How colinstephen/dual_tree_hvg works
This repository implements horizontal visibility graph (HVG) construction and benchmarks its dual-tree method against binary-search-tree and divide-and-conquer approaches. The main workflow configures algorithms and time-series sources, generates benchmark inputs, times construction, and records results; separate experiments measure chunked graph construction and merging on fractional Brownian, finance, and EEG data. The source excerpts show algorithm internals and several benchmark paths; some benchmark modules are represented from the README and file tree, so their internal wiring is not inferred.
Main entry points: experiment1.py Experiment one · run_experiment.py Experiment runner · experiment2.py Experiment two · merge_experiment3.py FBM merge benchmark
Components 13
HVG algorithms
- Dual-tree HVG
dt_hvg.py - BST HVG
bst_hvg.py - Divide-conquer HVG
dc_hvg.py
Time-series sources
- Series generators
streams.py - Finance and EEG data
data
Benchmark workflows
- Experiment runner
run_experiment.py - Benchmark utilities
experiment_utils.py - Experiment one
experiment1.py - Experiment two
experiment2.py - FBM merge benchmark
merge_experiment3.py - Finance merge benchmark
merge_experiment.py - EEG merge benchmark
merge_experiment2.py
Other components
- Researcher
Connections 24
- Researcher → Experiment runner: runs
run_experiment.py - Experiment runner → Dual-tree HVG: times
run_experiment.py - Experiment runner → BST HVG: times
run_experiment.py - Experiment runner → Divide-conquer HVG: times
run_experiment.py - Experiment runner → Series generators: generates inputs
run_experiment.py - Researcher → Experiment one: runs
experiment1.py - Experiment one → Benchmark utilities: uses
experiment1.py - Experiment one → Series generators: selects sources
experiment1.py - Experiment one → Dual-tree HVG: benchmarks
experiment1.py - Experiment one → BST HVG: benchmarks
experiment1.py - Experiment one → Divide-conquer HVG: benchmarks
experiment1.py - Researcher → Experiment two: runs
experiment2.py - Experiment two → Benchmark utilities: uses
experiment2.py - Experiment two → Series generators: selects sources
experiment2.py - Experiment two → Dual-tree HVG: benchmarks
experiment2.py - Experiment two → BST HVG: benchmarks
experiment2.py - Researcher → FBM merge benchmark: runs
merge_experiment3.py - FBM merge benchmark → Series generators: generates FBM
merge_experiment3.py - FBM merge benchmark → Dual-tree HVG: builds and merges
merge_experiment3.py - FBM merge benchmark → BST HVG: builds and merges
merge_experiment3.py - Researcher → Finance merge benchmark: runs
- Finance merge benchmark → Finance and EEG data: reads finance data
merge_experiment.py - Researcher → EEG merge benchmark: runs
- EEG merge benchmark → Finance and EEG data: reads EEG data
merge_experiment2.py