explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-06-27 16:33:18 +03:00
Code Issues Packages Projects Releases Wiki Activity

Let --data-path be specified when running download.py scripts

Browse Source 
Resolves https://github.com/explosion/spaCy/issues/637
This commit is contained in:
Mark Amery 2016-11-20 15:48:04 +00:00
parent c413ffe26d
commit 3871007c72
3 changed files with 20 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
spacy/de/download.py
View File

@ -4,9 +4,10 @@ from ..download import download
@plac.annotations( @plac.annotations(
force=("Force overwrite", "flag", "f", bool), force=("Force overwrite", "flag", "f", bool),
data_path=("Path to download model", "option", "d", str)
) )
def main(data_size='all', force=False): def main(data_size='all', force=False, data_path=None):
download('de', force) download('de', force=force, data_path=data_path)
if __name__ == '__main__': if __name__ == '__main__':

18
spacy/download.py
View File

@ -10,10 +10,19 @@ from . import about
from . import util from . import util
def download(lang, force=False, fail_on_exist=True): def download(lang, force=False, fail_on_exist=True, data_path=None):
if not data_path:
data_path = util.get_data_path()
# spaCy uses pathlib, and util.get_data_path returns a pathlib.Path object,
# but sputnik (which we're using below) doesn't use pathlib and requires
# its data_path parameters to be strings, so we coerce the data_path to a
# str here.
data_path = str(data_path)
try: try:
pkg = sputnik.package(about.__title__, about.__version__, pkg = sputnik.package(about.__title__, about.__version__,
about.__models__.get(lang, lang)) about.__models__.get(lang, lang), data_path)
if force: if force:
shutil.rmtree(pkg.path) shutil.rmtree(pkg.path)
elif fail_on_exist: elif fail_on_exist:
@ -24,15 +33,14 @@ def download(lang, force=False, fail_on_exist=True):
pass pass
package = sputnik.install(about.__title__, about.__version__, package = sputnik.install(about.__title__, about.__version__,
about.__models__.get(lang, lang)) about.__models__.get(lang, lang), data_path)
try: try:
sputnik.package(about.__title__, about.__version__, sputnik.package(about.__title__, about.__version__,
about.__models__.get(lang, lang)) about.__models__.get(lang, lang), data_path)
except (PackageNotFoundException, CompatiblePackageNotFoundException): except (PackageNotFoundException, CompatiblePackageNotFoundException):
print("Model failed to install. Please run 'python -m " print("Model failed to install. Please run 'python -m "
"spacy.%s.download --force'." % lang, file=sys.stderr) "spacy.%s.download --force'." % lang, file=sys.stderr)
sys.exit(1) sys.exit(1)
data_path = util.get_data_path()
print("Model successfully installed to %s" % data_path, file=sys.stderr) print("Model successfully installed to %s" % data_path, file=sys.stderr)

7
spacy/en/download.py
View File

@ -7,17 +7,18 @@ from .. import about
@plac.annotations( @plac.annotations(
force=("Force overwrite", "flag", "f", bool), force=("Force overwrite", "flag", "f", bool),
data_path=("Path to download model", "option", "d", str)
) )
def main(data_size='all', force=False): def main(data_size='all', force=False, data_path=None):
if force: if force:
sputnik.purge(about.__title__, about.__version__) sputnik.purge(about.__title__, about.__version__)
if data_size in ('all', 'parser'): if data_size in ('all', 'parser'):
print("Downloading parsing model") print("Downloading parsing model")
download('en', False) download('en', force=False, data_path=data_path)
if data_size in ('all', 'glove'): if data_size in ('all', 'glove'):
print("Downloading GloVe vectors") print("Downloading GloVe vectors")
download('en_glove_cc_300_1m_vectors', False) download('en_glove_cc_300_1m_vectors', force=False, data_path=data_path)
if __name__ == '__main__': if __name__ == '__main__':