MANIFEST.yaml

# Corpus Manifest: Pinned Repository Versions
#
# This file records the exact state of every repository in the benchmark corpus.
# Reproducibility requires cloning each repo at the specified commit. The graph
# databases (.knowing/graph.db) are derived artifacts: they can be rebuilt from
# the source repos using the commands in corpus-setup.sh.
#
# To verify your local corpus matches this manifest:
#   ./corpus-setup.sh verify
#
# To rebuild from scratch:
#   ./corpus-setup.sh clone     # clone all repos at pinned commits (~5 min)
#   ./corpus-setup.sh index     # tree-sitter + in-process resolvers (~5 min)
#   ./corpus-setup.sh enrich    # LSP enrichment (requires language servers, ~2 hours, optional)
#   # Embeddings are off by default (confirmed neutral). Embed step unnecessary.

format_version: 1
last_updated: "2026-06-01"
task_count: 328
repo_count: 16
language_count: 8

repos:
  - name: caddy
    url: https://github.com/caddyserver/caddy.git
    commit: 4d60d936edce2ab49cb36c47d0d1bdc0029a0ed2
    commit_date: "2026-05-27"
    language: Go
    tasks: 20
    enrichment: gopls
    nodes: 17881
    edges: 46218
    embeddings: 23759

  - name: cargo
    url: https://github.com/rust-lang/cargo
    commit: 2dbb1af80a2914475ba76827a312e29cedfa6b2f
    commit_date: "2024-08-20"
    language: Rust
    tasks: 19
    enrichment: rust-analyzer
    nodes: 81434
    edges: 136779
    embeddings: 18372

  - name: django
    url: https://github.com/django/django
    commit: 373cb3037fe4e67adbac9ac43340391e859aa957
    commit_date: "2024-08-07"
    language: Python
    tasks: 36
    enrichment: pyright
    nodes: 128482
    edges: 370498
    embeddings: 97230

  - name: fastapi
    url: https://github.com/fastapi/fastapi.git
    commit: 2bd97a6013cf29b0f1ecd8850e8ba4b89d00b89c
    commit_date: "2026-05-27"
    language: Python
    tasks: 20
    enrichment: pyright
    nodes: 18035
    edges: 38601
    embeddings: 19776

  - name: flask
    url: https://github.com/pallets/flask
    commit: ab8149664182b662453a563161aa89013c806dc9
    commit_date: "2024-11-13"
    language: Python
    tasks: 19
    enrichment: pyright
    nodes: 4679
    edges: 11003
    embeddings: 6264

  - name: jekyll
    url: https://github.com/jekyll/jekyll.git
    commit: 202df571314ba1d18e9fccd81d12aaad4a703c38
    commit_date: "2026-04-22"
    language: Ruby
    tasks: 20
    enrichment: ruby-lsp
    nodes: 14594
    edges: 34479
    embeddings: 1586

  - name: kafka
    url: https://github.com/apache/kafka.git
    commit: fa097fa593a2a3da17bd1c234bf140565561a1b8
    commit_date: "2026-05-24"
    language: Java
    tasks: 19
    enrichment: jdtls
    nodes: 105032
    edges: 957073
    embeddings: 32923

  - name: kubernetes
    url: https://github.com/kubernetes/kubernetes
    commit: 7c48c2bd72b9bf5c44d21d7338cc7bea77d0ad2a
    commit_date: "2024-04-17"
    language: Go
    tasks: 19
    enrichment: gopls
    nodes: 241909
    edges: 706349
    embeddings: 144842

  - name: ocelot
    url: https://github.com/ThreeMammals/Ocelot.git
    commit: a5efd3cc913883de017c9afaf9637402ce965606
    commit_date: "2026-05-21"
    language: "C#"
    tasks: 20
    enrichment: csharp-ls
    nodes: 17105
    edges: 52576
    embeddings: 13746

  - name: rails
    url: https://github.com/rails/rails.git
    commit: d2a0d991762d3aafafc28f7f6bd114ef33c21adf
    commit_date: "2026-05-29"
    language: Ruby
    tasks: 20
    enrichment: ruby-lsp
    nodes: 386186
    edges: 2429521
    embeddings: 0

  - name: ripgrep
    url: https://github.com/BurntSushi/ripgrep.git
    commit: 4857d6fa67db69a95cd4b6f2adda5d807d4d0119
    commit_date: "2026-05-26"
    language: Rust
    tasks: 20
    enrichment: rust-analyzer
    nodes: 10881
    edges: 21742
    embeddings: 2234

  - name: spark-java
    url: https://github.com/perwendel/spark.git
    commit: 1973e402f5d4c1442ad34a1d38ed0758079f7773
    commit_date: "2022-07-10"
    language: Java
    tasks: 20
    enrichment: jdtls
    nodes: 4460
    edges: 13238
    embeddings: 6143

  - name: terraform
    url: https://github.com/hashicorp/terraform.git
    commit: a2ad11c4f14ceea1e81c81a56c44428fc4110f20
    commit_date: "2026-05-23"
    language: Go
    tasks: 20
    enrichment: gopls
    nodes: 98915
    edges: 332394
    embeddings: 51676

  - name: vscode
    url: https://github.com/microsoft/vscode.git
    commit: 89de5a8d4d6205e5b11647eb6a74844ca23d2573
    commit_date: "2024-06-03"
    language: TypeScript
    tasks: 19
    enrichment: tsserver
    nodes: 552184
    edges: 4425658
    embeddings: 167380

  - name: saleor
    url: https://github.com/saleor/saleor.git
    commit: 8dc4b0baa32796b5e77b65cfe4a032abdf0b82fe
    commit_date: "2026-05-26"
    language: Python
    tasks: 11
    enrichment: none (unenriched, framework-using validation repo)
    nodes: 33676
    edges: 285408
    embeddings: 0
    note: "First framework-USING repo (Django e-commerce app). Validates equiv classes generalize to application code."

  - name: calcom
    url: https://github.com/calcom/cal.com.git
    commit: d1ad4ea
    commit_date: "2026-06-04"
    language: TypeScript
    tasks: 11
    enrichment: tsserver
    nodes: 80443
    edges: 245889
    embeddings: 0
    note: "Scheduling platform (Next.js/tRPC/Prisma). First typical TypeScript app repo. Domain equiv classes: booking, availability, calendar, webhook, attendee, limits, seats."

# Cross-cutting tasks (multi-repo) use symbols from multiple repos above.
# No separate repo entry needed.
cross_cutting:
  tasks: 9
  note: "Tasks that span multiple repos or test cross-language retrieval"

competitors:
  - name: codegraph
    version: "latest via npm (2026-05-24)"
    install: "npm install -g @nicolo-ribaudo/codegraph"
  - name: codebase-memory
    version: "latest via npm (2026-05-24)"
    install: "npm install -g codebase-memory"
  - name: GitNexus
    version: "latest via npm (2026-05-24)"
    install: "npm install -g gitnexus"
  - name: Gortex
    version: "latest via go install (2026-05-24)"
    install: "go install github.com/gortex-ai/gortex@latest"
  - name: Aider
    version: "latest via pip (2026-05-24)"
    install: "pip install aider-chat"
  - name: grep
    version: "system grep (BSD/GNU)"
    install: "built-in"