Setup and build

init

hydra-ctl init htest
hydra-ctl init htest --name "Breaking Bad" --prefix bb
hydra-ctl init htest.hydra

See the output of each example

Creates an archive in a new or empty folder from the built-in template, or opens a .hydra file into one.

At a terminal it asks for the archive's name (Enter keeps the folder's name) and its id prefix (Enter takes one made from the name, and none keeps a prefix per collection).

Option	Does
--name <name>	Sets the archive's name without asking
--prefix <prefix>	Sets the archive's id prefix without asking

activate, deactivate

hydra-ctl activate htest
hydra-ctl activate
hydra-ctl deactivate

See the output of each example

activate with a folder writes .hydra-activate here, and commands run from this folder (or below it) then use that archive. Without a folder it shows the active archive.

deactivate removes the marker.

import

hydra-ctl import ~/exports/opensanctions.json --format ftm
hydra-ctl import case.hydra-ftm.jsonl --format hydra-ftm --name fring
hydra-ctl import case-graph --format csv
hydra-ctl import case.graphml --format graphml --force

See the output of each example

Loads records into the archive as a collection, from the same formats export writes. Copying a file into collections/ yourself and running build does the same for ftm and hydra-ftm; import checks the file reads as records first, names the collection and rebuilds.

Format	Is
ftm	Standard FollowTheMoney JSON: a list, {"entities": [...]}, or JSON lines, as Aleph, OpenSanctions and other FtM tools write it. FtM's proof is read as sources
hydra-ftm	The same with links written as names, as export --format hydra-ftm writes it
csv	The folder export --format csv writes (nodes.csv and edges.csv)
graphml	The file export --format graphml writes

A graph carries each entity's id, type and name, and its links. Each relationship edge becomes the relationship (with its role and dates), and each link-property edge the property.

Other fields are not in a graph, so they don't come back. The tree and open merges (placedUnder, sameAs) are left out too; a build works them out again.

ftm and hydra-ftm files are copied as they are, and a graph is written as FtM JSON lines. The collection takes the file's name (case.ftm.jsonl becomes case, case-graph/ becomes case).

Option	Does
--format <format>	Says what the file is: ftm, hydra-ftm, csv or graphml
--name <name>	Chooses the collection's name
--force	Replaces a collection of that name
--no-build	Imports without rebuilding

build

hydra-ctl build
hydra-ctl build htest --force

See the output of each example

Reads every file in collections/ and .hydra/sources/entries/, merges records that are the same entity, places them, and writes .hydra/parsed/. It prints a summary of records, claims, nodes, merges, placements and references by status, and issue counts.

When nothing has changed since the last build, it prints that summary again without building.

Standard FtM data works as it comes. A link field that holds a record's id (as in owner: ["NK-p1"]) is followed by that id and shown as the record's name, and the id is kept for export.

Option	Does
--force	Builds even when nothing changed
--no-prefix	Keeps record ids as written, without prefixes
--verbose	Also prints the full summary: claims, reads and the review ledger