Archives/wting_autojump
1
0
Fork 0
mirror of https://github.com/wting/autojump synced 2024-10-27 20:34:07 +00:00
Code Issues Releases Wiki Activity

Implement approximate matching via Levenshtein distance.

Browse Source
This commit is contained in:
jez 2011-05-27 07:56:48 +00:00
parent 2a93e3c570
commit 9b977379eb
1 changed files with 60 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

82
autojump
View File

@ -30,6 +30,7 @@ import getopt
from sys import argv, stderr, version_info, exit
from tempfile import NamedTemporaryFile
from operator import itemgetter
from copy import copy
import os
MAX_KEYWEIGHT = 1000
MAX_STORED_PATHS = 600
@ -88,30 +89,62 @@ def clean_dict(sorted_dirs, path_dict):
return True
else: return False
def match(path, pattern, ignore_case=False, only_end=False):
"""Check whether a path matches a particular pattern"""
if only_end:
match_string = "/".join(path.split('/')[-1-pattern.count('/'):])
else:
match_string = path
if ignore_case:
does_match = (match_string.lower().find(pattern.lower()) != -1)
else:
does_match = (match_string.find(pattern) != -1)
#return True if there is a match and the path exists
#(useful in the case of external drives, for example)
return does_match and os.path.exists(path)
def approximatch(pat, text):
prev_col = list(range(0, len(pat)+1))
col = [0] * (len(pat) + 1)
errors = len(pat)
for char1 in text:
col[0] = 0
for i, char2 in enumerate(pat):
if char1 == char2:
col[i+1] = prev_col[i]
else:
col[i+1] = 1 + min(col[i], prev_col[i+1], prev_col[i])
prev_col = copy(col)
errors = min(errors, col[-1])
return errors
def find_matches(dirs, patterns, result_list, ignore_case, max_matches):
def find_matches(dirs, patterns, result_list, ignore_case, approx, max_matches):
"""Find max_matches paths that match the pattern,
and add them to the result_list"""
for path, count in dirs:
if len(result_list) >= max_matches :
break
def get_pattern_and_match(patterns, path):
#For the last pattern, only match the end of the pattern
if all(match(path, p, ignore_case,
only_end=(n == len(patterns)-1)) for n, p in enumerate(patterns)):
uniqadd(result_list, path)
for n, pattern in enumerate(patterns):
if n == len(patterns) - 1:
match_string = "/".join(path.split('/')[-1-pattern.count('/'):])
else:
match_string = path
if ignore_case:
pattern = pattern.lower()
match_string = match_string.lower()
yield (pattern, match_string)
if approx:
one_error_paths = []
two_error_paths = []
for path, count in dirs:
if len(one_error_paths) >= max_matches:
break
errors = sum(approximatch(pattern, match_string)
for pattern, match_string in get_pattern_and_match(patterns, path))
#Verify that the path exists
#(useful in the case of external drives, for example)
if errors <= 2 and os.path.exists(path):
if errors == 1:
uniqadd(one_error_paths, path)
elif errors == 2:
uniqadd(two_error_paths, path)
result_list.extend(one_error_paths)
result_list.extend(two_error_paths[:max_matches-len(one_error_paths)])
else:
for path, count in dirs:
if len(result_list) >= max_matches:
break
if all(match_string.find(pattern) != -1
for pattern, match_string in
get_pattern_and_match(patterns, path)) and os.path.exists(path):
uniqadd(result_list, path)
def open_dic(dic_file, error_recovery=False):
"""Try hard to open the database file, recovering
@ -206,12 +239,17 @@ def shell_utility():
max_matches = 9
else:
max_matches = 1
find_matches(dirs, patterns, results, False, max_matches)
find_matches(dirs, patterns, results, False, False, max_matches)
# If not found, try ignoring case.
# On completion always show all results
if completion or not results:
find_matches(dirs, patterns, results,
ignore_case=True, max_matches=max_matches)
ignore_case=True, approx=False, max_matches=max_matches)
if not results:
find_matches(dirs, patterns, results,
ignore_case=True, approx=True, max_matches=max_matches)
# Keep the database to a reasonable size
if not completion and clean_dict(dirs, path_dict):
save(path_dict, dic_file)