hydra-cli

Command-line interface for Hydra, the water infrastructure simulation platform. It drives Hydra's water distribution engine: reads EPANET .inp network descriptions from files or HTTP URLs, runs extended-period hydraulic and water quality simulation, and writes results to .rpt and .out files. Hydra's other engines — urban drainage and open channel — are planned and not yet available from the CLI.

→ Full documentation

Breaking changes in 3.0

The argument surface was redesigned so it can carry more than one engine.

• Subcommands. hydra run <model>, hydra report, hydra engines. A bare hydra <model> prints a hint naming the replacement.
• The EPANET positional triple is gone. hydra net.inp net.rpt net.out becomes hydra run net.inp --summary net.rpt --results net.out. The old order encoded one engine and one pair of artifacts; every added engine made it less true.
• One name per concept. The model is <MODEL>, the binary time series is --results, the native run log is --summary. Previously --output meant the binary results while hydra report's --out meant the report document — one letter apart, opposite meanings.
• --engine, defaulting to detection from the model's contents. There is no default engine; an unidentifiable model is an error, never a guess.
• -v is verbosity (repeatable), reclaimed at this major boundary. -V remains --version.
• Internal errors exit 4, no longer reusing 2 (solver). 0–3 unchanged.

Install

For most users, Cargo install is the recommended path.

Option 1 — Pre-built binary (no Rust required)

Download the hydra binary for your platform from the releases page.

macOS — Pre-built CLI binaries are currently not notarised. If Gatekeeper blocks the binary, remove the quarantine flag:

xattr -d com.apple.quarantine hydra

Option 2 — Cargo (recommended)

cargo install hydra-cli

Usage

# Run a simulation — summary goes to stdout
hydra run network.inp

# Write the summary and the binary time-series results
hydra run network.inp --summary report.rpt --results output.out

# JSON summary (chosen by the .json suffix)
hydra run network.inp --summary report.json

# Accept an HTTP URL as the model (redirects followed, up to 10; plain
# http:// accepted; bodies up to 1 GiB; 10 s connect / 300 s overall timeout)
hydra run https://example.com/network.inp

# Name the engine instead of detecting it from the model
hydra run network.inp --engine wds

# What engines does this build provide?
hydra engines

# Build a report document from a finished run
hydra report --model network.inp --results output.out -o report.html

# Suppress progress output / add detail
hydra run network.inp -q
hydra run network.inp -v

# Print version
hydra -V

Engine selection

The engine is decided by the model's contents, never its extension — .inp belongs to both EPANET and SWMM. Exactly one engine must identify the model for it to run. If none does, or the file is only shaped like some engine's format without identifying it, Hydra stops and asks for --engine rather than guessing: routing a stormwater model to a pressurised-pipe solver would return a confident wrong answer instead of an error.

Exit codes

License

AGPL v3 — see COMMERCIAL_LICENSE.md for commercial licensing options.