Rewriting Dagoba: Building an In‑Memory Graph Database in Python

The article introduces Dagoba, a small in‑memory graph database originally written in JavaScript, and explains why a dedicated graph database is needed because relational SQL queries become cumbersome for traversing nodes and edges.

It then describes the rewrite in Python, replacing prototype‑based code with plain dictionaries, using special fields like _id , _from , _to , and handling surrogate primary keys via an internal _next_id counter.

Implementation steps are presented: Step 0 creates the data model; Step 1 adds automatic primary‑key generation; Step 2 implements eager (immediate) queries with methods such as node() , outcome() , income() , and unique() ; Step 3 adds bidirectional edge support by storing outgoing and incoming edge lists on each node; Step 4 introduces lazy (deferred) queries using a pipeline of closure functions that execute only when run() is called; Step 5 adds a take() method to limit results and demonstrates how lazy queries reduce node‑visit counts compared to eager queries; Step 6 shows how custom query methods can be added at runtime, enabling domain‑specific language extensions.

Throughout the guide, unit tests using Python’s unittest framework verify functionality, including duplicate‑key detection, edge validation, node‑visit counting, and custom pipeline methods. The article concludes that the resulting Python implementation provides a functional, extensible graph database suitable for teaching and experimentation.