feat: Add MongoDB Collection Comparison Support

[ReginaldErzoah]

Summary

Add direct MongoDB collection comparison workflows.

---

Motivation

MongoDB data is organized as collections instead of relational tables.

Dift should support comparing MongoDB collections in a way that feels natural to NoSQL users.

This enables:

• collection-to-collection validation
• production vs staging document comparison
• NoSQL regression testing
• application data quality checks

---

Proposed Improvements

• Support collection-level comparison
• Compare old and new MongoDB collections
• Use a user-defined document key for matching
• Preserve existing Dift comparison behavior where possible

---

Suggested Files

Potential implementation areas:

dift/io/mongodb_reader.py
dift/core/comparator.py
dift/core/schema_diff.py
dift/core/row_diff.py
tests/test_mongodb_reader.py
tests/test_mongodb_integration.py
docs/connectors/mongodb.md

---

Suggested Tasks

• Define MongoDB collection URI format
• Load collections into comparable tabular structure
• Support document key matching
• Add tests for collection comparisons
• Add examples
• Update documentation

---

How to Test

Run:

pytest
ruff check .

Run targeted tests:

pytest tests/test_mongodb_reader.py
pytest tests/test_mongodb_integration.py

Manual validation example:

dift mongodb://localhost:27017/shop.customers_old \
     mongodb://localhost:27017/shop.customers_new \
     --key customer_id

Verify:

• collections compare successfully
• added and removed documents are detected
• key-based matching works
• reports remain valid

---

Documentation Impact

Update:

docs/connectors/mongodb.md
docs/examples.md
docs/connectors/index.md

Documentation should include:

• collection comparison examples
• document key usage
• supported URI format
• expected output behavior

---

Acceptance Criteria

• MongoDB collections can be compared directly
• Key-based document matching works
• Existing comparison reports work with MongoDB data
• Tests pass
• Documentation updated