# Deep Analysis - Behavioral Tests
Workflows for verifying the `thoughtbox_gateway` deep analysis operation.
**Tool:** `thoughtbox_gateway`
**Operation:** `deep_analysis`
**Required stage:** Stage 1 (init_complete)
**Analysis types:** `patterns`, `cognitive_load`, `decision_points`, `full`
**Options:** `includeTimeline`, `compareWith`
---
## DA-001: Patterns Analysis
**Goal:** Verify `patterns` analysis type returns statistical metrics.
**Prerequisite:** Stage 1, a session with thoughts/revisions/branches exists.
**Steps:**
1. Advance to Stage 1 via `start_new` or `load_context`
2. Call `{ operation: "deep_analysis", args: { sessionId: "<id>", analysisType: "patterns" } }`
3. Verify response includes:
- `totalThoughts` (integer ≥ 1)
- `revisionCount` (integer ≥ 0)
- `branchCount` (integer ≥ 0)
- `averageThoughtLength` (number > 0)
4. Verify counts match expected values for the session
**Expected:** Accurate statistical summary of session structure
---
## DA-002: Cognitive Load Analysis
**Goal:** Verify `cognitive_load` analysis type returns complexity indicators.
**Prerequisite:** Stage 1, a session exists.
**Steps:**
1. Advance to Stage 1
2. Call `{ operation: "deep_analysis", args: { sessionId: "<id>", analysisType: "cognitive_load" } }`
3. Verify response includes:
- `complexityScore` (number 0–100)
- `depthIndicator` (max thought number in session)
- `breadthIndicator` (count of unique branches)
4. Verify complexityScore is within 0–100 range
5. Verify depthIndicator ≥ 1
6. Verify breadthIndicator ≥ 0
**Expected:** Meaningful complexity assessment with bounded score
---
## DA-003: Decision Points Analysis
**Goal:** Verify `decision_points` analysis type returns array of revisions and branches with references.
**Prerequisite:** Stage 1, a session with at least one revision OR one branch.
**Steps:**
1. Advance to Stage 1
2. Use a session known to have revisions and branches
3. Call `{ operation: "deep_analysis", args: { sessionId: "<id>", analysisType: "decision_points" } }`
4. Verify response includes array of decision points
5. Each decision point should have:
- `thoughtNumber` (integer)
- `type` (`"revision"` or `"branch"`)
- `reference` (thought number being revised or branched from)
6. Verify revisions have type `"revision"` and branches have type `"branch"`
**Expected:** Each point where reasoning diverged is identified with type and reference
---
## DA-004: Full Analysis
**Goal:** Verify `full` analysis type combines all analysis types.
**Prerequisite:** Stage 1, a session exists.
**Steps:**
1. Advance to Stage 1
2. Call `{ operation: "deep_analysis", args: { sessionId: "<id>", analysisType: "full" } }`
3. Verify response includes all fields from DA-001 (patterns):
- `totalThoughts`, `revisionCount`, `branchCount`, `averageThoughtLength`
4. Verify response includes all fields from DA-002 (cognitive load):
- `complexityScore`, `depthIndicator`, `breadthIndicator`
5. Verify response includes all fields from DA-003 (decision points):
- Decision points array
**Expected:** Superset of all three individual analysis types
---
## DA-005: Include Timeline Option
**Goal:** Verify `options.includeTimeline` adds temporal data.
**Prerequisite:** Stage 1, a session exists.
**Steps:**
1. Advance to Stage 1
2. Call `{ operation: "deep_analysis", args: { sessionId: "<id>", analysisType: "full", options: { includeTimeline: true } } }`
3. Verify response includes timeline data:
- `createdAt` (session creation timestamp)
- `updatedAt` (last modification timestamp)
- `durationEstimate` (estimated session duration)
4. Call same analysis WITHOUT `includeTimeline`
5. Verify timeline fields are absent (or null) without the option
**Expected:** Timeline data included only when requested
---
## DA-006: Invalid Session ID
**Goal:** Verify clear error for nonexistent session.
**Steps:**
1. Advance to Stage 1
2. Call `{ operation: "deep_analysis", args: { sessionId: "nonexistent-session-xyz", analysisType: "patterns" } }`
3. Verify error response (not a crash)
4. Verify error message indicates session not found
5. Verify error is specific (not a generic 500)
**Expected:** Clear "session not found" error with the invalid ID echoed back
---
## Running These Tests
All deep analysis tests require Stage 1. Execute by calling `thoughtbox_gateway` with `operation: "deep_analysis"`.
**Setup:** Create a test session with thoughts, revisions, and branches before running these tests. Ideally:
- At least 5 thoughts
- At least 1 revision (isRevision: true, revisesThought: N)
- At least 1 branch (branchFromThought: N, branchId: "test-branch")
