656" y2="60.7656"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="154" x="26" y="53.6997">Ingestion (ingest.py)</text><!--MD5=[eee07272e247f727b0c9cb5ea2910cc3]
-cluster Search (search.py)--><polygon fill="#FFFFFF" filter="url(#f8n6pzrpnkj0p)" points="216,173.6328,361,173.6328,368,194.7656,438,194.7656,438,656.1328,216,656.1328,216,173.6328" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="216" x2="368" y1="194.7656" y2="194.7656"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="139" x="220" y="187.6997">Search (search.py)</text><!--MD5=[57a5b41ed6b3ba96fe9b4a740eb3fdc7]
-entity docs--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="70" x="73" y="91.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="68" y="96.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="68" y="116.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="50" x="83" y="113.6997">.txt files</text><!--MD5=[6fa1cc94bec0cf5ab264efe33d75d3c3]
-entity load--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="136" x="40" y="225.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="35" y="230.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="35" y="250.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="116" x="50" y="247.6997">load_documents()</text><!--MD5=[ff858b2fa29ed35ae3d882895b3a15f0]
-entity chunk--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="100" x="59" y="321.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="54" y="326.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="54" y="346.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="80" x="69" y="343.6997">chunk_text()</text><!--MD5=[1a7a766028402bec58ce31452d7a7648]
-entity embed--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="127" x="48.5" y="446.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="43.5" y="451.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="43.5" y="471.2656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="107" x="58.5" y="468.1997">embed_chunks()</text><!--MD5=[a096c1983bcf686c95ea726f28f7a0ed]
-entity query--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="58" x="334" y="225.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="329" y="230.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="329" y="250.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="38" x="344" y="247.6997">Query</text><!--MD5=[81edeb3f3a35c1dab7f46e9decfdfebe]
-entity embedq--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="118" x="304" y="321.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="299" y="326.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="299" y="346.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="98" x="314" y="343.6997">embed_query()</text><!--MD5=[8fb73a19f8d09b32c502350ebb5d2be5]
-entity cquery--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="130" x="260" y="446.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="255" y="451.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="255" y="471.2656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="110" x="270" y="468.1997">collection.query()</text><!--MD5=[b5cce2cdbd787adcd8934bb88f8254b3]
-entity results--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="50.2656" style="stroke: #5577AA; stroke-width: 1.5;" width="190" x="232" y="590.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="227" y="595.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="227" y="630.3984"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="86" x="242" y="612.1997">Top-K Results</text><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="170" x="242" y="627.3325">(text + source + distance)</text><!--MD5=[d85c0a1eb646255ad28a1eb3e90eaa32]
-entity openai--><path d="M663.835,442.4894 C667.1277,435.2659 672.3414,436.1026 676.6173,441.4931 C680.0176,435.0792 685.7228,435.5954 689.0973,441.5192 C691.6705,436.9846 696.1511,436.5453 699.0845,441.1244 C702.7275,434.8774 708.3475,432.647 713.2868,439.7448 C716.5627,434.4833 722.0088,434.6406 724.5873,440.4563 C727.487,434.6172 731.4132,435.2085 735.3898,439.2829 C738.5809,431.7226 743.7931,431.3201 749.0607,437.1652 C752.9629,433.2862 757.0274,433.6585 759.4121,438.8895 C763.0709,432.484 769.0669,433.1716 771.662,439.8829 C775.101,434.1339 781.5417,433.03 784.9154,439.9511 C787.8691,434.6344 792.9245,435.6335 795.3463,440.4961 C800.1533,436.0428 804.3976,436.0124 807.2952,442.5685 C811.7915,435.1647 815.8976,436.1596 820.8024,442.1567 C824.0394,437.2752 829.2042,436.3206 832.5035,442.0882 C838.6991,443.41 841.1537,448.8893 836.0195,453.655 C842.5332,454.8929 844.7109,457.9631 842.3017,464.4121 C846.637,470.97 843.7827,475.527 836.2549,475.7682 C839.9568,481.3819 838.6597,485.5913 832.1783,487.7365 C828.7109,493.8966 823.9772,493.6881 819.7924,488.5208 C815.9591,494.4228 811.7036,494.2377 807.6621,488.7424 C805.6099,495.3442 799.7881,496.7766 795.0442,491.6344 C792.0953,496.906 788.2131,496.5461 785.1212,491.7758 C782.9264,497.8229 777.5868,499.4607 773.0484,494.3002 C769.5195,501.0609 764.8674,501.6812 760.0009,495.6662 C756.23,501.3552 752.3189,501.3978 748.9306,495.3066 C744.55,501.9554 739.8568,502.2077 735.6966,495.0946 C731.0645,499.8501 725.8767,499.4683 723.6526,492.5997 C720.7068,497.941 715.0143,498.0348 712.0471,492.6227 C707.5735,497.7266 702.9539,499.1638 699.1543,491.919 C695.1953,496.5972 691.1095,495.38 689.1504,489.9415 C685.0524,495.8668 679.5274,495.6227 676.2618,489.081 C672.0738,494.2949 666.8131,494.5433 663.7083,487.9311 C655.9358,484.8511 656.5839,480.5434 660.9809,474.9136 C654.3407,474.2883 653.4529,470.4153 655.8453,465.075 C651.93,459.4246 653.5013,454.3598 660.9438,453.8839 C657.0052,448.7166 657.3805,445.0196 663.835,442.4894 " fill="#F0FFF0" style="stroke: #448844; stroke-width: 1.5;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="670" y="460.6997">OpenAI API</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="154" x="670" y="475.8325">text-embedding-3-small</text><!--MD5=[ba0e3141185c9f6ac85d8ff32b47afed]
-entity chroma--><path d="M103,595.6328 C103,585.6328 150,585.6328 150,585.6328 C150,585.6328 197,585.6328 197,595.6328 L197,634.8984 C197,644.8984 150,644.8984 150,644.8984 C150,644.8984 103,644.8984 103,634.8984 L103,595.6328 " fill="#FFF8E7" filter="url(#f8n6pzrpnkj0p)" style="stroke: #CC9900; stroke-width: 1.5;"/><path d="M103,595.6328 C103,605.6328 150,605.6328 150,605.6328 C150,605.6328 197,605.6328 197,595.6328 " fill="none" style="stroke: #CC9900; stroke-width: 1.5;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="74" x="113" y="621.6997">Chroma DB</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="74" x="113" y="636.8325">(persistent)</text><!--MD5=[34aeeab8a23e75e12ac7d0294eef783a]
-link docs to load--><path d="M108,126.8528 C108,150.6528 108,194.3328 108,220.5528 " fill="none" id="docs-&gt;load" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="108,225.5528,112,216.5528,108,220.5528,104,216.5528,108,225.5528" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[197124bfd89600884ad83f5bdb81fe2d]
-link load to chunk--><path d="M108.18,260.9428 C108.34,276.4428 108.59,299.5528 108.77,316.4428 " fill="none" id="load-&gt;chunk" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="108.82,321.4728,112.7169,312.4277,108.7629,316.4731,104.7174,312.5191,108.82,321.4728" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[f719c65e3cd5877eb9f397b030baf6d1]
-link chunk to embed--><path d="M109.4,356.6528 C109.94,378.3928 110.87,416.6528 111.46,440.7628 " fill="none" id="chunk-&gt;embed" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="111.59,445.9828,115.3833,436.8938,111.4757,440.9841,107.3854,437.0766,111.59,445.9828" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[14e589a052329e665bcf72ada14e2408]
-link embed to openai--><path d="M166.08,446.0928 C182.02,441.4528 199.59,436.8428 216,433.6328 C304.6,416.2828 532.95,373.5928 620.5,395.6328 C648.27,402.6228 676.58,417.0228 699.36,430.7828 " fill="none" id="embed-&gt;openai" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="703.84,433.5228,698.2517,425.4129,699.5754,430.9127,694.0755,432.2363,703.84,433.5228" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="45" x="574.5" y="399.6997">API call</text><!--MD5=[fad15655625d41508c0bd675faf1956d]
-link openai to embed--><path d="M684.82,433.6228 C651.11,419.0028 608.22,403.0128 568,395.6328 C422.88,368.9928 250.33,415.9728 165.09,444.3628 " fill="none" id="openai-&gt;embed" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="160.1,446.0428,169.9073,446.9471,164.8362,444.4402,167.3431,439.3692,160.1,446.0428" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="112" x="455" y="399.6997">1536-dim vectors</text><!--MD5=[b1ec60cbada152cf55f28d4273f8e0c5]
-link embed to chroma--><path d="M77.69,481.1428 C47.07,498.2528 9.67,526.5228 28,555.6328 C43.43,580.1528 72.31,594.6228 98.09,603.0328 " fill="none" id="embed-&gt;chroma" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="102.87,604.5328,95.5016,597.9977,98.1042,603.0203,93.0817,605.6229,102.87,604.5328" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="44" y="536.6997">upsert with</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="100" x="29" y="551.8325">ids + metadata</text><!--MD5=[8efdc581dcc4934b230519e9a1c62620]
-link query to embedq--><path d="M363,260.9428 C363,276.4428 363,299.5528 363,316.4428 " fill="none" id="query-&gt;embedq" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="363,321.4728,367,312.4728,363,316.4728,359,312.4728,363,321.4728" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[07d15e8df06b7fbd510547f2864e4d80]
-link embedq to openai--><path d="M422.04,343.1928 C523.6,349.0528 722.58,363.3828 744,386.6328 C754.12,397.6228 755.88,413.7328 754.7,428.3028 " fill="none" id="embedq-&gt;openai" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="754.14,433.5828,759.0691,425.0562,754.6686,428.6108,751.114,424.2104,754.14,433.5828" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="45" x="753" y="399.6997">API call</text><!--MD5=[3a9578d9bd486bb01d5041cfdd4d7983]
-link openai to embedq--><path d="M701.07,433.5828 C669.03,413.5128 630.55,389.8728 622,386.6328 C558.26,362.5028 480.85,350.7028 427.35,345.0628 " fill="none" id="openai-&gt;embedq" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="422.25,344.5328,430.7906,349.4377,427.2235,345.0473,431.6138,341.4801,422.25,344.5328" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="80" x="651" y="399.6997">query vector</text><!--MD5=[5c13e55db4cb974e2ddc415d002712d7]
-link embedq to cquery--><path d="M357.88,356.6528 C351.1,378.4828 339.16,416.9728 331.69,441.0728 " fill="none" id="embedq-&gt;cquery" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="330.17,445.9828,336.6446,438.5612,331.6434,441.2048,328.9998,436.2037,330.17,445.9828" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[f67d8fe0aaaf6a85deca3a91259f7f85]
-link cquery to chroma--><path d="M319.64,481.4628 C312.52,501.4428 298.2,534.7128 276,555.6328 C272.93,558.5328 235.47,575.7828 202.2,590.8028 " fill="none" id="cquery-&gt;chroma" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="197.32,593.0028,207.1687,592.9457,201.8772,590.9456,203.8772,585.6542,197.32,593.0028" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="76" x="301" y="544.1997">ANN search</text><!--MD5=[f593972c37969fbb6ae91af54d31458c]
-link chroma to cquery--><path d="M148.07,585.3728 C148.3,565.8828 151.79,540.6628 166,523.6328 C178.27,508.9228 219.17,493.9128 256.41,482.7228 " fill="none" id="chroma-&gt;cquery" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="261.4,481.2328,251.6364,479.9396,256.6039,482.646,253.8975,487.6134,261.4,481.2328" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="100" x="167" y="544.1997">nearest vectors</text><!--MD5=[a38503b39590097c9f1ad789f279155b]
-link cquery to results--><path d="M354.59,481.3528 C369.38,491.4828 385.82,505.8828 394,523.6328 C399.95,536.5528 399.88,542.6828 394,555.6328 C388.59,567.5428 379.41,577.9828 369.56,586.6128 " fill="none" id="cquery-&gt;results" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="365.61,589.9628,375.0638,587.2013,369.4264,586.7325,369.8952,581.0951,365.61,589.9628" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[583f7c257adc767560261a1a38599b51]
-@startuml pipeline-phase2
-
-skinparam backgroundColor #FFFFFF
-skinparam defaultFontName Arial
-skinparam defaultFontSize 13
-skinparam ArrowColor #555555
-skinparam componentStyle rectangle
-
-skinparam component {
-  BackgroundColor #EEF3FB
-  BorderColor #5577AA
-  FontColor #222222
-}
-
-skinparam database {
-  BackgroundColor #FFF8E7
-  BorderColor #CC9900
-}
-
-skinparam cloud {
-  BackgroundColor #F0FFF0
-  BorderColor #448844
-}
-
-title Phase 2 - - Chroma Vector Store Pipeline
-
-package "Ingestion (ingest.py)" {
-  [.txt files] as docs
-  [load_documents()] as load
-  [chunk_text()] as chunk
-  [embed_chunks()] as embed
-}
-
-cloud "OpenAI API\ntext-embedding-3-small" as openai
-
-database "Chroma DB\n(persistent)" as chroma
-
-package "Search (search.py)" {
-  [Query] as query
-  [embed_query()] as embedq
-  [collection.query()] as cquery
-  [Top-K Results\n(text + source + distance)] as results
-}
-
-docs - -> load
-load - -> chunk
-chunk - -> embed
-embed - -> openai : API call
-openai - -> embed : 1536-dim vectors
-embed - -> chroma : upsert with\nids + metadata
-
-query - -> embedq
-embedq - -> openai : API call
-openai - -> embedq : query vector
-embedq - -> cquery
-cquery - -> chroma : ANN search
-chroma - -> cquery : nearest vectors
-cquery - -> results
-
-@enduml
-
-PlantUML version 1.2020.02(Sun Mar 01 10:22:07 UTC 2020)
-(GPL source distribution)
-Java Runtime: OpenJDK Runtime Environment
-JVM: OpenJDK 64-Bit Server VM
-Java Version: 17.0.19+10
-Operating System: Linux
-Default Encoding: UTF-8
-Language: en
-Country: null
---></g></svg>
\ No newline at end of file
diff --git a/diagrams/docs/pipeline-phase2.svg b/diagrams/docs/pipeline-phase2.svg
deleted file mode 100644
index a1bd146..0000000
--- a/diagrams/docs/pipeline-phase2.svg
+++ /dev/null
@@ -1,97 +0,0 @@
-<?xml version="1.0" encoding="UTF-8" standalone="no"?><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" contentScriptType="application/ecmascript" contentStyleType="text/css" height="668px" preserveAspectRatio="none" style="width:858px;height:668px;" version="1.1" viewBox="0 0 858 668" width="858px" zoomAndPan="magnify"><defs><filter height="300%" id="f8n6pzrpnkj0p" width="300%" x="-1" y="-1"><feGaussianBlur result="blurOut" stdDeviation="2.0"/><feColorMatrix in="blurOut" result="blurOut2" type="matrix" values="0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .4 0"/><feOffset dx="4.0" dy="4.0" in="blurOut2" result="blurOut3"/><feBlend in="SourceGraphic" in2="blurOut3" mode="normal"/></filter></defs><g><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="258" x="299.5" y="12.0669">Phase 2 -- Chroma Vector Store Pipeline</text><!--MD5=[d43075e66b9401f5f099c90dff66d346]
-cluster Ingestion (ingest.py)--><polygon fill="#FFFFFF" filter="url(#f8n6pzrpnkj0p)" points="22,39.6328,182,39.6328,189,60.7656,192,60.7656,192,497.1328,22,497.1328,22,39.6328" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="22" x2="189" y1="60.7656" y2="60.7656"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="154" x="26" y="53.6997">Ingestion (ingest.py)</text><!--MD5=[eee07272e247f727b0c9cb5ea2910cc3]
-cluster Search (search.py)--><polygon fill="#FFFFFF" filter="url(#f8n6pzrpnkj0p)" points="216,173.6328,361,173.6328,368,194.7656,438,194.7656,438,656.1328,216,656.1328,216,173.6328" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="216" x2="368" y1="194.7656" y2="194.7656"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="139" x="220" y="187.6997">Search (search.py)</text><!--MD5=[57a5b41ed6b3ba96fe9b4a740eb3fdc7]
-entity docs--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="70" x="73" y="91.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="68" y="96.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="68" y="116.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="50" x="83" y="113.6997">.txt files</text><!--MD5=[6fa1cc94bec0cf5ab264efe33d75d3c3]
-entity load--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="136" x="40" y="225.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="35" y="230.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="35" y="250.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="116" x="50" y="247.6997">load_documents()</text><!--MD5=[ff858b2fa29ed35ae3d882895b3a15f0]
-entity chunk--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="100" x="59" y="321.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="54" y="326.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="54" y="346.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="80" x="69" y="343.6997">chunk_text()</text><!--MD5=[1a7a766028402bec58ce31452d7a7648]
-entity embed--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="127" x="48.5" y="446.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="43.5" y="451.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="43.5" y="471.2656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="107" x="58.5" y="468.1997">embed_chunks()</text><!--MD5=[a096c1983bcf686c95ea726f28f7a0ed]
-entity query--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="58" x="334" y="225.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="329" y="230.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="329" y="250.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="38" x="344" y="247.6997">Query</text><!--MD5=[81edeb3f3a35c1dab7f46e9decfdfebe]
-entity embedq--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="118" x="304" y="321.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="299" y="326.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="299" y="346.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="98" x="314" y="343.6997">embed_query()</text><!--MD5=[8fb73a19f8d09b32c502350ebb5d2be5]
-entity cquery--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="130" x="260" y="446.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="255" y="451.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="255" y="471.2656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="110" x="270" y="468.1997">collection.query()</text><!--MD5=[b5cce2cdbd787adcd8934bb88f8254b3]
-entity results--><rect fill="#EEF3FB" filter="url(#f8n6pzrpnkj0p)" height="50.2656" style="stroke: #5577AA; stroke-width: 1.5;" width="190" x="232" y="590.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="227" y="595.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="227" y="630.3984"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="86" x="242" y="612.1997">Top-K Results</text><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="170" x="242" y="627.3325">(text + source + distance)</text><!--MD5=[d85c0a1eb646255ad28a1eb3e90eaa32]
-entity openai--><path d="M663.835,442.4894 C667.1277,435.2659 672.3414,436.1026 676.6173,441.4931 C680.0176,435.0792 685.7228,435.5954 689.0973,441.5192 C691.6705,436.9846 696.1511,436.5453 699.0845,441.1244 C702.7275,434.8774 708.3475,432.647 713.2868,439.7448 C716.5627,434.4833 722.0088,434.6406 724.5873,440.4563 C727.487,434.6172 731.4132,435.2085 735.3898,439.2829 C738.5809,431.7226 743.7931,431.3201 749.0607,437.1652 C752.9629,433.2862 757.0274,433.6585 759.4121,438.8895 C763.0709,432.484 769.0669,433.1716 771.662,439.8829 C775.101,434.1339 781.5417,433.03 784.9154,439.9511 C787.8691,434.6344 792.9245,435.6335 795.3463,440.4961 C800.1533,436.0428 804.3976,436.0124 807.2952,442.5685 C811.7915,435.1647 815.8976,436.1596 820.8024,442.1567 C824.0394,437.2752 829.2042,436.3206 832.5035,442.0882 C838.6991,443.41 841.1537,448.8893 836.0195,453.655 C842.5332,454.8929 844.7109,457.9631 842.3017,464.4121 C846.637,470.97 843.7827,475.527 836.2549,475.7682 C839.9568,481.3819 838.6597,485.5913 832.1783,487.7365 C828.7109,493.8966 823.9772,493.6881 819.7924,488.5208 C815.9591,494.4228 811.7036,494.2377 807.6621,488.7424 C805.6099,495.3442 799.7881,496.7766 795.0442,491.6344 C792.0953,496.906 788.2131,496.5461 785.1212,491.7758 C782.9264,497.8229 777.5868,499.4607 773.0484,494.3002 C769.5195,501.0609 764.8674,501.6812 760.0009,495.6662 C756.23,501.3552 752.3189,501.3978 748.9306,495.3066 C744.55,501.9554 739.8568,502.2077 735.6966,495.0946 C731.0645,499.8501 725.8767,499.4683 723.6526,492.5997 C720.7068,497.941 715.0143,498.0348 712.0471,492.6227 C707.5735,497.7266 702.9539,499.1638 699.1543,491.919 C695.1953,496.5972 691.1095,495.38 689.1504,489.9415 C685.0524,495.8668 679.5274,495.6227 676.2618,489.081 C672.0738,494.2949 666.8131,494.5433 663.7083,487.9311 C655.9358,484.8511 656.5839,480.5434 660.9809,474.9136 C654.3407,474.2883 653.4529,470.4153 655.8453,465.075 C651.93,459.4246 653.5013,454.3598 660.9438,453.8839 C657.0052,448.7166 657.3805,445.0196 663.835,442.4894 " fill="#F0FFF0" style="stroke: #448844; stroke-width: 1.5;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="670" y="460.6997">OpenAI API</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="154" x="670" y="475.8325">text-embedding-3-small</text><!--MD5=[ba0e3141185c9f6ac85d8ff32b47afed]
-entity chroma--><path d="M103,595.6328 C103,585.6328 150,585.6328 150,585.6328 C150,585.6328 197,585.6328 197,595.6328 L197,634.8984 C197,644.8984 150,644.8984 150,644.8984 C150,644.8984 103,644.8984 103,634.8984 L103,595.6328 " fill="#FFF8E7" filter="url(#f8n6pzrpnkj0p)" style="stroke: #CC9900; stroke-width: 1.5;"/><path d="M103,595.6328 C103,605.6328 150,605.6328 150,605.6328 C150,605.6328 197,605.6328 197,595.6328 " fill="none" style="stroke: #CC9900; stroke-width: 1.5;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="74" x="113" y="621.6997">Chroma DB</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="74" x="113" y="636.8325">(persistent)</text><!--MD5=[34aeeab8a23e75e12ac7d0294eef783a]
-link docs to load--><path d="M108,126.8528 C108,150.6528 108,194.3328 108,220.5528 " fill="none" id="docs-&gt;load" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="108,225.5528,112,216.5528,108,220.5528,104,216.5528,108,225.5528" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[197124bfd89600884ad83f5bdb81fe2d]
-link load to chunk--><path d="M108.18,260.9428 C108.34,276.4428 108.59,299.5528 108.77,316.4428 " fill="none" id="load-&gt;chunk" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="108.82,321.4728,112.7169,312.4277,108.7629,316.4731,104.7174,312.5191,108.82,321.4728" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[f719c65e3cd5877eb9f397b030baf6d1]
-link chunk to embed--><path d="M109.4,356.6528 C109.94,378.3928 110.87,416.6528 111.46,440.7628 " fill="none" id="chunk-&gt;embed" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="111.59,445.9828,115.3833,436.8938,111.4757,440.9841,107.3854,437.0766,111.59,445.9828" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[14e589a052329e665bcf72ada14e2408]
-link embed to openai--><path d="M166.08,446.0928 C182.02,441.4528 199.59,436.8428 216,433.6328 C304.6,416.2828 532.95,373.5928 620.5,395.6328 C648.27,402.6228 676.58,417.0228 699.36,430.7828 " fill="none" id="embed-&gt;openai" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="703.84,433.5228,698.2517,425.4129,699.5754,430.9127,694.0755,432.2363,703.84,433.5228" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="45" x="574.5" y="399.6997">API call</text><!--MD5=[fad15655625d41508c0bd675faf1956d]
-link openai to embed--><path d="M684.82,433.6228 C651.11,419.0028 608.22,403.0128 568,395.6328 C422.88,368.9928 250.33,415.9728 165.09,444.3628 " fill="none" id="openai-&gt;embed" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="160.1,446.0428,169.9073,446.9471,164.8362,444.4402,167.3431,439.3692,160.1,446.0428" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="112" x="455" y="399.6997">1536-dim vectors</text><!--MD5=[b1ec60cbada152cf55f28d4273f8e0c5]
-link embed to chroma--><path d="M77.69,481.1428 C47.07,498.2528 9.67,526.5228 28,555.6328 C43.43,580.1528 72.31,594.6228 98.09,603.0328 " fill="none" id="embed-&gt;chroma" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="102.87,604.5328,95.5016,597.9977,98.1042,603.0203,93.0817,605.6229,102.87,604.5328" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="44" y="536.6997">upsert with</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="100" x="29" y="551.8325">ids + metadata</text><!--MD5=[8efdc581dcc4934b230519e9a1c62620]
-link query to embedq--><path d="M363,260.9428 C363,276.4428 363,299.5528 363,316.4428 " fill="none" id="query-&gt;embedq" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="363,321.4728,367,312.4728,363,316.4728,359,312.4728,363,321.4728" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[07d15e8df06b7fbd510547f2864e4d80]
-link embedq to openai--><path d="M422.04,343.1928 C523.6,349.0528 722.58,363.3828 744,386.6328 C754.12,397.6228 755.88,413.7328 754.7,428.3028 " fill="none" id="embedq-&gt;openai" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="754.14,433.5828,759.0691,425.0562,754.6686,428.6108,751.114,424.2104,754.14,433.5828" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="45" x="753" y="399.6997">API call</text><!--MD5=[3a9578d9bd486bb01d5041cfdd4d7983]
-link openai to embedq--><path d="M701.07,433.5828 C669.03,413.5128 630.55,389.8728 622,386.6328 C558.26,362.5028 480.85,350.7028 427.35,345.0628 " fill="none" id="openai-&gt;embedq" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="422.25,344.5328,430.7906,349.4377,427.2235,345.0473,431.6138,341.4801,422.25,344.5328" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="80" x="651" y="399.6997">query vector</text><!--MD5=[5c13e55db4cb974e2ddc415d002712d7]
-link embedq to cquery--><path d="M357.88,356.6528 C351.1,378.4828 339.16,416.9728 331.69,441.0728 " fill="none" id="embedq-&gt;cquery" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="330.17,445.9828,336.6446,438.5612,331.6434,441.2048,328.9998,436.2037,330.17,445.9828" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[f67d8fe0aaaf6a85deca3a91259f7f85]
-link cquery to chroma--><path d="M319.64,481.4628 C312.52,501.4428 298.2,534.7128 276,555.6328 C272.93,558.5328 235.47,575.7828 202.2,590.8028 " fill="none" id="cquery-&gt;chroma" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="197.32,593.0028,207.1687,592.9457,201.8772,590.9456,203.8772,585.6542,197.32,593.0028" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="76" x="301" y="544.1997">ANN search</text><!--MD5=[f593972c37969fbb6ae91af54d31458c]
-link chroma to cquery--><path d="M148.07,585.3728 C148.3,565.8828 151.79,540.6628 166,523.6328 C178.27,508.9228 219.17,493.9128 256.41,482.7228 " fill="none" id="chroma-&gt;cquery" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="261.4,481.2328,251.6364,479.9396,256.6039,482.646,253.8975,487.6134,261.4,481.2328" style="stroke: #555555; stroke-width: 1.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="100" x="167" y="544.1997">nearest vectors</text><!--MD5=[a38503b39590097c9f1ad789f279155b]
-link cquery to results--><path d="M354.59,481.3528 C369.38,491.4828 385.82,505.8828 394,523.6328 C399.95,536.5528 399.88,542.6828 394,555.6328 C388.59,567.5428 379.41,577.9828 369.56,586.6128 " fill="none" id="cquery-&gt;results" style="stroke: #555555; stroke-width: 1.0;"/><polygon fill="#555555" points="365.61,589.9628,375.0638,587.2013,369.4264,586.7325,369.8952,581.0951,365.61,589.9628" style="stroke: #555555; stroke-width: 1.0;"/><!--MD5=[583f7c257adc767560261a1a38599b51]
-@startuml pipeline-phase2
-
-skinparam backgroundColor #FFFFFF
-skinparam defaultFontName Arial
-skinparam defaultFontSize 13
-skinparam ArrowColor #555555
-skinparam componentStyle rectangle
-
-skinparam component {
-  BackgroundColor #EEF3FB
-  BorderColor #5577AA
-  FontColor #222222
-}
-
-skinparam database {
-  BackgroundColor #FFF8E7
-  BorderColor #CC9900
-}
-
-skinparam cloud {
-  BackgroundColor #F0FFF0
-  BorderColor #448844
-}
-
-title Phase 2 - - Chroma Vector Store Pipeline
-
-package "Ingestion (ingest.py)" {
-  [.txt files] as docs
-  [load_documents()] as load
-  [chunk_text()] as chunk
-  [embed_chunks()] as embed
-}
-
-cloud "OpenAI API\ntext-embedding-3-small" as openai
-
-database "Chroma DB\n(persistent)" as chroma
-
-package "Search (search.py)" {
-  [Query] as query
-  [embed_query()] as embedq
-  [collection.query()] as cquery
-  [Top-K Results\n(text + source + distance)] as results
-}
-
-docs - -> load
-load - -> chunk
-chunk - -> embed
-embed - -> openai : API call
-openai - -> embed : 1536-dim vectors
-embed - -> chroma : upsert with\nids + metadata
-
-query - -> embedq
-embedq - -> openai : API call
-openai - -> embedq : query vector
-embedq - -> cquery
-cquery - -> chroma : ANN search
-chroma - -> cquery : nearest vectors
-cquery - -> results
-
-@enduml
-
-PlantUML version 1.2020.02(Sun Mar 01 10:22:07 UTC 2020)
-(GPL source distribution)
-Java Runtime: OpenJDK Runtime Environment
-JVM: OpenJDK 64-Bit Server VM
-Java Version: 17.0.19+10
-Operating System: Linux
-Default Encoding: UTF-8
-Language: en
-Country: null
---></g></svg>
\ No newline at end of file
diff --git a/diagrams/docs/pipeline-semantic-search.svg b/diagrams/docs/pipeline-semantic-search.svg
deleted file mode 100644
index cc43a4d..0000000
--- a/diagrams/docs/pipeline-semantic-search.svg
+++ /dev/null
@@ -1,80 +0,0 @@
-<?xml version="1.0" encoding="UTF-8" standalone="no"?><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" contentScriptType="application/ecmascript" contentStyleType="text/css" height="668px" preserveAspectRatio="none" style="width:425px;height:668px;background:#FAFAFA;" version="1.1" viewBox="0 0 425 668" width="425px" zoomAndPan="magnify"><defs><filter height="300%" id="f1lmzks9slht9h" width="300%" x="-1" y="-1"><feGaussianBlur result="blurOut" stdDeviation="2.0"/><feColorMatrix in="blurOut" result="blurOut2" type="matrix" values="0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .4 0"/><feOffset dx="4.0" dy="4.0" in="blurOut2" result="blurOut3"/><feBlend in="SourceGraphic" in2="blurOut3" mode="normal"/></filter></defs><g><!--MD5=[ff4c7cbfc86cbafc5088fc7dc3b82e86]
-cluster Indexing — embed.py--><polygon fill="#EEF2FF" filter="url(#f1lmzks9slht9h)" points="22,24,189,24,196,45.1328,218,45.1328,218,454,22,454,22,24" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="22" x2="196" y1="45.1328" y2="45.1328"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="161" x="26" y="38.0669">Indexing — embed.py</text><!--MD5=[b7354f1b6b1a4ce6b3292a1a387e6ba1]
-cluster Search — search.py--><polygon fill="#FDF4FF" filter="url(#f1lmzks9slht9h)" points="242,258.5,393,258.5,400,279.6328,403,279.6328,403,656,242,656,242,258.5" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="242" x2="400" y1="279.6328" y2="279.6328"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="145" x="246" y="272.5669">Search — search.py</text><!--MD5=[57a5b41ed6b3ba96fe9b4a740eb3fdc7]
-entity docs--><polygon fill="#FEFECE" filter="url(#f1lmzks9slht9h)" points="67,58,113,58,120,79,173,79,173,111.2656,67,111.2656,67,58" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="67" x2="120" y1="79" y2="79"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="76" x="77" y="98.0669">documents/</text><!--MD5=[ff858b2fa29ed35ae3d882895b3a15f0]
-entity chunk--><rect fill="#EFF6FF" filter="url(#f1lmzks9slht9h)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="164" x="38" y="171"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="48" y="193.0669">chunk_text</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="144" x="48" y="208.1997">300 words · 50 overlap</text><!--MD5=[1a7a766028402bec58ce31452d7a7648]
-entity embed--><rect fill="#EFF6FF" filter="url(#f1lmzks9slht9h)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="117" x="65.5" y="281"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="97" x="75.5" y="303.0669">embed_chunks</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="75.5" y="318.1997">OpenAI API</text><!--MD5=[0686720d71d32ea68f81164171d509ce]
-entity store--><path d="M67.5,404 C67.5,394 132,394 132,394 C132,394 196.5,394 196.5,404 L196.5,428.1328 C196.5,438.1328 132,438.1328 132,438.1328 C132,438.1328 67.5,438.1328 67.5,428.1328 L67.5,404 " fill="#FEF9C3" filter="url(#f1lmzks9slht9h)" style="stroke: #EAB308; stroke-width: 1.5;"/><path d="M67.5,404 C67.5,414 132,414 132,414 C132,414 196.5,414 196.5,404 " fill="none" style="stroke: #EAB308; stroke-width: 1.5;"/><text fill="#713F12" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="109" x="77.5" y="430.0669">embeddings.json</text><ellipse cx="320.8701" cy="306.2005" fill="#F0FDF4" filter="url(#f1lmzks9slht9h)" rx="29.8701" ry="13.7005" style="stroke: #22C55E; stroke-width: 1.5;"/><text fill="#14532D" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="38" x="301.8701" y="310.2669">Query</text><!--MD5=[81edeb3f3a35c1dab7f46e9decfdfebe]
-entity embedq--><rect fill="#EFF6FF" filter="url(#f1lmzks9slht9h)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="108" x="267" y="391"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="88" x="277" y="413.0669">embed_query</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="277" y="428.1997">OpenAI API</text><!--MD5=[3db17c234f6b9cf849b9b762d4cff762]
-entity cosine--><rect fill="#EFF6FF" filter="url(#f1lmzks9slht9h)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="125" x="258.5" y="501"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="105" x="268.5" y="523.0669">cosine_similarity</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="81" x="268.5" y="538.1997">vs all chunks</text><ellipse cx="320.7736" cy="625.4547" fill="#F0FDF4" filter="url(#f1lmzks9slht9h)" rx="60.2736" ry="14.4547" style="stroke: #22C55E; stroke-width: 1.5;"/><text fill="#14532D" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="86" x="277.7736" y="629.5211">Top-K Results</text><!--MD5=[58d0584bf00a192df52be675af4060c3]
-link docs to chunk--><path d="M120,111.21 C120,111.21 120,165.66 120,165.66 " fill="none" id="docs-&gt;chunk" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="120,170.66,124,161.66,120,165.66,116,161.66,120,170.66" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[f719c65e3cd5877eb9f397b030baf6d1]
-link chunk to embed--><path d="M124,221.24 C124,221.24 124,275.88 124,275.88 " fill="none" id="chunk-&gt;embed" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="124,280.88,128,271.88,124,275.88,120,271.88,124,280.88" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[3fff7060fd0c4e8b3dcde3a0a24b4cc6]
-link embed to store--><path d="M125,331.24 C125,331.24 125,388.86 125,388.86 " fill="none" id="embed-&gt;store" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="125,393.86,129,384.86,125,388.86,121,384.86,125,393.86" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[8efdc581dcc4934b230519e9a1c62620]
-link query to embedq--><path d="M321,319.93 C321,319.93 321,385.88 321,385.88 " fill="none" id="query-&gt;embedq" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="321,390.88,325,381.88,321,385.88,317,381.88,321,390.88" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[07d0e59a576869aa907c1b73aa635231]
-link embedq to cosine--><path d="M321,441.24 C321,441.24 321,495.88 321,495.88 " fill="none" id="embedq-&gt;cosine" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="321,500.88,325,491.88,321,495.88,317,491.88,321,500.88" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[4684471f6051558c3413dfb8791f28b5]
-link cosine to results--><path d="M321,551.11 C321,551.11 321,605.9 321,605.9 " fill="none" id="cosine-&gt;results" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="321,610.9,325,601.9,321,605.9,317,601.9,321,610.9" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[a0c1a4ea2fae617c4a25a3db61a81add]
-link store to cosine--><path d="M196.61,416 C229.55,416 262.75,416 262.75,416 C262.75,416 262.75,495.76 262.75,495.76 " fill="none" id="store-&gt;cosine" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="262.75,500.76,266.75,491.76,262.75,495.76,258.75,491.76,262.75,500.76" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[ff44f41d877322434149af373054589d]
-@startuml pipeline-semantic-search
-skinparam direction left to right
-skinparam backgroundColor #FAFAFA
-skinparam defaultFontName Arial
-skinparam defaultFontSize 13
-skinparam ArrowColor #555555
-skinparam ArrowThickness 2
-skinparam linetype ortho
-
-skinparam rectangle {
-    BackgroundColor #EFF6FF
-    BorderColor #3B82F6
-    FontColor #1E3A8A
-    RoundCorner 12
-}
-
-skinparam database {
-    BackgroundColor #FEF9C3
-    BorderColor #EAB308
-    FontColor #713F12
-}
-
-skinparam usecase {
-    BackgroundColor #F0FDF4
-    BorderColor #22C55E
-    FontColor #14532D
-}
-
-package "Indexing — embed.py" #EEF2FF {
-    folder "documents/" as docs
-    rectangle "chunk_text\n300 words · 50 overlap" as chunk
-    rectangle "embed_chunks\nOpenAI API" as embed
-    database "embeddings.json" as store
-
-    docs - -> chunk
-    chunk - -> embed
-    embed - -> store
-}
-
-package "Search — search.py" #FDF4FF {
-    usecase "Query" as query
-    rectangle "embed_query\nOpenAI API" as embedq
-    rectangle "cosine_similarity\nvs all chunks" as cosine
-    usecase "Top-K Results" as results
-
-    query - -> embedq
-    embedq - -> cosine
-    cosine - -> results
-}
-
-store - -> cosine
-
-@enduml
-
-PlantUML version 1.2020.02(Sun Mar 01 10:22:07 UTC 2020)
-(GPL source distribution)
-Java Runtime: OpenJDK Runtime Environment
-JVM: OpenJDK 64-Bit Server VM
-Java Version: 17.0.19+10
-Operating System: Linux
-Default Encoding: UTF-8
-Language: en
-Country: null
---></g></svg>
\ No newline at end of file
diff --git a/diagrams/docs/pipeline-vector-store.svg b/diagrams/docs/pipeline-vector-store.svg
deleted file mode 100644
index 5651427..0000000
--- a/diagrams/docs/pipeline-vector-store.svg
+++ /dev/null
@@ -1,99 +0,0 @@
-<?xml version="1.0" encoding="UTF-8" standalone="no"?><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" contentScriptType="application/ecmascript" contentStyleType="text/css" height="668px" preserveAspectRatio="none" style="width:858px;height:668px;" version="1.1" viewBox="0 0 858 668" width="858px" zoomAndPan="magnify"><defs><filter height="300%" id="fha5baksoolmi" width="300%" x="-1" y="-1"><feGaussianBlur result="blurOut" stdDeviation="2.0"/><feColorMatrix in="blurOut" result="blurOut2" type="matrix" values="0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .4 0"/><feOffset dx="4.0" dy="4.0" in="blurOut2" result="blurOut3"/><feBlend in="SourceGraphic" in2="blurOut3" mode="normal"/></filter></defs><g><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="258" x="299.5" y="12.0669">Phase 2 -- Chroma Vector Store Pipeline</text><!--MD5=[d43075e66b9401f5f099c90dff66d346]
-cluster Ingestion (ingest.py)--><polygon fill="#FFFFFF" filter="url(#fha5baksoolmi)" points="22,39.6328,182,39.6328,189,60.7656,192,60.7656,192,497.1328,22,497.1328,22,39.6328" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="22" x2="189" y1="60.7656" y2="60.7656"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="154" x="26" y="53.6997">Ingestion (ingest.py)</text><!--MD5=[eee07272e247f727b0c9cb5ea2910cc3]
-cluster Search (search.py)--><polygon fill="#FFFFFF" filter="url(#fha5baksoolmi)" points="216,173.6328,361,173.6328,368,194.7656,438,194.7656,438,656.1328,216,656.1328,216,173.6328" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="216" x2="368" y1="194.7656" y2="194.7656"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="139" x="220" y="187.6997">Search (search.py)</text><!--MD5=[57a5b41ed6b3ba96fe9b4a740eb3fdc7]
-entity docs--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="70" x="73" y="91.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="68" y="96.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="68" y="116.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="50" x="83" y="113.6997">.txt files</text><!--MD5=[6fa1cc94bec0cf5ab264efe33d75d3c3]
-entity load--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="136" x="40" y="225.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="35" y="230.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="35" y="250.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="116" x="50" y="247.6997">load_documents()</text><!--MD5=[ff858b2fa29ed35ae3d882895b3a15f0]
-entity chunk--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="100" x="59" y="321.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="54" y="326.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="54" y="346.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="80" x="69" y="343.6997">chunk_text()</text><!--MD5=[1a7a766028402bec58ce31452d7a7648]
-entity embed--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="127" x="48.5" y="446.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="43.5" y="451.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="43.5" y="471.2656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="107" x="58.5" y="468.1997">embed_chunks()</text><!--MD5=[a096c1983bcf686c95ea726f28f7a0ed]
-entity query--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="58" x="334" y="225.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="329" y="230.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="329" y="250.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="38" x="344" y="247.6997">Query</text><!--MD5=[81edeb3f3a35c1dab7f46e9decfdfebe]
-entity embedq--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="118" x="304" y="321.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="299" y="326.6328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="299" y="346.7656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="98" x="314" y="343.6997">embed_query()</text><!--MD5=[8fb73a19f8d09b32c502350ebb5d2be5]
-entity cquery--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="35.1328" style="stroke: #5577AA; stroke-width: 1.5;" width="130" x="260" y="446.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="255" y="451.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="255" y="471.2656"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="110" x="270" y="468.1997">collection.query()</text><!--MD5=[b5cce2cdbd787adcd8934bb88f8254b3]
-entity results--><rect fill="#EEF3FB" filter="url(#fha5baksoolmi)" height="50.2656" style="stroke: #5577AA; stroke-width: 1.5;" width="190" x="232" y="590.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="227" y="595.1328"/><rect fill="#EEF3FB" height="5" style="stroke: #5577AA; stroke-width: 1.5;" width="10" x="227" y="630.3984"/><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="86" x="242" y="612.1997">Top-K Results</text><text fill="#222222" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="170" x="242" y="627.3325">(text + source + distance)</text><!--MD5=[d85c0a1eb646255ad28a1eb3e90eaa32]
-entity openai--><path d="M663.835,442.4894 C667.1277,435.2659 672.3414,436.1026 676.6173,441.4931 C680.0176,435.0792 685.7228,435.5954 689.0973,441.5192 C691.6705,436.9846 696.1511,436.5453 699.0845,441.1244 C702.7275,434.8774 708.3475,432.647 713.2868,439.7448 C716.5627,434.4833 722.0088,434.6406 724.5873,440.4563 C727.487,434.6172 731.4132,435.2085 735.3898,439.2829 C738.5809,431.7226 743.7931,431.3201 749.0607,437.1652 C752.9629,433.2862 757.0274,433.6585 759.4121,438.8895 C763.0709,432.484 769.0669,433.1716 771.662,439.8829 C775.101,434.1339 781.5417,433.03 784.9154,439.9511 C787.8691,434.6344 792.9245,435.6335 795.3463,440.4961 C800.1533,436.0428 804.3976,436.0124 807.2952,442.5685 C811.7915,435.1647 815.8976,436.1596 820.8024,442.1567 C824.0394,437.2752 829.2042,436.3206 832.5035,442.0882 C838.6991,443.41 841.1537,448.8893 836.0195,453.655 C842.5332,454.8929 844.7109,457.9631 842.3017,464.4121 C846.637,470.97 843.7827,475.527 836.2549,475.7682 C839.9568,481.3819 838.6597,485.5913 832.1783,487.7365 C828.7109,493.8966 823.9772,493.6881 819.7924,488.5208 C815.9591,494.4228 811.7036,494.2377 807.6621,488.7424 C805.6099,495.3442 799.7881,496.7766 795.0442,491.6344 C792.0953,496.906 788.2131,496.5461 785.1212,491.7758 C782.9264,497.8229 777.5868,499.4607 773.0484,494.3002 C769.5195,501.0609 764.8674,501.6812 760.0009,495.6662 C756.23,501.3552 752.3189,501.3978 748.9306,495.3066 C744.55,501.9554 739.8568,502.2077 735.6966,495.0946 C731.0645,499.8501 725.8767,499.4683 723.6526,492.5997 C720.7068,497.941 715.0143,498.0348 712.0471,492.6227 C707.5735,497.7266 702.9539,499.1638 699.1543,491.919 C695.1953,496.5972 691.1095,495.38 689.1504,489.9415 C685.0524,495.8668 679.5274,495.6227 676.2618,489.081 C672.0738,494.2949 666.8131,494.5433 663.7083,487.9311 C655.9358,484.8511 656.5839,480.5434 660.9809,474.9136 C654.3407,474.2883 653.4529,470.4153 655.8453,465.075 C651.93,459.4246 653.5013,454.3598 660.9438,453.8839 C657.0052,448.7166 657.3805,445.0196 663.835,442.4894 " fill="#F0FFF0" style="stroke: #448844; stroke-width: 1.5;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="670" y="460.6997">OpenAI API</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="154" x="670" y="475.8325">text-embedding-3-small</text><!--MD5=[ba0e3141185c9f6ac85d8ff32b47afed]
-entity chroma--><path d="M103,595.6328 C103,585.6328 150,585.6328 150,585.6328 C150,585.6328 197,585.6328 197,595.6328 L197,634.8984 C197,644.8984 150,644.8984 150,644.8984 C150,644.8984 103,644.8984 103,634.8984 L103,595.6328 " fill="#FFF8E7" filter="url(#fha5baksoolmi)" style="stroke: #CC9900; stroke-width: 1.5;"/><path d="M103,595.6328 C103,605.6328 150,605.6328 150,605.6328 C150,605.6328 197,605.6328 197,595.6328 " fill="none" style="stroke: #CC9900; stroke-width: 1.5;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="74" x="113" y="621.6997">Chroma DB</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="74" x="113" y="636.8325">(persistent)</text><!--MD5=[34aeeab8a23e75e12ac7d0294eef783a]
-link docs to load--><path d="M108,126.6528 C108,126.6528 108,220.6128 108,220.6128 " fill="none" id="docs-&gt;load" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="108,225.6128,112,216.6128,108,220.6128,104,216.6128,108,225.6128" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[197124bfd89600884ad83f5bdb81fe2d]
-link load to chunk--><path d="M109,260.7428 C109,260.7428 109,316.3128 109,316.3128 " fill="none" id="load-&gt;chunk" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="109,321.3128,113,312.3128,109,316.3128,105,312.3128,109,321.3128" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[f719c65e3cd5877eb9f397b030baf6d1]
-link chunk to embed--><path d="M109,356.7028 C109,356.7028 109,440.7428 109,440.7428 " fill="none" id="chunk-&gt;embed" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="109,445.7428,113,436.7428,109,440.7428,105,436.7428,109,445.7428" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[14e589a052329e665bcf72ada14e2408]
-link embed to openai--><path d="M164.5,445.9328 C164.5,444.8228 164.5,444.1328 164.5,444.1328 C164.5,444.1328 649.7,444.1328 649.7,444.1328 " fill="none" id="embed-&gt;openai" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="654.7,444.1328,645.7,440.1328,649.7,444.1328,645.7,448.1328,654.7,444.1328" style="stroke: #555555; stroke-width: 2.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="45" x="574.5" y="399.6997">API call</text><!--MD5=[fad15655625d41508c0bd675faf1956d]
-link openai to embed--><path d="M654.93,441.1328 C492.03,441.1328 170,441.1328 170,441.1328 C170,441.1328 170,441.6128 170,441.6128 " fill="none" id="openai-&gt;embed" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="170,445.9028,174,436.9028,170,440.9028,166,436.9028,170,445.9028" style="stroke: #555555; stroke-width: 2.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="112" x="455" y="399.6997">1536-dim vectors</text><!--MD5=[b1ec60cbada152cf55f28d4273f8e0c5]
-link embed to chroma--><path d="M139.25,481.3128 C139.25,481.3128 139.25,580.6228 139.25,580.6228 " fill="none" id="embed-&gt;chroma" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="139.25,585.6228,143.25,576.6228,139.25,580.6228,135.25,576.6228,139.25,585.6228" style="stroke: #555555; stroke-width: 2.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="47" y="536.6997">upsert with</text><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="100" x="32" y="551.8325">ids + metadata</text><!--MD5=[8efdc581dcc4934b230519e9a1c62620]
-link query to embedq--><path d="M363,260.7428 C363,260.7428 363,316.3128 363,316.3128 " fill="none" id="query-&gt;embedq" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="363,321.3128,367,312.3128,363,316.3128,359,312.3128,363,321.3128" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[07d15e8df06b7fbd510547f2864e4d80]
-link embedq to openai--><path d="M400.67,357.0128 C400.67,385.7728 400.67,439.1328 400.67,439.1328 C400.67,439.1328 649.64,439.1328 649.64,439.1328 " fill="none" id="embedq-&gt;openai" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="654.64,439.1328,645.64,435.1328,649.64,439.1328,645.64,443.1328,654.64,439.1328" style="stroke: #555555; stroke-width: 2.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="45" x="784" y="399.6997">API call</text><!--MD5=[3a9578d9bd486bb01d5041cfdd4d7983]
-link openai to embedq--><path d="M654.83,436.1328 C555.46,436.1328 411.33,436.1328 411.33,436.1328 C411.33,436.1328 411.33,361.9228 411.33,361.9228 " fill="none" id="openai-&gt;embedq" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="411.33,356.9228,407.33,365.9228,411.33,361.9228,415.33,365.9228,411.33,356.9228" style="stroke: #555555; stroke-width: 2.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="80" x="662" y="399.6997">query vector</text><!--MD5=[5c13e55db4cb974e2ddc415d002712d7]
-link embedq to cquery--><path d="M347,356.7028 C347,356.7028 347,440.7428 347,440.7428 " fill="none" id="embedq-&gt;cquery" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="347,445.7428,351,436.7428,347,440.7428,343,436.7428,347,445.7428" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[f67d8fe0aaaf6a85deca3a91259f7f85]
-link cquery to chroma--><path d="M259.87,470.1328 C225.35,470.1328 190,470.1328 190,470.1328 C190,470.1328 190,580.4428 190,580.4428 " fill="none" id="cquery-&gt;chroma" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="190,585.4428,194,576.4428,190,580.4428,186,576.4428,190,585.4428" style="stroke: #555555; stroke-width: 2.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="76" x="316" y="544.1997">ANN search</text><!--MD5=[f593972c37969fbb6ae91af54d31458c]
-link chroma to cquery--><path d="M183,585.6028 C183,540.1228 183,458.1328 183,458.1328 C183,458.1328 254.82,458.1328 254.82,458.1328 " fill="none" id="chroma-&gt;cquery" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="259.82,458.1328,250.82,454.1328,254.82,458.1328,250.82,462.1328,259.82,458.1328" style="stroke: #555555; stroke-width: 2.0;"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="100" x="174" y="544.1997">nearest vectors</text><!--MD5=[a38503b39590097c9f1ad789f279155b]
-link cquery to results--><path d="M325,481.3128 C325,481.3128 325,585.0328 325,585.0328 " fill="none" id="cquery-&gt;results" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="325,590.0328,329,581.0328,325,585.0328,321,581.0328,325,590.0328" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[308ceff5917ee9d8aca216a7e4a31942]
-@startuml pipeline-vector-store
-
-skinparam backgroundColor #FFFFFF
-skinparam defaultFontName Arial
-skinparam defaultFontSize 13
-skinparam ArrowColor #555555
-skinparam ArrowThickness 2
-skinparam componentStyle rectangle
-skinparam linetype ortho
-
-skinparam component {
-  BackgroundColor #EEF3FB
-  BorderColor #5577AA
-  FontColor #222222
-}
-
-skinparam database {
-  BackgroundColor #FFF8E7
-  BorderColor #CC9900
-}
-
-skinparam cloud {
-  BackgroundColor #F0FFF0
-  BorderColor #448844
-}
-
-title Phase 2 - - Chroma Vector Store Pipeline
-
-package "Ingestion (ingest.py)" {
-  [.txt files] as docs
-  [load_documents()] as load
-  [chunk_text()] as chunk
-  [embed_chunks()] as embed
-}
-
-cloud "OpenAI API\ntext-embedding-3-small" as openai
-
-database "Chroma DB\n(persistent)" as chroma
-
-package "Search (search.py)" {
-  [Query] as query
-  [embed_query()] as embedq
-  [collection.query()] as cquery
-  [Top-K Results\n(text + source + distance)] as results
-}
-
-docs - -> load
-load - -> chunk
-chunk - -> embed
-embed - -> openai : API call
-openai - -> embed : 1536-dim vectors
-embed - -> chroma : upsert with\nids + metadata
-
-query - -> embedq
-embedq - -> openai : API call
-openai - -> embedq : query vector
-embedq - -> cquery
-cquery - -> chroma : ANN search
-chroma - -> cquery : nearest vectors
-cquery - -> results
-
-@enduml
-
-PlantUML version 1.2020.02(Sun Mar 01 10:22:07 UTC 2020)
-(GPL source distribution)
-Java Runtime: OpenJDK Runtime Environment
-JVM: OpenJDK 64-Bit Server VM
-Java Version: 17.0.19+10
-Operating System: Linux
-Default Encoding: UTF-8
-Language: en
-Country: null
---></g></svg>
\ No newline at end of file
diff --git a/diagrams/docs/pipeline.svg b/diagrams/docs/pipeline.svg
deleted file mode 100644
index aacf16b..0000000
--- a/diagrams/docs/pipeline.svg
+++ /dev/null
@@ -1,80 +0,0 @@
-<?xml version="1.0" encoding="UTF-8" standalone="no"?><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" contentScriptType="application/ecmascript" contentStyleType="text/css" height="668px" preserveAspectRatio="none" style="width:425px;height:668px;background:#FAFAFA;" version="1.1" viewBox="0 0 425 668" width="425px" zoomAndPan="magnify"><defs><filter height="300%" id="ft2u3n2i3ka5f" width="300%" x="-1" y="-1"><feGaussianBlur result="blurOut" stdDeviation="2.0"/><feColorMatrix in="blurOut" result="blurOut2" type="matrix" values="0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .4 0"/><feOffset dx="4.0" dy="4.0" in="blurOut2" result="blurOut3"/><feBlend in="SourceGraphic" in2="blurOut3" mode="normal"/></filter></defs><g><!--MD5=[ff4c7cbfc86cbafc5088fc7dc3b82e86]
-cluster Indexing — embed.py--><polygon fill="#EEF2FF" filter="url(#ft2u3n2i3ka5f)" points="22,24,189,24,196,45.1328,218,45.1328,218,454,22,454,22,24" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="22" x2="196" y1="45.1328" y2="45.1328"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="161" x="26" y="38.0669">Indexing — embed.py</text><!--MD5=[b7354f1b6b1a4ce6b3292a1a387e6ba1]
-cluster Search — search.py--><polygon fill="#FDF4FF" filter="url(#ft2u3n2i3ka5f)" points="242,258.5,393,258.5,400,279.6328,403,279.6328,403,656,242,656,242,258.5" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="242" x2="400" y1="279.6328" y2="279.6328"/><text fill="#000000" font-family="Arial" font-size="13" font-weight="bold" lengthAdjust="spacingAndGlyphs" textLength="145" x="246" y="272.5669">Search — search.py</text><!--MD5=[57a5b41ed6b3ba96fe9b4a740eb3fdc7]
-entity docs--><polygon fill="#FEFECE" filter="url(#ft2u3n2i3ka5f)" points="67,58,113,58,120,79,173,79,173,111.2656,67,111.2656,67,58" style="stroke: #000000; stroke-width: 1.5;"/><line style="stroke: #000000; stroke-width: 1.5;" x1="67" x2="120" y1="79" y2="79"/><text fill="#000000" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="76" x="77" y="98.0669">documents/</text><!--MD5=[ff858b2fa29ed35ae3d882895b3a15f0]
-entity chunk--><rect fill="#EFF6FF" filter="url(#ft2u3n2i3ka5f)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="164" x="38" y="171"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="48" y="193.0669">chunk_text</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="144" x="48" y="208.1997">300 words · 50 overlap</text><!--MD5=[1a7a766028402bec58ce31452d7a7648]
-entity embed--><rect fill="#EFF6FF" filter="url(#ft2u3n2i3ka5f)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="117" x="65.5" y="281"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="97" x="75.5" y="303.0669">embed_chunks</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="75.5" y="318.1997">OpenAI API</text><!--MD5=[0686720d71d32ea68f81164171d509ce]
-entity store--><path d="M67.5,404 C67.5,394 132,394 132,394 C132,394 196.5,394 196.5,404 L196.5,428.1328 C196.5,438.1328 132,438.1328 132,438.1328 C132,438.1328 67.5,438.1328 67.5,428.1328 L67.5,404 " fill="#FEF9C3" filter="url(#ft2u3n2i3ka5f)" style="stroke: #EAB308; stroke-width: 1.5;"/><path d="M67.5,404 C67.5,414 132,414 132,414 C132,414 196.5,414 196.5,404 " fill="none" style="stroke: #EAB308; stroke-width: 1.5;"/><text fill="#713F12" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="109" x="77.5" y="430.0669">embeddings.json</text><ellipse cx="320.8701" cy="306.2005" fill="#F0FDF4" filter="url(#ft2u3n2i3ka5f)" rx="29.8701" ry="13.7005" style="stroke: #22C55E; stroke-width: 1.5;"/><text fill="#14532D" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="38" x="301.8701" y="310.2669">Query</text><!--MD5=[81edeb3f3a35c1dab7f46e9decfdfebe]
-entity embedq--><rect fill="#EFF6FF" filter="url(#ft2u3n2i3ka5f)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="108" x="267" y="391"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="88" x="277" y="413.0669">embed_query</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="70" x="277" y="428.1997">OpenAI API</text><!--MD5=[3db17c234f6b9cf849b9b762d4cff762]
-entity cosine--><rect fill="#EFF6FF" filter="url(#ft2u3n2i3ka5f)" height="50.2656" rx="6" ry="6" style="stroke: #3B82F6; stroke-width: 1.5;" width="125" x="258.5" y="501"/><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="105" x="268.5" y="523.0669">cosine_similarity</text><text fill="#1E3A8A" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="81" x="268.5" y="538.1997">vs all chunks</text><ellipse cx="320.7736" cy="625.4547" fill="#F0FDF4" filter="url(#ft2u3n2i3ka5f)" rx="60.2736" ry="14.4547" style="stroke: #22C55E; stroke-width: 1.5;"/><text fill="#14532D" font-family="Arial" font-size="13" lengthAdjust="spacingAndGlyphs" textLength="86" x="277.7736" y="629.5211">Top-K Results</text><!--MD5=[58d0584bf00a192df52be675af4060c3]
-link docs to chunk--><path d="M120,111.21 C120,111.21 120,165.66 120,165.66 " fill="none" id="docs-&gt;chunk" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="120,170.66,124,161.66,120,165.66,116,161.66,120,170.66" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[f719c65e3cd5877eb9f397b030baf6d1]
-link chunk to embed--><path d="M124,221.24 C124,221.24 124,275.88 124,275.88 " fill="none" id="chunk-&gt;embed" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="124,280.88,128,271.88,124,275.88,120,271.88,124,280.88" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[3fff7060fd0c4e8b3dcde3a0a24b4cc6]
-link embed to store--><path d="M125,331.24 C125,331.24 125,388.86 125,388.86 " fill="none" id="embed-&gt;store" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="125,393.86,129,384.86,125,388.86,121,384.86,125,393.86" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[8efdc581dcc4934b230519e9a1c62620]
-link query to embedq--><path d="M321,319.93 C321,319.93 321,385.88 321,385.88 " fill="none" id="query-&gt;embedq" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="321,390.88,325,381.88,321,385.88,317,381.88,321,390.88" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[07d0e59a576869aa907c1b73aa635231]
-link embedq to cosine--><path d="M321,441.24 C321,441.24 321,495.88 321,495.88 " fill="none" id="embedq-&gt;cosine" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="321,500.88,325,491.88,321,495.88,317,491.88,321,500.88" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[4684471f6051558c3413dfb8791f28b5]
-link cosine to results--><path d="M321,551.11 C321,551.11 321,605.9 321,605.9 " fill="none" id="cosine-&gt;results" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="321,610.9,325,601.9,321,605.9,317,601.9,321,610.9" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[a0c1a4ea2fae617c4a25a3db61a81add]
-link store to cosine--><path d="M196.61,416 C229.55,416 262.75,416 262.75,416 C262.75,416 262.75,495.76 262.75,495.76 " fill="none" id="store-&gt;cosine" style="stroke: #555555; stroke-width: 2.0;"/><polygon fill="#555555" points="262.75,500.76,266.75,491.76,262.75,495.76,258.75,491.76,262.75,500.76" style="stroke: #555555; stroke-width: 2.0;"/><!--MD5=[b5895bd08e5963fd690290115f3d9b71]
-@startuml pipeline
-skinparam direction left to right
-skinparam backgroundColor #FAFAFA
-skinparam defaultFontName Arial
-skinparam defaultFontSize 13
-skinparam ArrowColor #555555
-skinparam ArrowThickness 2
-skinparam linetype ortho
-
-skinparam rectangle {
-    BackgroundColor #EFF6FF
-    BorderColor #3B82F6
-    FontColor #1E3A8A
-    RoundCorner 12
-}
-
-skinparam database {
-    BackgroundColor #FEF9C3
-    BorderColor #EAB308
-    FontColor #713F12
-}
-
-skinparam usecase {
-    BackgroundColor #F0FDF4
-    BorderColor #22C55E
-    FontColor #14532D
-}
-
-package "Indexing — embed.py" #EEF2FF {
-    folder "documents/" as docs
-    rectangle "chunk_text\n300 words · 50 overlap" as chunk
-    rectangle "embed_chunks\nOpenAI API" as embed
-    database "embeddings.json" as store
-
-    docs - -> chunk
-    chunk - -> embed
-    embed - -> store
-}
-
-package "Search — search.py" #FDF4FF {
-    usecase "Query" as query
-    rectangle "embed_query\nOpenAI API" as embedq
-    rectangle "cosine_similarity\nvs all chunks" as cosine
-    usecase "Top-K Results" as results
-
-    query - -> embedq
-    embedq - -> cosine
-    cosine - -> results
-}
-
-store - -> cosine
-
-@enduml
-
-PlantUML version 1.2020.02(Sun Mar 01 10:22:07 UTC 2020)
-(GPL source distribution)
-Java Runtime: OpenJDK Runtime Environment
-JVM: OpenJDK 64-Bit Server VM
-Java Version: 17.0.19+10
-Operating System: Linux
-Default Encoding: UTF-8
-Language: en
-Country: null
---></g></svg>
\ No newline at end of file
diff --git a/docs/pipeline-vector-store.puml b/docs/pipeline-vector-store.puml
index 954dd62..0322a30 100644
--- a/docs/pipeline-vector-store.puml
+++ b/docs/pipeline-vector-store.puml
@@ -3,60 +3,51 @@
 skinparam backgroundColor #FFFFFF
 skinparam defaultFontName Arial
 skinparam defaultFontSize 13
-skinparam ArrowColor #555555
-skinparam ArrowThickness 2
-skinparam componentStyle rectangle
-skinparam linetype ortho
-
-skinparam component {
-  BackgroundColor #EEF3FB
-  BorderColor #5577AA
-  FontColor #222222
-}
-
-skinparam database {
-  BackgroundColor #FFF8E7
-  BorderColor #CC9900
-}
-
-skinparam cloud {
-  BackgroundColor #F0FFF0
-  BorderColor #448844
-}
-
-title Phase 2 -- Chroma Vector Store Pipeline
-
-package "Ingestion (ingest.py)" {
-  [.txt files] as docs
-  [load_documents()] as load
-  [chunk_text()] as chunk
-  [embed_chunks()] as embed
-}
-
-cloud "OpenAI API\ntext-embedding-3-small" as openai
-
-database "Chroma DB\n(persistent)" as chroma
-
-package "Search (search.py)" {
-  [Query] as query
-  [embed_query()] as embedq
-  [collection.query()] as cquery
-  [Top-K Results\n(text + source + distance)] as results
-}
-
-docs --> load
-load --> chunk
-chunk --> embed
-embed --> openai : API call
+skinparam sequenceArrowThickness 2
+skinparam SequenceBoxBackgroundColor #F8F9FF
+skinparam SequenceBoxBorderColor #AABBDD
+skinparam ParticipantBackgroundColor #EEF3FB
+skinparam ParticipantBorderColor #5577AA
+skinparam ParticipantFontColor #222222
+skinparam DatabaseBackgroundColor #FFF8E7
+skinparam DatabaseBorderColor #CC9900
+skinparam ActorBackgroundColor #F0FFF0
+skinparam ActorBorderColor #448844
+
+title Chroma Vector Store Pipeline
+
+box "Ingestion" #EEF6FF
+  participant "load_documents()" as load
+  participant "chunk_text()" as chunk
+  participant "embed_chunks()" as embed
+end box
+
+participant "OpenAI API" as openai
+database "Chroma DB" as chroma
+
+box "Search" #FFF0F8
+  actor "Query" as query
+  participant "embed_query()" as embedq
+  participant "collection.query()" as cquery
+end box
+
+== Ingestion ==
+
+[-> load : .txt files
+load -> chunk : text
+chunk -> embed : chunks
+embed -> openai : embed request
 openai --> embed : 1536-dim vectors
-embed --> chroma : upsert with\nids + metadata
+embed -> chroma : upsert(ids, embeddings, metadata)
 
-query --> embedq
-embedq --> openai : API call
+== Search ==
+
+query -> embedq : query string
+embedq -> openai : embed request
 openai --> embedq : query vector
-embedq --> cquery
-cquery --> chroma : ANN search
+embedq -> cquery : query vector
+cquery -> chroma : ANN search
 chroma --> cquery : nearest vectors
-cquery --> results
+cquery --> query : Top-K results\n(text + source + distance)
 
 @enduml