Computational reproducibility package accompanying Item 7, "In silico validation of topological modulation," of Chapter 11 of the book Movimento da água no corpo da planta.

This deposit contains the computational reproducibility package accompanying Item 7, “In silico validation of topological modulation,” of Chapter 11 of the book Movimento da água no corpo da planta. The experiment is based on the Woody Crown Hydraulic Network model (WCHN; RHLC in Portuguese), which represents the woody crown as a rooted, directed, loop-free network. The main script generates 40 independent random WCHNs, each containing 101 nodes, and applies two successive rounds of topological modulation. During each round, every final node identified at the beginning of the round is detached from its parent and laterally reinserted at its grandparent. This operation moves the terminal portion of each affected woody axis closer to the initial node while keeping the total number of nodes constant. Final nodes whose parent is already the initial node have no grandparent and therefore remain unchanged. After the original state and each modulation round, the scripts recalculate network decomposition, topology, and eight WCHN properties: navigability, vulnerability, symmetry, complexity, laterality, basitony, node sharing, and potential topological plasticity. Symmetry and complexity are calculated using their coefficient-of-variation-normalized formulations. Vulnerability is deliberately calculated using the original formulation, in which the initial node and all final nodes are excluded from the set of attackable nodes. This historical formulation is retained to reproduce and illustrate the apparent vulnerability paradox discussed in the chapter: vulnerability increases after modulation because internal nodes are reclassified as final nodes and the number of attackable nodes decreases, rather than because the network becomes intrinsically more fragile. Paired Wilcoxon signed-rank tests compare the original networks with the first modulation state and the first with the second modulation state. The scripts can export the results for every replicate and state as a CSV file and generate a panel of boxplots corresponding to the properties analyzed in the chapter. The package includes: · rhlc_random_wcn.py: base random WCHN generator and implementation of the network properties; · rhlc_prado_calibrado.py: calibrated version of the generator using an edge-acceptance weight of approximately 0.55 to better approximate the width and depth of the random networks measured by Prado (2024); · rhlc_experimento_modulacao_101.py: main script that generates the 40 networks, performs the two modulation rounds, recalculates the properties, conducts the paired Wilcoxon tests, and optionally exports the CSV file and boxplot figure; · Redes aleatórias no Excel.xlsx: measurements of 58 random networks from Prado (2024), with sizes ranging from 10 to 248 nodes, used as the empirical reference for generator calibration; · resultados_exemplo_replicas.csv: example output containing property values for each replicate and experimental state; · figura_exemplo_reprodutibilidade.png: example boxplot figure produced from the reference execution; · README.md: methodological details, execution instructions, equations, and interpretation notes. This is not the original script, with the original random seeds, that produced the exact numerical values reported in Tables 6–8 and Figure 15 of the chapter. That original file and its seeds were not preserved. The present package is a reproducibility implementation reconstructed from the documented methodology. It uses a fixed set of new seeds to generate a new sample of 40 networks and allows readers to verify the qualitative concordance of the experiment: the direction and general pattern of the topological responses, the delayed response of basitony, and the vulnerability reclassification artifact. Exact means, medians, coefficients of variation, p-values, and node counts may differ from those reported in the chapter. Very small floating-point differences may also occur among Python and library versions. The package requires Python, SciPy for the paired Wilcoxon tests, and Matplotlib for figure generation. It can be executed with: python rhlc_experimento_modulacao_101.py --csv resultados.csv --fig figura15.png The files are distributed under the Creative Commons Attribution 4.0 International license (CC BY 4.0).