aprender-graph 0.29.0

GPU-first embedded graph database for code analysis (call graphs, dependencies, AST traversals)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118

# trueno-graph Roadmap
# Aligned with docs/specifications/graph-db-spec.md

project:
  name: trueno-graph
  version: 0.1.0
  description: GPU-first embedded graph database for code analysis

milestones:
  - name: Phase 1 - Foundation
    status: complete
    completion_date: 2025-11-22
    tasks:
      - id: p1-csr-storage
        title: CSR graph storage implementation
        status: complete
        labels: [foundation, core]
      - id: p1-parquet-io
        title: Parquet I/O for graph persistence
        status: complete
        labels: [foundation, io]
      - id: p1-testing
        title: Property tests and 95% coverage
        status: complete
        labels: [quality, testing]

  - name: Phase 2 - Algorithm Integration
    status: in_progress
    tasks:
      - id: p2-pagerank
        title: PageRank implementation via aprender
        status: complete
        labels: [algorithms, core]
      - id: p2-bfs
        title: BFS and find_callers implementation
        status: complete
        labels: [algorithms, traversal]
      - id: p2-benchmarks
        title: Criterion benchmarks for CSR + algorithms
        status: complete
        labels: [performance, benchmarks]
      - id: p2-networkx-comparison
        title: NetworkX baseline comparison benchmarks
        status: todo
        priority: high
        labels: [validation, benchmarks, python]
        description: |
          Create Python benchmarks to validate performance claims:
          - BFS 10x faster than NetworkX CPU baseline
          - PageRank 5x faster than NetworkX
          - Use identical graph structures for fair comparison
        acceptance_criteria:
          - Python script benchmarks/compare_networkx.py created
          - BFS speedup measured and documented
          - PageRank speedup measured and documented
          - Results match spec claims (±20% tolerance)
        estimated_hours: 4
        dependencies: [p2-pagerank, p2-bfs, p2-benchmarks]

  - name: Phase 3 - GPU Acceleration
    status: in_progress
    tasks:
      - id: p3-reverse-csr
        title: Reverse CSR for O(1) incoming neighbors
        status: complete
        completion_date: 2025-11-22
        labels: [optimization, core]
      - id: p3-gpu-setup
        title: wgpu integration and GPU device initialization
        status: complete
        completion_date: 2025-11-22
        priority: medium
        labels: [gpu, infrastructure]
        dependencies: [p3-reverse-csr]
        estimated_hours: 8
      - id: p3-gpu-bfs
        title: GPU BFS kernel implementation
        status: todo
        priority: medium
        labels: [gpu, algorithms, traversal]
        dependencies: [p3-gpu-setup]
        estimated_hours: 12
      - id: p3-gpu-pagerank
        title: GPU PageRank implementation
        status: todo
        priority: medium
        labels: [gpu, algorithms]
        dependencies: [p3-gpu-setup]
        estimated_hours: 10

  - name: Phase 4 - Advanced Algorithms
    status: todo
    tasks:
      - id: p4-louvain
        title: Louvain community detection
        status: todo
        priority: low
        labels: [algorithms, clustering]
        estimated_hours: 16
      - id: p4-betweenness
        title: Betweenness centrality
        status: todo
        priority: low
        labels: [algorithms, centrality]
        estimated_hours: 12
      - id: p4-triangle-count
        title: Triangle counting
        status: todo
        priority: low
        labels: [algorithms]
        estimated_hours: 8

notes: |
  - Phase 1 complete: 23 tests, 96.96% coverage
  - Phase 2 complete: 37 tests, 97.86% coverage (+ NetworkX benchmarks)
  - Phase 3.1 complete: 40 tests, 98.11% coverage, O(1) incoming neighbors
  - Phase 3.2 complete: 27 tests (with --features gpu), GPU device initialization
  - Next priority: GPU buffer management (Phase 3.3) then GPU BFS/PageRank