# Content Pipeline v2 — Integration Test Guide
**Date:** 2026-04-22  
**Status:** Ready for testing  
**Prerequisites:** Backend running, Gaming PC online

---

## Test 1: API Connectivity

**Purpose:** Verify all endpoints respond

```bash
cd /home/hoffmann_admin/.openclaw/workspace-socrates/hoffdesk-api

# Start the API (if not running)
python -m uvicorn main:app --host 127.0.0.1 --port 8000 --reload

# In another terminal, test connectivity
curl -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/briefs

# Expected: {"briefs": [], "count": 0} or list of existing briefs
```

**✅ Pass:** HTTP 200 with JSON response
**❌ Fail:** Connection refused, 401 (check token), 500 (check logs)

---

## Test 2: Style Reference Endpoint

**Purpose:** Verify dns-night.json loads correctly

```bash
curl -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/style-references

# Expected: [{"slug": "dns-night", "title": "...", "excerpt": "..."}]
```

**✅ Pass:** Returns style example with correct metadata
**❌ Fail:** Empty array (check file path), 404 (check endpoint)

---

## Test 3: Create Brief

**Purpose:** Verify brief creation and validation

```bash
# Create test brief
curl -X POST \
  -H "Content-Type: application/json" \
  -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/briefs \
  -d '{
    "title": "Test: The Morning I Broke Coffee",
    "struggle_angle": "I broke the coffee maker at 6 AM when Aundrea needed caffeine",
    "origin_story": "It was Monday morning. Aundrea was already running late. I was 'just cleaning' the coffee maker...",
    "attempts": [
      {"attempt": "Took it apart to clean", "why_failed": "Couldn't remember how to reassemble"},
      {"attempt": "Watched YouTube tutorial", "why_failed": "Our model was different"}
    ],
    "the_moment": "I found the broken plastic piece under the sink. It had been broken the whole time.",
    "the_fix": "Bought a new coffee maker on Amazon Same-Day. Delivered by 9 AM.",
    "reflection": "Never disassemble critical appliances before morning coffee. Also: always have backup caffeine.",
    "style_reference": "dns-night",
    "target_length": 800
  }'

# Expected: {"brief_id": "...", "status": "draft", "created_at": "..."}
# Save the brief_id for next test
```

**✅ Pass:** Returns brief_id, status "draft"
**❌ Fail:** 400 validation error (check required fields), 500 (server error)

---

## Test 4: Submit for Approval

**Purpose:** Verify Telegram notification triggers

```bash
# Replace BRIEF_ID with actual ID from Test 3
curl -X POST \
  -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/briefs/BRIEF_ID/submit

# Expected: {"success": true, "status": "submitted"}
```

**✅ Pass:** Status changes to "submitted", you receive Telegram DM
**❌ Fail:** 400 (brief not complete), Telegram error (check token)

**Check Telegram:** You should receive DM:
> "New brief: Test: The Morning I Broke Coffee — Approve? [Yes] [No] [Edit]"

---

## Test 5: Approve Brief

**Purpose:** Verify generation triggers

```bash
# Option A: Click [Yes] in Telegram (recommended)
# Option B: API call
curl -X POST \
  -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/briefs/BRIEF_ID/approve

# Expected: {"success": true, "status": "generating"}
```

**✅ Pass:** Status changes to "generating"
**❌ Fail:** 403 (not authorized), Gaming PC unavailable

---

## Test 6: Monitor Generation

**Purpose:** Verify generation completes

```bash
# Poll every 10 seconds
curl -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/briefs/BRIEF_ID

# Watch status: generating → complete
```

**Expected timeline:**
- T+0s: Approved, generation started
- T+60-120s: Generation completes
- T+120s: Status "complete", output available

**✅ Pass:** Status becomes "complete", struggle_score present
**❌ Fail:** Stuck on "generating" (check Gaming PC), timeout (check model)

---

## Test 7: View Output

**Purpose:** Verify content generated with struggle-first voice

```bash
curl -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/briefs/BRIEF_ID/output

# Expected: {
#   "content": "# Test: The Morning...",
#   "html": "<h1>Test: The Morning...</h1>...",
#   "struggle_score": 82,
#   "interpretation": "Strong struggle narrative...",
#   "checks": {...}
# }
```

**✅ Pass:** Content present, struggle_score >75, first-person voice
**❌ Fail:** Low score (<50), generic tone, missing voice markers

---

## Test 8: End-to-End Live Script

**Purpose:** Full pipeline test with actual generation

```bash
cd /home/hoffmann_admin/.openclaw/workspace-socrates/hoffdesk-api

# Run live test (creates job, generates content, saves output)
python test_pipeline_live.py "How I learned to stop worrying and love local DNS"

# Check Gaming PC only
python test_pipeline_live.py --check-only
```

**Expected output:**
```
✅ Gaming PC available
Created job: uuid
📄 TITLE: [generated title]
📝 WORD COUNT: ~1500
⏱️ READING TIME: 6 min
✅ COMPLIANCE: replacements_made: 2, warnings: []
```

**✅ Pass:** Complete content, word count reasonable, saved to JSON
**❌ Fail:** Gaming PC unavailable, generation timeout, empty output

---

## Test 9: Web UI — StruggleBriefForm

**Purpose:** Verify frontend components render

**Steps:**
1. Open browser to `http://127.0.0.1:8000/admin/pipeline/new`
2. Verify: 4-step wizard visible
3. Fill Step 1 (title, struggle_angle, origin_story)
4. Click Continue → Step 2
5. Add 2 attempts, click Continue → Step 3
6. Fill moment/fix/reflection, click Continue → Step 4
7. Verify VoiceChecklist shows all 6 items
8. Check boxes → Submit enabled
9. Submit → "Brief Submitted" confirmation

**✅ Pass:** Form validates, wizard advances, checklist gates submit, API success
**❌ Fail:** 404 (templates not served), JavaScript error (check console), API 401/500

---

## Test 10: Web UI — BriefReview

**Purpose:** Verify approval UI and content preview

**Steps:**
1. After Test 4 (submitted brief), open `/admin/pipeline/brief/BRIEF_ID`
2. Verify: Status shows "Submitted — Awaiting Matt's approval"
3. Verify: [Approve & Generate] and [Reject] buttons visible
4. Click [Approve & Generate]
5. Verify: "Generating..." state appears
6. Wait 60-120s (or poll status)
7. Verify: Status "Complete", struggle score visible
8. Verify: Side-by-side preview (brief left, generated right)
9. Verify: [Create Post from Output] button

**✅ Pass:** Approval triggers generation, score displays, preview renders
**❌ Fail:** Approval fails, generation stuck, preview missing

---

## Success Criteria

| Test | Required | Score |
|------|----------|-------|
| API connectivity | ✅ | — |
| Brief creation | ✅ | — |
| Telegram approval | ✅ | — |
| Generation | ✅ | — |
| **Struggle score** | ✅ | **≥75** |
| First-person voice | ✅ | Detected |
| **Side-by-side preview** | ✅ | Renders |
| **3 successful posts** | ✅ | All score >75 |

---

## Troubleshooting

| Symptom | Likely Cause | Fix |
|---------|--------------|-----|
| 401 Unauthorized | Token mismatch | Check `X-Admin-Token: hoffdesk-admin-2025` |
| Gaming PC unavailable | Tailscale down | Check `ping matt-pc.tail864e81.ts.net` |
| Generation timeout | phi4:14b not loaded | `curl http://matt-pc.tail864e81.ts.net:11434/api/tags` |
| No Telegram DM | Bot token missing | Check `blog/notifications/telegram.py` config |
| Low struggle score | Prompt injection failed | Verify `dns-night.json` loads, check logs |
| Template 404 | Static files not served | Check `app.mount("/blog/static", ...)` |

---

## Quick Health Check

```bash
cd /home/hoffmann_admin/.openclaw/workspace-socrates/hoffdesk-api

# 1. API up?
curl -s http://127.0.0.1:8000/health || echo "❌ API down"

# 2. Gaming PC reachable?
curl -s http://matt-pc.tail864e81.ts.net:11434/api/tags | head -1 || echo "❌ Gaming PC unreachable"

# 3. Style example loads?
curl -s -H "X-Admin-Token: hoffdesk-admin-2025" \
  http://127.0.0.1:8000/api/v1/content/style-references | grep "dns-night" || echo "❌ Style not loaded"

echo "✅ All systems ready for testing"
```

---

**Ready to begin?** Start with Test 1 (API connectivity), or jump to Test 8 (live script) for immediate feedback.