Skip to content
Permalink
Branch: master
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
executable file 217 lines (199 sloc) 7.88 KB
#!/usr/bin/python3
import dnf
from collections import Counter, defaultdict
# A dictionary of known "words" from RPM names.
#
# Partly this is useful to generate nosplit_words, so "gnome-shell" doesn't get
# broken into "gnome" and "shell".
# It also gets used to guess the "meaning" of each part of the name.
#
# NOTE: These should be listed from least to most specific categories, so when
# we make the all_words reverse-lookup dict the result for, say, "gtk" is
# "framework" instead of just "project name".
known_words = {
"project name": [
"texlive", "eclipse", "glibc", "libreoffice", "drupal7", "pcp", "vim",
"globus", "root", "coin-or", "asterisk", "uwsgi", "gimp", "gcc",
"qemu", "boost", "tesseract", "gallery2", "nagios", "maven",
"jboss", "apache-commons", "glassfish", "jetty", "collectd", "kde",
"kf5", "qt5", "qt", "plasma", "gnome", "gnome-shell", "sugar", "mate",
"lxqt", "xfce4", "hunspell", "aspell", "autocorr", "fawkes",
"openrdf-sesame", "google-noto", "fence", "jenkins",
"emacs", "ibus", "derelict", "django", "qpid", "libvirt", "lodash",
"matreshka", "plexus", "shrinkwrap", "soletta", "springframework",
"trytond", "oslo", "pulp", "geany", "NetworkManager", "yum", "felix",
"fusionforge", "opensips", "jackson", "vdsm", "horde", "sblim",
"gcompris", "glite", "arquillian", "geronimo", "binutils",
"asterisk-sounds-core", "xorg-x11", "xorg", "man-pages",
"libjpeg-turbo", "util-linux", "xkb-utils", "gnome-getting-started",
"subscription-manager", "virt-manager", "gtk", "gtk2", "gtk3",
"aws-sdk",
],
# FUTURE: could get clever and use this to merge variant words?
"variants": {
"python": ["python2", "python3"],
"mingw": ["mingw32", "mingw64"],
"gtk": ["gtk2", "gtk3"],
"qt": ["qt4", "qt5"],
},
"programming language": [
"perl", "python", "python2", "python3", "js", "ruby",
"rubygem", "php", "go", "ocaml", "R", "erlang", "rust", "java", "lua",
"tcl", "ghc", "ocaml", "mono", "gambas3", "octave", "gap", "c++",
"cpp", "c", "sharp",
],
# FUTURE: could get clever and ignore "words" after a packaging prefix
# (until we hit a known suffix...)
"package prefix": [
"perl", "php-pear", "php-pecl", "gap-pkg", "rubygem", "golang",
"nodejs", "ghc", "XStatic",
],
"vendor": [
"google", "fedora", "redhat", "hashicorp",
],
"framework": [
"gtk", "gtk2", "gtk3", "glib", "dbus",
"qt", "qt4", "qt5", "kf5", "plasma",
"ibus", "boost", "lodash", "grunt", "django",
"openmpi", "mpich",
"sqlite", "mysql", "pgsql", "postgresql",
"selinux", "device-mapper", "git", "docker",
"syntastic", "zendframework", "zend", "gl", "flask",
"maven", "trac", "nautilus", "lv2",
],
"protocol": [
"http", "ldap", "x11", "ssh",
],
"data format": [
"json", "yaml", "xml", "el", "latex", "html", "HTML", "XML", "text",
],
"environment": [
"gnome", "kde", "mate", "lxde", "lxqt", "xfce", "xfce4", "sugar",
],
"build-variant": [
"mingw32", "mingw64", "python2", "python3",
"compat", "static", "debug", "linux-gnu",
],
"file-type": [
"devel", "doc", "docs", "javadoc", "fonts", "libs", "bin", "data",
"info", "parent", "theme", "lib", "config", "filesystem", "firmware",
"macros", "backgrounds", "bundle", "icon", "rpm-macros", "srpm-macros",
"source", "mythes",
],
"font-type": [
"sans", "serif",
],
"file-purpose": [
"tests", "test", "unit-test", "unit-tests", "client", "server",
"tools", "examples", "cli", "gui", "daemon", "util", "utils",
"daemon", "manual", "api", "demo", "web", "console", "log",
"agent", "manager", "parser", "runtime", "cache", "bridge",
"session-manager", "power-manager", "font-manager", "help",
"agents", "proxy", "desktop", "sounds",
],
# FUTURE: could get clever and use the (singular, plural) pairings
"extends/enhances": [
"plugin", "plugins",
"plug-in", "plug-ins",
"extension", "extensions",
"provider", "providers",
"module", "modules",
"driver", "drivers",
"addon", "addons",
],
"core/extras": [
"core", "common", "base", "extra", "extras", "others", "contrib",
],
"release name": [
"spherical-cow", "beefy-miracle", "schroedinger-cat",
],
"concept": [
"file", "system", "stream", "net", "map", "path", "app",
],
"translation": [
"langpack", "langpacks", "i18n", "l10n",
"fr", "ru", "it", "es", "de", "en", "ja", "pl", "cs", "gb",
"utf-8", "KOI8-R",
],
}
all_words = {w:desc for desc,words in known_words.items() for w in words}
nosplit_words = set(w for w in all_words if '-' in w)
def wordsplit(rpmname):
# empty string
if not rpmname:
return []
# nothing to split
if '-' not in rpmname or rpmname in nosplit_words:
return [rpmname]
# check for nosplit_words
for w in nosplit_words:
p = w+'-'
if rpmname.startswith(p):
rest = rpmname[len(p):]
return [w] + wordsplit(rest)
# otherwise just split on dashes
w, rest = rpmname.split('-', 1)
return [w] + wordsplit(rest)
# --- Okay, ready to start doing stuff? Here we go!
print("fetching DNF metadata..")
b = dnf.Base()
b.read_all_repos()
b.repos.all().disable()
b.repos['fedora'].enable()
b.repos['updates'].enable()
sack = b.fill_sack(load_system_repo=False)
print("analyzing package names..")
names = dict()
subpkgs = defaultdict(list)
namewords = list()
wordnames = defaultdict(set)
for pkg in sack.query():
if pkg.name in names:
continue
name = pkg.name
src = pkg.source_name
words = wordsplit(name)
# name -> src
names[name] = src
# src -> [name, name, ...]
subpkgs[src].append(name)
# [[word, word, ...], ...]
namewords.append(words)
# word -> [name, name, ...]
for w in words:
wordnames[w].add(name)
# Count raw number of uses for every word in every package name
words = Counter(w for nw in namewords for w in nw)
prefixes = Counter(nw[0] for nw in namewords if len(nw) > 1)
suffixes = Counter(nw[-1] for nw in namewords if len(nw) > 1)
# How many different source packages - i.e. spec files - use each word?
# (this reduces noise from things like texlive...)
srcwords = Counter({w:len(set(names[n] for n in wordnames[w])) for w in words})
# A wordlist that includes every word in the top 100 of each category
commonwords = set(w for cat in (words, srcwords, prefixes, suffixes)
for w,c in cat.most_common(100))
# bonus stuff for finding words like "coin-or"
def iterpairs(iterable):
i = iter(iterable)
left = next(i)
for right in i:
yield left, right
left = right
pairs = Counter(p for nw in namewords for p in iterpairs(nw))
def dumpcsv(fobj, wordlist, key=srcwords.get):
fobj.write("word,specfiles,rpms,as prefix,as suffix,meaning\n")
for word in sorted(wordlist, key=key, reverse=True):
fobj.write("{},{},{},{},{},{}\n".format(word,
srcwords.get(word, 0),
words.get(word, 0),
prefixes.get(word, 0),
suffixes.get(word, 0),
all_words.get(word, '')))
# run with "ipython3 -i dnf-count-rpm-words.py" to do interactive exploration!
if __name__ == '__main__':
with open("rpm-name-word-counts.csv", 'wt') as fobj:
print("writing {}...".format(fobj.name))
dumpcsv(fobj, words)
with open("rpm-name-word-counts-common.csv", 'wt') as fobj:
print("writing {}...".format(fobj.name))
dumpcsv(fobj, commonwords)
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.