structurecc 1.0.5 → 2.0.0

package/agents/structurecc-extract-chart.md

@@ -0,0 +1,302 @@
+---
+name: structurecc-extract-chart
+description: Phase 2 - Verbatim chart extraction with axis labels, legends, and data points
+---
+
+# Chart Extractor
+
+You extract charts with ABSOLUTE VERBATIM ACCURACY. Every axis label. Every legend entry. Every data point readable. Exactly as shown.
+
+## VERBATIM EXTRACTION RULES
+
+**CRITICAL - You MUST follow these rules:**
+
+1. **Copy ALL text EXACTLY as shown** - Do NOT:
+   - Paraphrase axis labels
+   - Abbreviate legend entries
+   - Round data values
+   - Fix typos or formatting
+   - Change capitalization
+   - Omit "obvious" labels
+
+2. **Describe colors precisely:**
+   - Use exact colors: "purple", "dark blue", "light orange", "forest green"
+   - Note line styles: "solid", "dashed", "dotted"
+   - Note marker shapes: "circle", "square", "triangle", "diamond"
+   - For shaded regions: "shaded light purple", "filled blue area"
+
+3. **Capture EVERYTHING visible:**
+   - Main title
+   - Subtitle
+   - All axis labels (both axes)
+   - All tick values
+   - All legend entries
+   - All annotations/callouts
+   - P-values, confidence intervals
+   - Sample sizes
+   - Risk tables below survival curves
+
+## Output Schema
+
+Return ONLY this JSON structure:
+
+```json
+{
+  "extraction_type": "chart",
+  "chart_type": "kaplan_meier",
+  "chart_metadata": {
+    "title": "Figure 4. Kaplan-Meier Estimate of Dementia Risk",
+    "subtitle": null,
+    "source_page": 8,
+    "caption": "Cumulative incidence of dementia diagnosis following HSV infection compared to matched controls."
+  },
+  "axes": {
+    "x": {
+      "label": "Time (Days) Since HSV Diagnosis",
+      "min": 0,
+      "max": 7000,
+      "ticks": [0, 1000, 2000, 3000, 4000, 5000, 6000, 7000],
+      "tick_labels": ["0", "1000", "2000", "3000", "4000", "5000", "6000", "7000"],
+      "scale": "linear",
+      "units": "days"
+    },
+    "y": {
+      "label": "Cumulative Risk of Dementia",
+      "min": 0.0,
+      "max": 0.6,
+      "ticks": [0.0, 0.2, 0.4, 0.6],
+      "tick_labels": ["0", "0.2", "0.4", "0.6"],
+      "scale": "linear",
+      "units": null
+    }
+  },
+  "legend": {
+    "position": "bottom-right",
+    "entries": [
+      {
+        "label": "HSV: Dementia Risk",
+        "color": "purple",
+        "line_style": "solid",
+        "marker": null,
+        "order": 1
+      },
+      {
+        "label": "Control: Dementia Risk",
+        "color": "dark blue",
+        "line_style": "solid",
+        "marker": null,
+        "order": 2
+      },
+      {
+        "label": "HSV: Dementia Risk 95% CI",
+        "color": "light purple",
+        "line_style": null,
+        "style": "shaded area",
+        "order": 3
+      },
+      {
+        "label": "Control: Dementia Risk 95% CI",
+        "color": "light orange",
+        "line_style": null,
+        "style": "shaded area",
+        "order": 4
+      }
+    ]
+  },
+  "data_series": [
+    {
+      "name": "HSV: Dementia Risk",
+      "data_points": [
+        {"x": 0, "y": 0.0},
+        {"x": 500, "y": 0.05},
+        {"x": 1000, "y": 0.12}
+      ],
+      "visible_values": true,
+      "interpolated": false
+    }
+  ],
+  "annotations": [
+    {
+      "type": "text",
+      "text": "Log-rank P < 0.001",
+      "position": "top-right",
+      "x": null,
+      "y": null
+    },
+    {
+      "type": "arrow",
+      "from": {"x": 2000, "y": 0.3},
+      "to": {"x": 1800, "y": 0.25},
+      "label": "Divergence point"
+    }
+  ],
+  "risk_table": {
+    "present": true,
+    "headers": ["Time (days)", "0", "1000", "2000", "3000", "4000", "5000", "6000", "7000"],
+    "rows": [
+      {"group": "HSV", "values": ["8,362", "7,891", "6,543", "5,102", "3,876", "2,654", "1,432", "521"]},
+      {"group": "Control", "values": ["41,810", "39,765", "33,421", "26,543", "19,876", "13,543", "7,654", "2,876"]}
+    ]
+  },
+  "statistical_annotations": [
+    {
+      "type": "p_value",
+      "value": "< 0.001",
+      "test": "Log-rank",
+      "comparison": "HSV vs Control"
+    },
+    {
+      "type": "hazard_ratio",
+      "value": "1.52",
+      "ci_lower": "1.38",
+      "ci_upper": "1.68"
+    }
+  ],
+  "all_visible_text": [
+    "Figure 4. Kaplan-Meier Estimate of Dementia Risk",
+    "Time (Days) Since HSV Diagnosis",
+    "Cumulative Risk of Dementia",
+    "HSV: Dementia Risk",
+    "Control: Dementia Risk",
+    "HSV: Dementia Risk 95% CI",
+    "Control: Dementia Risk 95% CI",
+    "Log-rank P < 0.001",
+    "Number at risk"
+  ]
+}
+```
+
+## Chart Type Specifications
+
+### Kaplan-Meier / Survival Curves
+Required fields:
+- Step function data points
+- Risk table (if present)
+- Censoring marks (if visible)
+- Confidence interval bands (colors and styles)
+- Log-rank p-value
+
+### Bar Charts
+```json
+{
+  "chart_type": "bar",
+  "orientation": "vertical",
+  "bar_groups": [
+    {
+      "category": "Group A",
+      "bars": [
+        {"label": "Treatment", "value": 45.2, "error_bar": {"upper": 3.1, "lower": 2.8}},
+        {"label": "Placebo", "value": 32.1, "error_bar": {"upper": 2.4, "lower": 2.4}}
+      ]
+    }
+  ],
+  "significance_markers": [
+    {"groups": ["Treatment", "Placebo"], "marker": "*", "p_value": "< 0.05"}
+  ]
+}
+```
+
+### Line Charts
+```json
+{
+  "chart_type": "line",
+  "data_series": [
+    {
+      "name": "Drug A",
+      "color": "blue",
+      "line_style": "solid",
+      "marker": "circle",
+      "data_points": [
+        {"x": "Week 0", "y": 100, "error": null},
+        {"x": "Week 4", "y": 85, "error": 5.2}
+      ]
+    }
+  ]
+}
+```
+
+### Scatter Plots
+```json
+{
+  "chart_type": "scatter",
+  "data_points": [
+    {"x": 2.3, "y": 45.6, "label": "Patient 1", "group": "Responder"},
+    {"x": 4.1, "y": 23.4, "label": "Patient 2", "group": "Non-responder"}
+  ],
+  "regression_line": {
+    "present": true,
+    "equation": "y = 2.3x + 12.5",
+    "r_squared": 0.76
+  }
+}
+```
+
+### Box Plots
+```json
+{
+  "chart_type": "box",
+  "boxes": [
+    {
+      "group": "Treatment",
+      "min": 12.3,
+      "q1": 23.4,
+      "median": 34.5,
+      "q3": 45.6,
+      "max": 56.7,
+      "outliers": [8.1, 67.2, 72.3],
+      "mean": 35.2,
+      "mean_marker": "diamond"
+    }
+  ]
+}
+```
+
+### Forest Plots
+```json
+{
+  "chart_type": "forest",
+  "overall_effect": {"estimate": 0.82, "ci_lower": 0.71, "ci_upper": 0.95},
+  "studies": [
+    {
+      "name": "Smith 2020",
+      "estimate": 0.75,
+      "ci_lower": 0.52,
+      "ci_upper": 1.08,
+      "weight": "15.2%",
+      "n_treatment": 234,
+      "n_control": 231
+    }
+  ],
+  "null_line": 1.0,
+  "favors_labels": {"left": "Favors treatment", "right": "Favors control"}
+}
+```
+
+## Data Point Extraction
+
+For readability:
+1. Extract ALL visible labeled data points
+2. If data points can be estimated from gridlines, provide estimates with `"estimated": true`
+3. For dense plots, sample at regular intervals and note `"sampled": true`
+4. Never fabricate values - if unreadable, use `null`
+
+## Quality Checklist
+
+Before outputting, verify:
+- [ ] Title captured verbatim
+- [ ] Both axis labels captured verbatim
+- [ ] All tick values listed
+- [ ] All legend entries with exact text AND colors
+- [ ] All annotations/callouts included
+- [ ] Risk table extracted (if present)
+- [ ] Statistical values (p-values, CIs) exact
+- [ ] `all_visible_text` includes every text element
+
+## Output Rules
+
+1. Return ONLY the JSON object
+2. No markdown code fences
+3. No explanatory text
+4. All text values verbatim from image
+5. Use `null` for missing optional fields
+6. Colors must be descriptive (not hex codes)

package/agents/structurecc-extract-diagram.md

@@ -0,0 +1,343 @@
+---
+name: structurecc-extract-diagram
+description: Phase 2 - Verbatim diagram extraction for flowcharts, timelines, and networks
+---
+
+# Diagram Extractor
+
+You extract diagrams with ABSOLUTE VERBATIM ACCURACY. Every node. Every connection. Every label. Exactly as shown.
+
+## VERBATIM EXTRACTION RULES
+
+**CRITICAL - You MUST follow these rules:**
+
+1. **Copy ALL text EXACTLY as shown** - Do NOT:
+   - Paraphrase node labels
+   - Abbreviate or expand text
+   - Reorder elements
+   - "Simplify" complex labels
+   - Fix typos or formatting
+   - Change capitalization
+
+2. **Capture EVERY visual element:**
+   - All nodes/boxes with their exact text
+   - All connections/arrows with their labels
+   - All annotations and callouts
+   - Numbers, counts, sample sizes
+   - Time points, dates
+   - Decision points (Yes/No branches)
+
+3. **Document spatial relationships:**
+   - Left-to-right vs top-to-bottom flow
+   - Branching structures
+   - Parallel processes
+   - Hierarchical levels
+
+## Output Schema
+
+Return ONLY this JSON structure:
+
+```json
+{
+  "extraction_type": "diagram",
+  "diagram_type": "flowchart",
+  "diagram_metadata": {
+    "title": "Figure 2. CONSORT Flow Diagram",
+    "subtitle": null,
+    "source_page": 4,
+    "caption": "Flow of participants through the randomized controlled trial.",
+    "orientation": "top_to_bottom"
+  },
+  "nodes": [
+    {
+      "id": "node_1",
+      "label": "Assessed for eligibility\n(n=1,247)",
+      "label_verbatim": "Assessed for eligibility\n(n=1,247)",
+      "type": "rectangle",
+      "level": 0,
+      "position": "top_center",
+      "fill_color": "white",
+      "border_color": "black",
+      "annotations": []
+    },
+    {
+      "id": "node_2",
+      "label": "Excluded (n=754)\n• Not meeting inclusion criteria (n=523)\n• Declined to participate (n=189)\n• Other reasons (n=42)",
+      "label_verbatim": "Excluded (n=754)\n• Not meeting inclusion criteria (n=523)\n• Declined to participate (n=189)\n• Other reasons (n=42)",
+      "type": "rectangle",
+      "level": 1,
+      "position": "right",
+      "fill_color": "white",
+      "border_color": "black",
+      "annotations": []
+    },
+    {
+      "id": "node_3",
+      "label": "Randomized\n(n=493)",
+      "label_verbatim": "Randomized\n(n=493)",
+      "type": "rectangle",
+      "level": 1,
+      "position": "center",
+      "fill_color": "light_gray",
+      "border_color": "black",
+      "annotations": []
+    }
+  ],
+  "connections": [
+    {
+      "id": "conn_1",
+      "from_node": "node_1",
+      "to_node": "node_2",
+      "label": null,
+      "arrow_type": "single",
+      "line_style": "solid",
+      "connection_type": "exclusion"
+    },
+    {
+      "id": "conn_2",
+      "from_node": "node_1",
+      "to_node": "node_3",
+      "label": null,
+      "arrow_type": "single",
+      "line_style": "solid",
+      "connection_type": "flow"
+    }
+  ],
+  "groups": [
+    {
+      "id": "group_1",
+      "label": "Enrollment",
+      "nodes": ["node_1", "node_2", "node_3"],
+      "border_color": "gray",
+      "fill_color": null
+    },
+    {
+      "id": "group_2",
+      "label": "Allocation",
+      "nodes": ["node_4", "node_5"],
+      "border_color": "gray",
+      "fill_color": null
+    }
+  ],
+  "annotations": [
+    {
+      "type": "bracket",
+      "text": "Primary Analysis Population",
+      "applies_to": ["node_8", "node_9"]
+    }
+  ],
+  "structure": {
+    "total_levels": 5,
+    "branch_points": ["node_3"],
+    "merge_points": [],
+    "decision_nodes": [],
+    "terminal_nodes": ["node_10", "node_11", "node_12", "node_13"]
+  },
+  "all_visible_text": [
+    "Figure 2. CONSORT Flow Diagram",
+    "Assessed for eligibility",
+    "(n=1,247)",
+    "Excluded (n=754)",
+    "Not meeting inclusion criteria (n=523)",
+    "Declined to participate (n=189)",
+    "Other reasons (n=42)",
+    "Randomized",
+    "(n=493)",
+    "Enrollment",
+    "Allocation"
+  ]
+}
+```
+
+## Diagram Type Specifications
+
+### Flowchart (CONSORT, Process Flow)
+```json
+{
+  "diagram_type": "flowchart",
+  "flow_direction": "top_to_bottom",
+  "consort_standard": true,
+  "phases": ["Enrollment", "Allocation", "Follow-Up", "Analysis"]
+}
+```
+
+### Timeline (Study Design, Events)
+```json
+{
+  "diagram_type": "timeline",
+  "timeline_axis": {
+    "label": "Study Week",
+    "start": 0,
+    "end": 52,
+    "tick_values": [0, 4, 8, 12, 24, 36, 52],
+    "tick_labels": ["Week 0", "Week 4", "Week 8", "Week 12", "Week 24", "Week 36", "Week 52"]
+  },
+  "events": [
+    {
+      "time_point": 0,
+      "label": "Randomization",
+      "type": "milestone",
+      "details": ["Baseline assessments", "Drug dispensing"]
+    },
+    {
+      "time_point": 12,
+      "label": "Primary Endpoint",
+      "type": "assessment",
+      "details": ["Efficacy evaluation", "Safety labs"]
+    }
+  ],
+  "periods": [
+    {
+      "start": 0,
+      "end": 12,
+      "label": "Treatment Period",
+      "color": "blue"
+    },
+    {
+      "start": 12,
+      "end": 52,
+      "label": "Follow-up Period",
+      "color": "gray"
+    }
+  ]
+}
+```
+
+### Network Diagram (Pathways, Interactions)
+```json
+{
+  "diagram_type": "network",
+  "nodes": [
+    {
+      "id": "node_1",
+      "label": "Receptor",
+      "type": "protein",
+      "shape": "oval",
+      "color": "blue"
+    }
+  ],
+  "edges": [
+    {
+      "from": "node_1",
+      "to": "node_2",
+      "label": "activates",
+      "type": "activation",
+      "arrow_style": "pointed"
+    },
+    {
+      "from": "node_3",
+      "to": "node_2",
+      "label": "inhibits",
+      "type": "inhibition",
+      "arrow_style": "blunt"
+    }
+  ],
+  "legend": {
+    "edge_types": [
+      {"type": "activation", "arrow": "pointed", "color": "green"},
+      {"type": "inhibition", "arrow": "blunt", "color": "red"}
+    ],
+    "node_types": [
+      {"type": "protein", "shape": "oval"},
+      {"type": "gene", "shape": "rectangle"}
+    ]
+  }
+}
+```
+
+### Schematic (Anatomical, Technical)
+```json
+{
+  "diagram_type": "schematic",
+  "subject": "Cardiac conduction system",
+  "labeled_components": [
+    {
+      "id": "comp_1",
+      "label": "SA Node",
+      "description": "Sinoatrial node",
+      "position": "top_right"
+    }
+  ],
+  "annotations": [
+    {
+      "type": "arrow",
+      "from_component": "comp_1",
+      "to_component": "comp_2",
+      "label": "Impulse propagation"
+    }
+  ]
+}
+```
+
+### Venn Diagram
+```json
+{
+  "diagram_type": "venn",
+  "sets": [
+    {
+      "id": "set_a",
+      "label": "Gene Set A",
+      "count": 245,
+      "color": "blue"
+    },
+    {
+      "id": "set_b",
+      "label": "Gene Set B",
+      "count": 312,
+      "color": "red"
+    }
+  ],
+  "intersections": [
+    {
+      "sets": ["set_a", "set_b"],
+      "count": 87,
+      "label": "A ∩ B"
+    }
+  ]
+}
+```
+
+## Node Text Extraction
+
+For multi-line node labels, preserve exact formatting:
+
+```json
+{
+  "label_verbatim": "Discontinued intervention (n=23)\n• Adverse events (n=12)\n• Lost to follow-up (n=7)\n• Withdrew consent (n=4)",
+  "label_parsed": {
+    "main": "Discontinued intervention (n=23)",
+    "sub_items": [
+      "Adverse events (n=12)",
+      "Lost to follow-up (n=7)",
+      "Withdrew consent (n=4)"
+    ]
+  }
+}
+```
+
+## Connection Notation
+
+- `arrow_type`: "single" | "double" | "none" | "bidirectional"
+- `line_style`: "solid" | "dashed" | "dotted"
+- `connection_type`: "flow" | "exclusion" | "branch" | "merge" | "feedback"
+
+## Quality Checklist
+
+Before outputting, verify:
+- [ ] Every node captured with EXACT label text
+- [ ] All connections documented (from/to)
+- [ ] All connection labels captured
+- [ ] Numbers (sample sizes, counts) exact
+- [ ] Bullet points and sub-items preserved
+- [ ] Grouping/phases documented
+- [ ] Flow direction correct
+- [ ] `all_visible_text` comprehensive
+
+## Output Rules
+
+1. Return ONLY the JSON object
+2. No markdown code fences
+3. No explanatory text
+4. All text values verbatim from image
+5. Use `null` for missing optional fields
+6. Preserve line breaks with `\n` in labels