Add AI research redline packet

ai-research-redline-packet/README.md

@@ -0,0 +1,69 @@
+# AI Research Redline Packet
+
+This module adds a focused AI-assisted research-tool slice for SCIBASE. It does
+not try to be another broad paper summarizer or generic citation formatter.
+Instead, it produces an evidence-backed pre-submission redline packet that tells
+researchers what must be fixed before sharing a manuscript.
+
+The workflow is deterministic and dependency-free, making it suitable for local
+review checks, future API handlers, or reviewer-facing demo artifacts.
+
+## What It Covers
+
+- Generates abstract, executive, and layperson summary deltas from a structured
+  manuscript sample.
+- Maps manuscript claims to evidence spans and flags unsupported claims.
+- Detects statistical redlines such as missing confidence intervals and
+  inconsistent p-value language.
+- Checks compliance items such as ethics statements and data availability.
+- Ranks citation gap recommendations with transparent reasons.
+- Routes issues to reviewer templates and creates insertion-ready revision
+  tasks.
+- Includes tests, synthetic sample data, a CLI demo, requirement mapping, and a
+  short demo video.
+
+## Quick Start
+
+```bash
+npm run check
+npm test
+npm run demo
+```
+
+Expected demo summary:
+
+```text
+Packet decision: revise-before-submit
+Redlines: 5
+Revision tasks: 5
+Citation recommendations: 3
+```
+
+## Repository Layout
+
+```text
+ai-research-redline-packet/
+  data/sample-manuscript.json
+  docs/demo.svg
+  docs/demo.mp4
+  docs/requirement-map.md
+  scripts/demo.js
+  src/redline-packet.js
+  test/redline-packet.test.js
+```
+
+## Design Notes
+
+The redline packet sits between AI summarization, AI peer review, and AI
+citation assistance:
+
+1. A manuscript is parsed into claims, evidence spans, results, compliance
+   statements, and citations.
+2. Summary modes are generated from the same evidence inventory.
+3. Unsupported claims, statistical issues, compliance gaps, and citation gaps
+   become reviewer-facing redlines.
+4. Redlines are converted into concrete revision tasks with insertion targets.
+5. The final packet has an audit digest so reviewers can reproduce the decision.
+
+This gives researchers an actionable pre-review artifact rather than a loose AI
+chat transcript.

ai-research-redline-packet/data/sample-manuscript.json

@@ -0,0 +1,168 @@
+{
+  "schemaVersion": "ai-redline-packet.v1",
+  "manuscript": {
+    "id": "ms-neuro-organoid-response",
+    "title": "Transcriptomic Response of Neural Organoids to Microfluidic Oxygen Gradients",
+    "domain": "biology",
+    "targetJournal": "Nature Biomedical Engineering",
+    "stage": "pre-submission",
+    "sections": [
+      {
+        "id": "abstract",
+        "heading": "Abstract",
+        "text": "We report a microfluidic oxygen-gradient platform for neural organoids. Treated organoids showed increased maturation markers and improved viability after seven days."
+      },
+      {
+        "id": "results",
+        "heading": "Results",
+        "text": "Marker GFAP increased by 22 percent after treatment. Organoids had significantly better viability with p = 0.08. The platform eliminates batch effects across all donor lines."
+      },
+      {
+        "id": "methods",
+        "heading": "Methods",
+        "text": "Organoids from three donor lines were cultured in a controlled oxygen gradient. Analysis code and raw count tables were deposited with the project repository."
+      },
+      {
+        "id": "ethics",
+        "heading": "Ethics",
+        "text": "Human induced pluripotent stem cell lines were obtained from a certified biobank."
+      }
+    ],
+    "claims": [
+      {
+        "id": "claim-marker-maturation",
+        "text": "GFAP expression increased by 22 percent after seven days of treatment.",
+        "sectionId": "results",
+        "evidenceSpanIds": ["span-gfap-table"],
+        "citationIds": ["smith-2024-organoid-maturation"],
+        "importance": "key-finding"
+      },
+      {
+        "id": "claim-viability-significance",
+        "text": "Treated organoids had significantly better viability.",
+        "sectionId": "results",
+        "evidenceSpanIds": ["span-viability-pvalue"],
+        "citationIds": [],
+        "importance": "key-finding"
+      },
+      {
+        "id": "claim-eliminates-batch-effects",
+        "text": "The platform eliminates batch effects across all donor lines.",
+        "sectionId": "results",
+        "evidenceSpanIds": [],
+        "citationIds": [],
+        "importance": "broad-claim"
+      },
+      {
+        "id": "claim-data-deposited",
+        "text": "Analysis code and raw count tables were deposited with the project repository.",
+        "sectionId": "methods",
+        "evidenceSpanIds": ["span-repository-link"],
+        "citationIds": ["project-repository-v2"],
+        "importance": "reproducibility"
+      }
+    ],
+    "evidenceSpans": [
+      {
+        "id": "span-gfap-table",
+        "sectionId": "results",
+        "quote": "Marker GFAP increased by 22 percent after treatment.",
+        "artifact": "results/gfap-differential-expression.csv",
+        "confidence": 0.91
+      },
+      {
+        "id": "span-viability-pvalue",
+        "sectionId": "results",
+        "quote": "Organoids had significantly better viability with p = 0.08.",
+        "artifact": "results/viability-analysis.json",
+        "confidence": 0.72
+      },
+      {
+        "id": "span-repository-link",
+        "sectionId": "methods",
+        "quote": "Analysis code and raw count tables were deposited with the project repository.",
+        "artifact": "metadata/repository-export.json",
+        "confidence": 0.86
+      }
+    ],
+    "statisticalChecks": [
+      {
+        "id": "stat-viability-pvalue",
+        "claimId": "claim-viability-significance",
+        "metric": "viability",
+        "pValue": 0.08,
+        "claimsSignificant": true,
+        "confidenceInterval": null,
+        "sampleSize": 3
+      },
+      {
+        "id": "stat-gfap-effect",
+        "claimId": "claim-marker-maturation",
+        "metric": "GFAP expression",
+        "pValue": 0.012,
+        "claimsSignificant": true,
+        "confidenceInterval": [0.14, 0.31],
+        "sampleSize": 3
+      }
+    ],
+    "compliance": {
+      "ethicsStatementPresent": true,
+      "irbProtocolId": null,
+      "dataAvailabilityStatementPresent": true,
+      "codeAvailabilityStatementPresent": true,
+      "conflictOfInterestPresent": false,
+      "fundingStatementPresent": true
+    },
+    "citations": [
+      {
+        "id": "smith-2024-organoid-maturation",
+        "title": "Organoid maturation markers under controlled oxygen gradients",
+        "year": 2024,
+        "style": "Nature",
+        "supportsClaimIds": ["claim-marker-maturation"]
+      },
+      {
+        "id": "project-repository-v2",
+        "title": "SCIBASE project repository export preprint-v2",
+        "year": 2026,
+        "style": "DataCite",
+        "supportsClaimIds": ["claim-data-deposited"]
+      }
+    ],
+    "candidateCitations": [
+      {
+        "id": "lee-2025-donor-batch-effects",
+        "title": "Donor-line batch effects in neural organoid transcriptomics",
+        "year": 2025,
+        "reason": "Needed before making a broad batch-effect claim.",
+        "claimIds": ["claim-eliminates-batch-effects"]
+      },
+      {
+        "id": "nguyen-2024-viability-statistics",
+        "title": "Statistical reporting standards for organoid viability assays",
+        "year": 2024,
+        "reason": "Supports corrected viability reporting and confidence intervals.",
+        "claimIds": ["claim-viability-significance"]
+      },
+      {
+        "id": "icmje-2025-ethics-reporting",
+        "title": "Research ethics and disclosure reporting guidance",
+        "year": 2025,
+        "reason": "Helps complete ethics protocol and disclosure statements.",
+        "claimIds": []
+      }
+    ]
+  },
+  "reviewTemplates": [
+    {
+      "id": "biology-pre-submission",
+      "name": "Biology pre-submission review",
+      "requiredTopics": ["evidence", "statistics", "ethics", "data-availability", "citations"]
+    },
+    {
+      "id": "statistical-methods",
+      "name": "Statistical methods review",
+      "requiredTopics": ["statistics", "sample-size", "effect-size"]
+    }
+  ]
+}

ai-research-redline-packet/docs/requirement-map.md

@@ -0,0 +1,40 @@
+# Requirement Map
+
+Issue: SCIBASE-AI/SCIBASE.AI#13, AI-Assisted Research Tools (MVP Level).
+
+## AI Paper Summarizer
+
+The packet emits abstract, executive, and layperson summary modes from the same
+structured manuscript evidence. It also reports summary deltas such as how many
+key findings, reproducibility claims, and broad claims require review.
+
+## AI Peer Review Aid
+
+The redline engine flags unsupported claims, statistical reporting issues,
+missing confidence intervals, ethics protocol gaps, and missing disclosure
+statements. Each redline carries severity, insertion target, and evidence digest
+so reviewers can audit why it was raised.
+
+## AI Citation Tool
+
+Citation recommendations are ranked against the redlines they would help fix.
+The output includes transparent reasons and claim ids, making citations
+insertion-ready rather than opaque suggestions.
+
+## Review Templates
+
+The packet routes redlines to domain-specific reviewer templates such as biology
+pre-submission review and statistical methods review. This maps AI findings into
+review workflows that institutions can standardize.
+
+## Actionable Output
+
+Every redline becomes a revision task with an insertion target and action text.
+This supports one-click manuscript editing or PR-like review queues in a future
+SCIBASE interface.
+
+## Scope Boundary
+
+This is not another generic summarizer or chatbot transcript. It is a
+deterministic pre-submission redline packet that ties summaries, peer-review
+diagnostics, citation recommendations, and revision tasks together.

ai-research-redline-packet/package.json

@@ -0,0 +1,15 @@
+{
+  "name": "ai-research-redline-packet",
+  "version": "0.1.0",
+  "description": "Dependency-free evidence redline packet for AI-assisted research review workflows.",
+  "private": true,
+  "scripts": {
+    "check": "node scripts/demo.js --json > /dev/null",
+    "demo": "node scripts/demo.js",
+    "test": "node --test test/*.test.js"
+  },
+  "engines": {
+    "node": ">=18"
+  },
+  "license": "MIT"
+}

ai-research-redline-packet/scripts/demo.js

@@ -0,0 +1,17 @@
+const fs = require("node:fs");
+const path = require("node:path");
+const { createRedlinePacket, formatPacketReport } = require("../src/redline-packet");
+
+const manifestPath = path.join(__dirname, "..", "data", "sample-manuscript.json");
+const manifest = JSON.parse(fs.readFileSync(manifestPath, "utf8"));
+const packet = createRedlinePacket(manifest);
+
+if (process.argv.includes("--json")) {
+  process.stdout.write(`${JSON.stringify(packet, null, 2)}\n`);
+} else {
+  process.stdout.write(`${formatPacketReport(packet)}\n`);
+}
+
+if (packet.decision === "invalid-manifest") {
+  process.exitCode = 1;
+}