The Zinc Search Indexer crawls a folder for files and triggers indexing them via the HTTP API of Zinc. There is no logic or processing. The files aren't parsed but send to the API as they are. Using JSON files seems to be most reasonable as Zinc quite nicely indexes those providing a full text search and a simple navigation on the web interface.
Settings are applied by config files which look like the one below.
[db]
url = "https://zincsearch.box"
user = "username"
pass = "super_secret_pass"
[[indexers]]
folder = "${HOME}/loads_of_json_files_about_nature"
rxmatcher = ".*/animal_.*.json$"
index = "animals"
[[indexers]]
folder = "${HOME}/loads_of_json_files_about_nature"
rxmatcher = ".*/vegetable_.*.json$"
index = "vegetables"If there is a configuration file named zsi_conf.toml you would have to pass the path to the config file as positional argument like this:
zsi ${HOME}/zsi/zsi_confNot providing an absolute path will make ZSI look for the file automatically. These folders are used for the lookup in the following order.
1. inside the binary's folder
2. ${HOME}/.config/zsi
3. ${HOME}/.conf/zsi
Executing a command like zsi animals will make zsi look for a config file called animals.toml in the folders listed above. The first found will be processed.
a zinc search indexer
Arguments:
[<config-file>] file to process, positional arg required
Flags:
-h, --help Show context-sensitive help.
-t, --threads=16 max indexer threads, default no of avail.
cpu threads
-l, --log-file="/dev/stdout" log file
-e, --log-level="info" log level
-n, --log-no-colors disable output colours, print plain text
-j, --log-json enable json log, instead of text one
-d, --debug debug mode
-V, --version-flag display version