aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorMartin Czygan <martin.czygan@gmail.com>2020-10-21 10:57:55 +0200
committerMartin Czygan <martin.czygan@gmail.com>2020-10-21 10:57:55 +0200
commitf9f28047f1f30f75afdce91e3a18adb3eb7f5dfb (patch)
tree22e9fd0cb9ba0a8fc6b9c62ab52f880e1e367662
parent0cc68900cc160dc934233825d02ee7deaff3b439 (diff)
downloadfuzzycat-f9f28047f1f30f75afdce91e3a18adb3eb7f5dfb.tar.gz
fuzzycat-f9f28047f1f30f75afdce91e3a18adb3eb7f5dfb.zip
use custom tmp prefix
-rw-r--r--fuzzycat/cluster.py13
1 files changed, 7 insertions, 6 deletions
diff --git a/fuzzycat/cluster.py b/fuzzycat/cluster.py
index 9b97509..32faec6 100644
--- a/fuzzycat/cluster.py
+++ b/fuzzycat/cluster.py
@@ -24,13 +24,14 @@ import orjson as json
import fuzzy
DEFAULT_CACHE_DIR = os.path.join(os.path.expanduser("~"), ".cache", "fuzzycat")
+TMP_PREFIX = 'fuzzycat-'
-def sort_by_column(filename, mode="w", opts="-k 2", fast=True):
+def sort_by_column(filename, mode="w", opts="-k 2", fast=True, prefix=TMP_PREFIX):
"""
Sort tabular file with sort(1), returns the filename of the sorted file.
"""
- with tempfile.NamedTemporaryFile(delete=False, mode=mode) as tf:
+ with tempfile.NamedTemporaryFile(delete=False, mode=mode, prefix=prefix) as tf:
env = os.environ.copy()
if fast:
env["LC_ALL"] = "C"
@@ -60,12 +61,12 @@ def cut(f=0, sep='\t'):
"""
return lambda v: v.split(sep)[f]
-def cluster_by_title(args):
+def cluster_by_title(args, prefix=TMP_PREFIX):
"""
Basic example for a three stage process: extract, sort, group. Speed is
about: 20K/s (json roundtrip, sorting, grouping).
"""
- with tempfile.NamedTemporaryFile(delete=False, mode="w") as tf:
+ with tempfile.NamedTemporaryFile(delete=False, mode="w", prefix=prefix) as tf:
for line in fileinput.input(files=args.files if len(args.files) > 0 else ('-', )):
doc = json.loads(line)
try:
@@ -91,7 +92,7 @@ def cluster_by_title_normalized(args):
Normalize title, e.g. analysisofheritability. 17k/s.
"""
pattern = re.compile('[\W_]+', re.UNICODE)
- with tempfile.NamedTemporaryFile(delete=False, mode="w") as tf:
+ with tempfile.NamedTemporaryFile(delete=False, mode="w", prefix=prefix) as tf:
for line in fileinput.input(files=args.files if len(args.files) > 0 else ('-', )):
doc = json.loads(line)
try:
@@ -117,7 +118,7 @@ def cluster_by_title_nysiis(args):
"""
Soundex on title.
"""
- with tempfile.NamedTemporaryFile(delete=False, mode="w") as tf:
+ with tempfile.NamedTemporaryFile(delete=False, mode="w", prefix=prefix) as tf:
for line in fileinput.input(files=args.files if len(args.files) > 0 else ('-', )):
doc = json.loads(line)
try: