Test Dashboard

“Yeah, I’m gonna need you to go ahead and come in on Saturday…”

TPS = Testing Painful Stuff

Because extracting text from a .doc file from 1997 is painful. We test it so you don’t have to debug it.

The Numbers

53

Tests Passing

Unit tests covering all 20 tools

99.3%

Torture Test Success

299/301 random Office documents

20%

Code Coverage

Focus on critical paths

1.39s

Test Suite Runtime

Fast feedback loop

---

Torture Test Results

We grabbed 301 random Office documents from a real filesystem and threw them at mcwaddams.

By Format

Format	Tested	Passed	Failed	Success Rate
.docx	142	142	0	100%
.xlsx	89	89	0	100%
.pptx	34	34	0	100%
.doc	18	17	1*	94.4%
.xls	12	12	0	100%
.ppt	4	4	0	100%
.csv	2	1	1*	50%

*Failed files were empty (0 bytes) or corrupt. Not extraction failures.

What We Found

• 1,293 resources indexed across all documents
• Zero crashes — Every file handled gracefully
• Clear error messages for the 2 failures
• Average extraction time: 0.045s per document

---

Test Categories

Unit Tests (53)

Category	Count	Description
Universal Tools	12	extract_text, extract_images, detect_format, etc.
Word Tools	18	convert_to_markdown, extract_tables, structure analysis
Excel Tools	9	analyze_data, extract_formulas, chart generation
MCP Resources	8	Resource store, URI parsing, format conversion
Validation	6	File validation, error handling, edge cases

What We Test

1. Happy Path — Normal documents extract correctly
2. Legacy Formats — .doc, .xls, .ppt from the basement
3. Large Documents — Pagination triggers at 25k tokens
4. Malformed Files — Graceful errors, no crashes
5. Edge Cases — Empty files, Unicode, special characters
6. URL Processing — HTTP downloads, caching

---

Run Tests Yourself

# Clone the repo
git clone https://github.com/ryanmalloy/mcwaddams.git
cd mcwaddams

# Install dev dependencies
uv sync --dev

# Run tests
uv run pytest

# With coverage
uv run pytest --cov=mcwaddams

Makefile Shortcuts

make test          # Run tests + generate HTML dashboard
make test-pytest   # Just pytest, no dashboard
make view-dashboard  # Open the HTML report

---

The HTML Dashboard

We built a visual test dashboard because staring at pytest output gets old.

Features:

• Pass/fail stats at a glance
• Expandable test details
• MS Office-inspired theme (Word blue, Excel green, PowerPoint orange)
• Detailed I/O for debugging

View It

After running make test, open reports/dashboard.html in your browser.

---

Coverage Philosophy

Our coverage is 20% — and that’s intentional.

We focus on:

• Critical extraction paths — The code that touches your documents
• Error handling — Making sure failures are graceful
• Edge cases — The weird stuff that breaks other tools

We don’t test:

• Boilerplate and configuration
• Third-party library internals
• UI/formatting code

Coverage ≠ Quality

100% coverage means nothing if you’re testing the wrong things. We test the paths that matter.

---

CI/CD

Every push triggers:

1. Lint — ruff check
2. Format — black --check
3. Type Check — mypy
4. Tests — pytest with coverage
5. Build — Verify package builds

---

“I could set the building on fire…”

But we’d rather just run the tests.