build graph from codebase

[rekino]

a graph of nodes and edges with definition:

Node:

• function name: str
• namespace: str
• definition: str

Edge:

• line number: int
• arguments: list[str]
• returns: list[str]

[amrhssn]

@rekino I implemented a Python code that implements an example scenario for testing Lattice. You can find it here: https://github.com/soverant/lattice/blob/dev-%238-retrieval-setup-preparation/src/lattice/retrieve/example_scenario_1.py

Based on the definitions above, I manually created the code graph for the example scenario: https://github.com/soverant/lattice/blob/dev-%238-retrieval-setup-preparation/src/lattice/retrieve/example_data_1.py

Note that I haven't used an LLM yet to compute the embeddings for nodes and edges and index the graph. The enrichment of this graph data is something that @mehdibalouchi is going to consider.

@farhoud @mehdibalouchi @rekino Please have look at the scenario and the extracted graph, and let me know what you think 🙏🏻 These implementations are related to this issue #8 that I created.

[rekino]

This is a good point to start; we can use it as a case study.