Awesome
Gzemnid
Notice: some commands might require --max-old-space-size=2000
or above.
Using pre-built datasets
See documentation on Using pre-built datasets page.
Data structures
All data files are stored inside the ./pool/
dir by default.
See documentation on Data structures page.
Commands
The main script is invoked as gzemnid command [subcommand]
(or ./gzemnid.js command [subcommand]
),
where [subcommand]
is optional.
Here is the list of the current commands:
fetch
— buildsbyField.info.json
. Should take about 2 minutes.fetch run byField.json
— buildsbyField.info.json
locally without network connection, but you need to downloadbyField.json
manually from the registry to run that.
stats
— runs subcommandrebuild
.stats rebuild
— rebuildsstats.json
, downloading stats for all packages present inbyField.info.json
. Should take about 25 minutes.stats update
— updatesstats.json
for only newly added packages, keeping the numbers for already present packages.
meta
— buildsmeta/
directory, downloading meta info for all packages present inbyField.info.json
. Outdated files that were present in themeta/
directory are moved tometa.old/
.depsdb
— runs subcommandsplain
,resolved
,nested
,stats
,depsdb plain
— buildsdeps/deps.json
. Requiresmeta/
dir contents.depsdb resolved
— buildsdeps/deps-resolved.json
. Requiresdeps/deps.json
.depsdb nested
— buildsdeps/deps-nested.json
. Requiresstats.json
anddeps/deps-resolved.json
. Should take about 6 minutes.depsdb stats
— buildsdeps/deps-nested.txt
. Requiresstats.json
anddeps/deps-nested.json
. Should take about 30 seconds.
packages
— buildscurrent/
directory, downloadinglatest
versions for all packages present inbyField.info.json
. Outdated files that were present in thecurrent/
directory are moved tooutdated/
.extract
— runs subcommandspartials
,totals
,extract partials
—extract totals
—
code search {regex}
— performs a code search over a specified regular expression using the pre-built dataset.ast execute {file.js}
— performs an AST search using the pre-built dataset. Example script is located inexamples/ast_status.js
, run withgzemnid ast execute ./examples/ast_status.js
.server
— starts the web server providing the search API endpoints.
Times are given for reference, could depend significantly on the internet connection speed and/or CPU speed, and increase over time with npm registry growth.
Server
TODO: document server.
Started via gzemnid server
.
Deception
Note: think twice before relying on the data obtained from Gzemnid or using it to decide on something.
Code search has both false negatives and false positives — some files are ignored, some files are unused, and some lines could be in a middle of a comment block. Also, your regexps are never ideal.
AST tree also ignores a list of excluded files and directories and minified code and includes unused code and files if those are present in the package for some reason.
Downloads/month are not equal to popularity, and you can't see which version is being used.
Code and AST search, among other things, takes only latest
released package versions into an account. That could be significantly different from master
, beta branches, also older versions could be much more popular that latest
.
All datasets get out of date the moment you build them.
Scoped packages are ignored completely.
Gzemnid deceives you, keep that in mind. But it's still better than nothing.