Revision history for "LLM::Graph"
0.1.1 2026-09-03
- First version of the agent skill "raku-llm-graph-making" (placed in "./resources").
- Fix: Making sure arguments that are promises for node test-functions are awaited/evaluated.
0.1.0 2026-09-01
- Implemented threaded LLM-functions.
0.0.12 2026-08-27
- Refactor/fix: using the renamed Graph-class attribute adjacency-map.
0.0.11 2026-04-05
- Bug fix: proper handling of user-specified arguments that are undefined.
- E.g. Whatever.
- Documentation:
- Summary of large text with Python notebook (for comparison)
0.0.10 2026-01-22
- Fix: Proper passing of parameters that are for the eval-function and test-function.
- Fix: Preventing graph-node dependencies when input arguments have hashmaps.
- Added example using a "custom LLM" in order to trace iterative generation refinements using LLM-graph.
- It illustrates that cyclic LLM-graphs are not that needed.
0.0.9 2026-01-20
- Fix: proper passing of the default LLM-evaluator object.
- Added comparison notebook with Python's "LangGraph".
0.0.8 2025-09-03
- Handling of spec synonyms during normalization
- Like "llm-sub", "eval-sub", etc.
0.0.7 2025-09-02
- Refactored / fixed slurpy named parameters handling