Deploy sha:591a9f67fc1c

_version.json

@@ -1,6 +1,6 @@
 {
   "tag": "",
-  "sha": "e92eb09bd0db",
-  "date": "2026-03-31T08:57:50Z",
+  "sha": "591a9f67fc1c",
+  "date": "2026-03-31T09:21:23Z",
   "deploy": "space"
 }

analysis/mimic_iv_ecg/report.html

@@ -129,9 +129,10 @@
       </div>
       <div class="legend" style="margin-top:14px;">
         <div class="legend-item"><span class="legend-swatch" style="background:#1a2233;border-radius:50%;"></span> Root / Directory</div>
-        <div class="legend-item"><span class="legend-swatch" style="background:#4f8ef7;"></span> CSV Metadata File</div>
+        <div class="legend-item"><span class="legend-swatch" style="background:#4f8ef7;"></span> CSV Metadata File <em style="font-size:0.72rem;color:#888">(click to expand columns ▾)</em></div>
         <div class="legend-item"><span class="legend-swatch" style="background:#10b981;"></span> WFDB Waveform File (.hea / .dat)</div>
         <div class="legend-item"><span class="legend-swatch" style="background:#94a3b8;"></span> Other File</div>
+        <div class="legend-item"><span class="legend-swatch" style="background:#cbd5e1;border-radius:50%;width:10px;height:10px;"></span> Column name <em style="font-size:0.72rem;color:#888">(monospace, hover for type)</em></div>
       </div>
     </div>
   </section>
@@ -5788,17 +5789,82 @@ sections.forEach(s => observer.observe(s));
 
 // ── Dataset Structure Tree ────────────────────────────────────────────────────
 (function() {
+  // Column metadata: name → description shown in tooltip
+  const COL_META = {
+    "subject_id":            "Patient identifier (integer)",
+    "study_id":              "ECG study identifier — primary key (integer)",
+    "file_name":             "WFDB record filename without extension (string)",
+    "ecg_time":              "Acquisition datetime — date-shifted for de-identification",
+    "path":                  "Relative path to WFDB record within files/",
+    "cart_id":               "ECG cart / device identifier (string)",
+    "report_0 … report_17":  "18 machine-generated diagnostic phrase fields (string, often blank)",
+    "bandwidth":             "ECG device bandwidth setting (string)",
+    "filtering":             "ECG device filter setting (string)",
+    "rr_interval":           "RR interval — ms (float)",
+    "p_onset":               "P-wave onset — ms (float)",
+    "p_end":                 "P-wave end — ms (float)",
+    "qrs_onset":             "QRS complex onset — ms (float)",
+    "qrs_end":               "QRS complex end — ms (float)",
+    "t_end":                 "T-wave end — ms (float)",
+    "p_axis":                "P-wave electrical axis — degrees (float)",
+    "qrs_axis":              "QRS electrical axis — degrees (float)",
+    "t_axis":                "T-wave electrical axis — degrees (float)",
+    "waveform_path":         "Path used to link waveform to clinical note",
+    "note_id":               "MIMIC-IV clinical note identifier (integer)",
+    "note_seq":              "Note sequence number within admission (integer)",
+    "charttime":             "Datetime of linked clinical note",
+    "variable":              "Column / variable name (string)",
+    "label":                 "Short human-readable label (string)",
+    "description":           "Full description of the variable (string)",
+    "unit":                  "Measurement unit (string)",
+  };
+  const col = name => ({ name, type: "column", desc: COL_META[name] || "" });
+
+  // Shared measurement columns (machine_measurements + original)
+  const measCols = [
+    "subject_id","study_id","cart_id","ecg_time",
+    "report_0 … report_17",
+    "bandwidth","filtering",
+    "rr_interval","p_onset","p_end","qrs_onset","qrs_end","t_end",
+    "p_axis","qrs_axis","t_axis",
+  ].map(col);
+
   const treeData = {
     name: "mimic-iv-ecg-v1.0/", type: "root", desc: "Root dataset directory (800,035 studies)",
     children: [
-      { name: "record_list.csv",                          type: "csv",  desc: "800,035 rows — all ECG record paths + subject/study IDs + ecg_time" },
-      { name: "machine_measurements.csv",                 type: "csv",  desc: "789,481 rows × 33 cols — automated interval measurements + text report phrases" },
-      { name: "machine_measurements_original.csv",        type: "csv",  desc: "789,481 rows — pre-processed raw measurement values" },
-      { name: "machine_measurements_data_dictionary.csv", type: "csv",  desc: "18 rows — variable descriptions and units" },
-      { name: "waveform_note_links.csv",                  type: "csv",  desc: "609,272 rows — links ECG studies to MIMIC-IV clinical notes" },
-      { name: "RECORDS",                                  type: "file", desc: "1000 patient-group directory names (p1000–p1999)" },
-      { name: "SHA256SUMS.txt",                           type: "file", desc: "File integrity checksums (SHA-256)" },
-      { name: "LICENSE.txt",                              type: "file", desc: "ODC Open Database License (ODbL)" },
+      {
+        name: "record_list.csv", type: "csv",
+        desc: "800,035 rows — all ECG record paths + subject/study IDs + ecg_time",
+        _collapsed: true,
+        children: ["subject_id","study_id","file_name","ecg_time","path"].map(col),
+      },
+      {
+        name: "machine_measurements.csv", type: "csv",
+        desc: "789,481 rows × 33 cols — automated interval measurements + text report phrases",
+        _collapsed: true,
+        children: measCols,
+      },
+      {
+        name: "machine_measurements_original.csv", type: "csv",
+        desc: "789,481 rows — pre-processed raw measurement values (same 33 columns)",
+        _collapsed: true,
+        children: measCols,
+      },
+      {
+        name: "machine_measurements_data_dictionary.csv", type: "csv",
+        desc: "18 rows — variable descriptions and units",
+        _collapsed: true,
+        children: ["variable","label","description","unit"].map(col),
+      },
+      {
+        name: "waveform_note_links.csv", type: "csv",
+        desc: "609,272 rows — links ECG studies to MIMIC-IV clinical notes",
+        _collapsed: true,
+        children: ["subject_id","study_id","waveform_path","note_id","note_seq","charttime"].map(col),
+      },
+      { name: "RECORDS",        type: "file", desc: "1000 patient-group directory names (p1000–p1999)" },
+      { name: "SHA256SUMS.txt", type: "file", desc: "File integrity checksums (SHA-256)" },
+      { name: "LICENSE.txt",    type: "file", desc: "ODC Open Database License (ODbL)" },
       {
         name: "files/", type: "dir", desc: "Waveform data organised in a 4-level hierarchy",
         _collapsed: false,
@@ -5806,207 +5872,203 @@ sections.forEach(s => observer.observe(s));
           name: "pXXXX/", type: "dir", desc: "1,000 patient-group folders (p1000–p1999)",
           _collapsed: false,
           children: [{
-            name: "pNNNNNNNN/", type: "dir", desc: "Individual patient directory (subject_id — 160,862 unique patients)",
+            name: "pNNNNNNNN/", type: "dir",
+            desc: "Individual patient directory (subject_id — 160,862 unique patients)",
             _collapsed: false,
             children: [{
-              name: "sNNNNNNNN/", type: "dir", desc: "Study directory (study_id — one per ECG acquisition)",
+              name: "sNNNNNNNN/", type: "dir",
+              desc: "Study directory (study_id — one per ECG acquisition)",
               _collapsed: false,
               children: [
                 { name: "NNNNNNNN.hea", type: "hea", desc: "WFDB header — 12 leads, 500 Hz, 10 s (plain text, ~600 B)" },
-                { name: "NNNNNNNN.dat", type: "dat", desc: "WFDB signal — 16-bit signed binary, 5,000 samples/lead (~120 KB)" }
-              ]
-            }]
-          }]
-        }]
-      }
-    ]
+                { name: "NNNNNNNN.dat", type: "dat", desc: "WFDB signal — 16-bit signed binary, 5,000 samples/lead (~120 KB)" },
+              ],
+            }],
+          }],
+        }],
+      },
+    ],
   };
 
-  const nodeColor = n => {
-    if (n.data.type === "root" || n.data.type === "dir") return "#1a2233";
-    if (n.data.type === "csv")  return "#4f8ef7";
-    if (n.data.type === "hea" || n.data.type === "dat") return "#10b981";
+  // ── Helpers ──────────────────────────────────────────────────────────────────
+  const isDir  = d => d.data.type === "root" || d.data.type === "dir";
+  const isCSV  = d => d.data.type === "csv";
+  const isCol  = d => d.data.type === "column";
+
+  const nodeColor = d => {
+    if (isDir(d))  return "#1a2233";
+    if (isCSV(d))  return "#4f8ef7";
+    if (d.data.type === "hea" || d.data.type === "dat") return "#10b981";
+    if (isCol(d))  return "#94a3b8";
     return "#94a3b8";
   };
-  const nodeShape = n => (n.data.type === "root" || n.data.type === "dir") ? "diamond" : "rect";
 
-  const margin = {top: 20, right: 320, bottom: 20, left: 160};
-  const baseWidth  = 1100;
-  let   svgHeight  = 600;
+  // Approximate pixel width of a label string
+  const labelPx = (s, mono) => (s || "").length * (mono ? 7.2 : 7.5);
+
+  // ── SVG setup ────────────────────────────────────────────────────────────────
+  const margin = { top: 24, right: 360, bottom: 24, left: 160 };
+  const baseWidth = 1100;
+  let svgHeight = 600;
 
   const svg = d3.select("#ds-tree-svg")
     .attr("width", baseWidth + margin.left + margin.right)
-    .attr("height", svgHeight + margin.top + margin.bottom)
+    .attr("height", svgHeight)
     .style("font", "13px 'Segoe UI', system-ui, sans-serif");
 
   const g = svg.append("g").attr("transform", `translate(${margin.left},${margin.top})`);
-
-  // link path generator
   const linkPath = d3.linkHorizontal().x(d => d.y).y(d => d.x);
 
-  // Build hierarchy with collapse state
+  // ── Hierarchy builder ──────────────────────────────────────────────────────���──
   function buildHierarchy(data) {
-    const root = d3.hierarchy(data, d => d._collapsed ? null : d.children);
-    root.x0 = 0; root.y0 = 0;
-    return root;
+    return d3.hierarchy(data, d => d._collapsed ? null : d.children);
   }
 
-  // Collapse/expand node
-  function toggle(d) {
-    if (!d.data.children) return;
-    if (d.children) {
-      d.data._collapsed = true;
-    } else {
-      d.data._collapsed = false;
-    }
-  }
-
-  function setAllCollapsed(node, collapsed) {
-    if (node.children) node.data.children = node.children.map(c => c.data);
-    if (collapsed) node.data._collapsed = true;
-    else           node.data._collapsed = false;
-    (node.children || node._children || []).forEach(c => setAllCollapsed(c, collapsed));
-  }
-
-  let root = buildHierarchy(treeData);
-
+  // ── Render ────────────────────────────────────────────────────────────────────
   function update() {
-    root = buildHierarchy(treeData);
+    const root = buildHierarchy(treeData);
 
-    const treeLayout = d3.tree().nodeSize([34, 220]);
+    // Use tighter row spacing for column nodes; standard for others
+    const treeLayout = d3.tree().nodeSize([26, 220]);
     treeLayout(root);
 
-    // centre vertically
     const nodes = root.descendants();
-    const minX = d3.min(nodes, d => d.x);
-    const maxX = d3.max(nodes, d => d.x);
-    svgHeight = Math.max(400, maxX - minX + margin.top + margin.bottom + 40);
+    const minX  = d3.min(nodes, d => d.x);
+    const maxX  = d3.max(nodes, d => d.x);
+    svgHeight = Math.max(420, maxX - minX + margin.top + margin.bottom + 40);
     svg.attr("height", svgHeight);
     const offsetY = margin.top - minX;
 
     g.selectAll("*").remove();
 
-    // Links
-    g.selectAll(".link")
+    // ── Links ──────────────────────────────────────────────────────────────────
+    g.selectAll(".lnk")
       .data(root.links())
       .join("path")
-        .attr("class","link")
-        .attr("fill","none")
-        .attr("stroke","#c8d0e0")
-        .attr("stroke-width",1.5)
+        .attr("fill", "none")
+        .attr("stroke",       d => isCol(d.target) ? "#dde3ef" : "#c8d0e0")
+        .attr("stroke-width", d => isCol(d.target) ? 1          : 1.5)
+        .attr("stroke-dasharray", d => isCol(d.target) ? "3,3" : null)
         .attr("d", d => linkPath({
-          source: {x: d.source.x + offsetY, y: d.source.y},
-          target: {x: d.target.x + offsetY, y: d.target.y}
+          source: { x: d.source.x + offsetY, y: d.source.y },
+          target: { x: d.target.x + offsetY, y: d.target.y },
         }));
 
-    // Nodes
-    const node = g.selectAll(".node")
+    // ── Node groups ────────────────────────────────────────────────────────────
+    const node = g.selectAll(".nd")
       .data(nodes)
       .join("g")
-        .attr("class","node")
         .attr("transform", d => `translate(${d.y},${d.x + offsetY})`)
         .style("cursor", d => d.data.children ? "pointer" : "default")
         .on("click", (event, d) => {
           if (!d.data.children) return;
-          toggle(d);
+          d.data._collapsed = !d.data._collapsed;
           update();
         })
         .on("mouseenter", (event, d) => {
-          const icon = d.data.type === "dir" || d.data.type === "root" ? "📁" :
-                       d.data.type === "csv" ? "📄" :
-                       d.data.type === "hea" ? "📋" :
-                       d.data.type === "dat" ? "💾" : "📎";
-          showTip(`<b>${icon} ${d.data.name}</b><br>${d.data.desc}`, event);
+          const icons = { root:"📁", dir:"📁", csv:"📄", hea:"📋", dat:"💾", column:"⬡", file:"📎" };
+          showTip(`<b>${icons[d.data.type]||"📎"} ${d.data.name}</b><br>${d.data.desc}`, event);
         })
-        .on("mousemove", (event) => {
-          tip.style("left",(event.pageX+14)+"px").style("top",(event.pageY-28)+"px");
-        })
-        .on("mouseleave", () => hideTip());
+        .on("mousemove",  event => tip.style("left",(event.pageX+14)+"px").style("top",(event.pageY-28)+"px"))
+        .on("mouseleave", ()    => hideTip());
 
-    // Shape: diamond for dirs, rect for files
+    // ── Node shapes ────────────────────────────────────────────────────────────
     node.each(function(d) {
-      const sel = d3.select(this);
-      const color = nodeColor(d);
-      if (d.data.type === "root" || d.data.type === "dir") {
-        sel.append("polygon")
-          .attr("points", "-9,0 0,-9 9,0 0,9")
-          .attr("fill", color)
-          .attr("stroke", "#fff")
-          .attr("stroke-width", 1.5);
+      const s = d3.select(this), c = nodeColor(d);
+      if (isCol(d)) {
+        // Small filled circle for column nodes
+        s.append("circle").attr("r", 4).attr("fill", c).attr("stroke","#fff").attr("stroke-width",1);
+      } else if (isDir(d)) {
+        s.append("polygon").attr("points","-9,0 0,-9 9,0 0,9")
+          .attr("fill",c).attr("stroke","#fff").attr("stroke-width",1.5);
       } else {
-        sel.append("rect")
-          .attr("x", -7).attr("y", -7)
-          .attr("width", 14).attr("height", 14)
-          .attr("rx", 3)
-          .attr("fill", color)
-          .attr("stroke", "#fff")
-          .attr("stroke-width", 1.5);
+        s.append("rect").attr("x",-7).attr("y",-7).attr("width",14).attr("height",14)
+          .attr("rx",3).attr("fill",c).attr("stroke","#fff").attr("stroke-width",1.5);
       }
     });
 
-    // Collapse indicator (+ / -)
-    node.filter(d => d.data.children)
+    // ── Expand/collapse +/− badge (non-column expandable nodes only) ───────────
+    node.filter(d => d.data.children && !isCol(d))
       .append("text")
-        .attr("dy", "0.35em")
-        .attr("text-anchor", "middle")
-        .attr("fill", "#fff")
-        .attr("font-size", "11px")
-        .attr("font-weight", "700")
-        .attr("pointer-events", "none")
+        .attr("dy","0.35em").attr("text-anchor","middle")
+        .attr("fill","#fff").attr("font-size","11px").attr("font-weight","700")
+        .attr("pointer-events","none")
         .text(d => d.children ? "−" : "+");
 
-    // Labels
-    node.append("text")
-      .attr("dy", "0.35em")
-      .attr("x", d => (d.data.type === "root" || d.data.type === "dir") ? 14 : 12)
-      .attr("text-anchor", "start")
-      .attr("fill", d => (d.data.type === "root" || d.data.type === "dir") ? "#1a2233" : "#334")
-      .attr("font-weight", d => (d.data.type === "root" || d.data.type === "dir") ? "700" : "400")
-      .attr("font-size", "12.5px")
-      .text(d => d.data.name)
-      .clone(true).lower()
-        .attr("stroke","#f4f6f9")
-        .attr("stroke-width",3)
-        .attr("stroke-linejoin","round");
+    // ── Labels ─────────────────────────────────────────────────────────────────
+    node.each(function(d) {
+      const s      = d3.select(this);
+      const col_   = isCol(d);
+      const xOff   = col_ ? 8 : isDir(d) ? 14 : 12;
+      const fSize  = col_ ? "11px"  : "12.5px";
+      const fFam   = col_ ? "'Courier New', Courier, monospace" : null;
+      const fColor = col_ ? "#475569" : isDir(d) ? "#1a2233" : "#334";
+      const fWeight= (!col_ && isDir(d)) ? "700" : "400";
+      const name   = d.data.name;
 
-    // Description labels (grey, smaller)
-    node.append("text")
-      .attr("dy", "0.35em")
-      .attr("x", d => (d.data.type === "root" || d.data.type === "dir") ? 14 : 12)
-      .attr("dx", d => {
-        // approximate label width (7px/char)
-        const nameLen = (d.data.name || "").length;
-        return nameLen * 7 + 6;
-      })
-      .attr("text-anchor", "start")
-      .attr("fill", "#888")
-      .attr("font-size", "11px")
-      .text(d => "— " + d.data.desc)
-      .clone(true).lower()
-        .attr("stroke","#f4f6f9")
-        .attr("stroke-width",3)
-        .attr("stroke-linejoin","round");
+      // Readability halo
+      s.append("text")
+        .attr("dy","0.35em").attr("x",xOff).attr("text-anchor","start")
+        .attr("stroke","#f4f6f9").attr("stroke-width", col_ ? 2 : 3)
+        .attr("stroke-linejoin","round").attr("fill","none")
+        .attr("font-size",fSize).attr("font-family",fFam||null)
+        .text(name);
+
+      // Name label
+      s.append("text")
+        .attr("dy","0.35em").attr("x",xOff).attr("text-anchor","start")
+        .attr("fill",fColor).attr("font-weight",fWeight)
+        .attr("font-size",fSize).attr("font-family",fFam||null)
+        .text(name);
+
+      if (col_) return; // column nodes: name only, no extra labels
+
+      const nameW = labelPx(name, false);
+
+      // Description (grey, smaller)
+      s.append("text")
+        .attr("dy","0.35em").attr("x",xOff).attr("dx", nameW + 6)
+        .attr("text-anchor","start").attr("fill","#888").attr("font-size","11px")
+        .text("— " + d.data.desc)
+        .clone(true).lower()
+          .attr("stroke","#f4f6f9").attr("stroke-width",3)
+          .attr("stroke-linejoin","round").attr("fill","none");
+
+      // "▾ N cols" badge on collapsed CSV nodes
+      if (isCSV(d) && d.data._collapsed) {
+        const n = d.data.children.length;
+        const descW = labelPx("— " + d.data.desc, false);
+        s.append("text")
+          .attr("dy","0.35em").attr("x",xOff).attr("dx", nameW + 6 + descW + 10)
+          .attr("text-anchor","start")
+          .attr("fill","#4f8ef7").attr("font-size","10.5px").attr("font-weight","600")
+          .text(`[▾ ${n} cols]`);
+      }
+    });
 
-    // update container height
     document.getElementById("ds-tree-container").style.minHeight = svgHeight + "px";
   }
 
   update();
 
+  // ── Toolbar buttons ───────────────────────────────────────────────────────────
   document.getElementById("tree-expand-all").addEventListener("click", () => {
+    // Expand dirs/files but NOT CSV column children (too noisy)
     function expandAll(n) {
-      n.data._collapsed = false;
-      if (n.data.children) n.data.children.forEach(c => expandAll({data: c}));
+      if (n.type === "csv") return;   // leave CSV column children collapsed
+      n._collapsed = false;
+      if (n.children) n.children.forEach(expandAll);
     }
-    expandAll({data: treeData});
+    expandAll(treeData);
     update();
   });
+
   document.getElementById("tree-collapse-all").addEventListener("click", () => {
     function collapseAll(n, depth) {
-      if (depth > 0) n.data._collapsed = true;
-      if (n.data.children) n.data.children.forEach(c => collapseAll({data: c}, depth + 1));
+      if (depth > 0) n._collapsed = true;
+      if (n.children) n.children.forEach(c => collapseAll(c, depth + 1));
     }
-    collapseAll({data: treeData}, 0);
+    collapseAll(treeData, 0);
     update();
   });
 })();