Commit: f7d8b46585e1878ff9a9d1380b053e5981715063 - allura (git)

allura

Commit MetaInfo

• Revisión: f7d8b46585e1878ff9a9d1380b053e5981715063 (tree)
• Tiempo: 2012-05-15 22:25:27
• Autor: bolkimen <bolkimen@yaho...>
• Commiter: bolkimen

Log Message

ticket:53 replace patience with difflib

Cambiar Resumen

• modified: Allura/allura/controllers/repository.py
• delete: Allura/allura/lib/patience.py
• modified: Allura/allura/model/repository.py
• delete: Allura/allura/tests/test_patience.py
• modified: ForgeBlog/forgeblog/model/blog.py
• modified: ForgeTracker/forgetracker/data/ticket_changed_tmpl
• modified: ForgeTracker/forgetracker/model/ticket.py
• modified: ForgeWiki/forgewiki/model/wiki.py

Diferencia incremental

--- a/Allura/allura/controllers/repository.py
+++ b/Allura/allura/controllers/repository.py
@@ -1,6 +1,7 @@
 import os
 import json
 import logging
+import difflib
 from urllib import quote, unquote
 from collections import defaultdict
 
@@ -15,7 +16,6 @@ from ming.base import Object
 from ming.orm import ThreadLocalORMSession, session
 
 import allura.tasks
-from allura.lib import patience
 from allura.lib import security
 from allura.lib import helpers as h
 from allura.lib import widgets as w
@@ -514,7 +514,7 @@ class FileBrowser(BaseController):
         b = self._blob
         la = list(a)
         lb = list(b)
-        diff = ''.join(patience.unified_diff(
+        diff = ''.join(difflib.unified_diff(
                 la, lb,
                 ('a' + apath).encode('utf-8'),
                 ('b' + b.path()).encode('utf-8')))

--- a/Allura/allura/lib/patience.py
+++ /dev/null
@@ -1,205 +0,0 @@
-import sys
-import difflib
-from itertools import chain
-
-class Region(object):
-    '''Simple utility class that keeps track of subsequences'''
-    __slots__=('seq', # sequence
-               'l',   # lower bound
-               'h',   # upper bound
-               )
-    def __init__(self, seq, l=0, h=None):
-        if h is None: h = len(seq)
-        self.seq, self.l, self.h = seq, l, h
-
-    def __iter__(self):
-        '''Iterates over the subsequence only'''
-        for i in xrange(self.l, self.h):
-            yield self.seq[i]
-
-    def __getitem__(self, i):
-        '''works like getitem on the subsequence.  Slices return new
-        regions.'''
-        if isinstance(i, slice):
-            start, stop, step = i.indices(len(self))
-            assert step == 1
-            return self.clone(l=self.l+start,h=self.l+stop)
-        elif i >= 0:
-            return self.seq[i+self.l]
-        else:
-            return self.seq[i+self.h]
-
-    def __len__(self):
-        return self.h - self.l
-
-    def __repr__(self):
-        if len(self) > 8:
-            srepr = '[%s,...]' % (','.join(repr(x) for x in self[:5]))
-        else:
-            srepr = repr(list(self))
-        return '<Region (%s,%s): %s>' % (self.l, self.h, srepr)
-
-    def clone(self, **kw):
-        '''Return a new Region based on this one with the
-        provided kw arguments replaced in the constructor.
-        '''
-        kwargs = dict(seq=self.seq, l=self.l, h=self.h)
-        kwargs.update(kw)
-        return Region(**kwargs)
-
-def region_diff(ra, rb):
-    '''generator yielding up to two matching blocks, one at the
-    beginning of the region and one at the end.  This function
-    mutates the a and b regions, removing any matched blocks.
-    '''
-    # Yield match at the beginning
-    i = 0
-    while i < len(ra) and i < len(rb) and ra[i] == rb[i]:
-        i += 1
-    if i:
-        yield ra.l, rb.l, i
-        ra.l+=i
-        rb.l+=i
-    # Yield match at the end
-    j = 0
-    while j < len(ra) and j < len(rb) and ra[-j-1]==rb[-j-1]:
-        j+=1
-    if j:
-        yield ra.h-j, rb.h-j, j
-        ra.h-=j
-        rb.h-=j
-
-def unique(a):
-    '''generator yielding unique lines of a sequence and their positions'''
-    count = {}
-    for aa in a:
-        count[aa] = count.get(aa, 0) + 1
-    for i, aa in enumerate(a):
-        if count[aa] == 1: yield aa, i
-
-def common_unique(a, b):
-    '''generator yielding pairs i,j where
-    a[i] == b[j] and a[i] and b[j] are unique within a and b,
-    in increasing j order.'''
-    uq_a = dict(unique(a))
-    for bb, j in unique(b):
-        try:
-            yield uq_a[bb], j
-        except KeyError, ke:
-            continue
-
-def patience(seq):
-    stacks = []
-    for i, j in seq:
-        last_top = None
-        for stack in stacks:
-            top_i, top_j, top_back = stack[-1]
-            if top_i > i:
-                stack.append((i, j, last_top))
-                break
-            last_top = len(stack)-1
-        else:
-            stacks.append([(i, j, last_top)])
-    if not stacks: return []
-    prev = len(stacks[-1])-1
-    seq = []
-    for stack in reversed(stacks):
-        top_i, top_j, top_back = stack[prev]
-        seq.append((top_i, top_j))
-        prev = top_back
-    return reversed(seq)
-
-def match_core(a, b):
-    '''Returns blocks that match between sequences a and b as
-    (index_a, index_b, size)
-    '''
-    ra = Region(a)
-    rb = Region(b)
-    # beginning/end match
-    for block in region_diff(ra,rb): yield block
-    # patience core
-    last_i = last_j = None
-    for i, j in chain(
-        patience(common_unique(ra, rb)),
-        [(ra.h, rb.h)]):
-        if last_i is not None:
-            for block in region_diff(ra[last_i:i], rb[last_j:j]):
-                yield block
-        last_i = i
-        last_j = j
-
-def diff_gen(a, b, opcode_gen):
-    '''Convert a sequence of SequenceMatcher opcodes to
-    unified diff-like output
-    '''
-    def _iter():
-        for op, i1, i2, j1, j2 in opcode_gen:
-            if op == 'equal':
-                yield '  ', Region(a, i1, i2)
-            if op in ('delete', 'replace'):
-                yield '- ', Region(a, i1, i2)
-            if op in ('replace', 'insert'):
-                yield '+ ', Region(b, j1, j2)
-    for prefix, rgn in _iter():
-        for line in rgn:
-            yield prefix, line
-
-def unified_diff(
-    a, b, fromfile='', tofile='', fromfiledate='',
-    tofiledate='', n=3, lineterm='\n'):
-    started = False
-    for group in difflib.SequenceMatcher(None,a,b).get_grouped_opcodes(n):
-        if not started:
-            yield '--- %s %s%s' % (fromfile, fromfiledate, lineterm)
-            yield '+++ %s %s%s' % (tofile, tofiledate, lineterm)
-            started = True
-        i1, i2, j1, j2 = group[0][1], group[-1][2], group[0][3], group[-1][4]
-        yield "@@ -%d,%d +%d,%d @@%s" % (i1+1, i2-i1, j1+1, j2-j1, lineterm)
-        for tag, i1, i2, j1, j2 in group:
-            if tag == 'equal':
-                for line in a[i1:i2]:
-                    yield ' ' + line
-                continue
-            if tag == 'replace' or tag == 'delete':
-                for line in a[i1:i2]:
-                    yield '-' + line
-            if tag == 'replace' or tag == 'insert':
-                for line in b[j1:j2]:
-                    yield '+' + line
-
-class SequenceMatcher(difflib.SequenceMatcher):
-
-    def get_matching_blocks(self):
-        '''Uses patience diff algorithm to find matching blocks.'''
-        if self.matching_blocks is not None:
-            return self.matching_blocks
-        matching_blocks = list(match_core(self.a, self.b))
-        matching_blocks.append((len(self.a), len(self.b), 0))
-        self.matching_blocks = sorted(matching_blocks)
-        return self.matching_blocks
-
-def test(): # pragma no cover
-    if len(sys.argv) == 3:
-        fn_a = sys.argv[1]
-        fn_b = sys.argv[2]
-    else:
-        fn_a = 'a.c'
-        fn_b = 'b.c'
-    a = open(fn_a).readlines()
-    b = open(fn_b).readlines()
-    # print '====', fn_a
-    # sys.stdout.write(''.join(a))
-    # print '====', fn_b
-    # sys.stdout.write(''.join(b))
-    sm = SequenceMatcher(None, a, b)
-    # print 'Patience opcodes:', sm.get_opcodes()
-    print ''.join(unified_diff(a, b)) #pragma:printok
-    # for prefix, line in diff_gen(a, b, sm.get_opcodes()):
-    #     sys.stdout.write(''.join((prefix, line)))
-    # sm = difflib.SequenceMatcher(None, a, b)
-    # print 'Difflib opcodes:', sm.get_opcodes()
-    # for prefix, line in diff_gen(a, b, sm.get_opcodes()):
-    #     sys.stdout.write(''.join((prefix, line)))
-
-if __name__ == '__main__': # pragma no cover
-    test()

--- a/Allura/allura/model/repository.py
+++ b/Allura/allura/model/repository.py
@@ -5,6 +5,7 @@ import mimetypes
 import logging
 import string
 import re
+from difflib import SequenceMatcher
 from hashlib import sha1
 from datetime import datetime
 from collections import defaultdict
@@ -20,7 +21,6 @@ from ming.utils import LazyProperty
 from ming.orm import FieldProperty, session, Mapper
 from ming.orm.declarative import MappedClass
 
-from allura.lib.patience import SequenceMatcher
 from allura.lib import helpers as h
 from allura.lib import utils
 

--- a/Allura/allura/tests/test_patience.py
+++ /dev/null
@@ -1,76 +0,0 @@
-from os import path, environ
-from collections import defaultdict
-
-from allura.lib import patience
-
-def text2lines(text):
-    return [l + '\n' for l in text.split('\n')]
-
-def test_region():
-    r = patience.Region('foobar')
-    r2 = r.clone()
-    assert id(r) != id(r2)
-    assert '-'.join(r) == '-'.join(r2)
-    subr = r[1:5]
-    assert type(subr) is type(r)
-    assert ''.join(subr) == ''.join(r)[1:5]
-    repr(r)
-    repr(patience.Region('fffffffffffffffffffffffffffffffffffffffff'))
-
-def test_unified_diff():
-    text1 = '''\
-from paste.deploy import loadapp
-from paste.deploy import loadapp
-from paste.deploy import loadapp
-from paste.deploy import loadapp
-from paste.script.appinstall import SetupCommand
-from paste.script.appinstall import SetupCommand
-from paste.script.appinstall import SetupCommand
-from paste.script.appinstall import SetupCommand
-from paste.deploy import appconfig
-'''
-    text2 = '''\
-from paste.deploy import loadapp
-from paste.deploy import loadapp
-from paste.deploy import loadapp
-from paste.deploy import loadapp
-from paste.script.appinstall import SetupCommand2
-from paste.script.appinstall import SetupCommand3
-from paste.script.appinstall import SetupCommand4
-from paste.deploy import appconfig
-'''
-    line_uni_diff = '''\
- from paste.deploy import loadapp
- from paste.deploy import loadapp
- from paste.deploy import loadapp
--from paste.script.appinstall import SetupCommand
--from paste.script.appinstall import SetupCommand
--from paste.script.appinstall import SetupCommand
--from paste.script.appinstall import SetupCommand
-+from paste.script.appinstall import SetupCommand2
-+from paste.script.appinstall import SetupCommand3
-+from paste.script.appinstall import SetupCommand4
- from paste.deploy import appconfig'''
-
-    line_diff = '''\
- from paste.deploy import loadapp
-''' + line_uni_diff
-
-    lines1 = text2lines(text1)
-    lines2 = text2lines(text2)
-    diff = patience.unified_diff(lines1, lines2)
-    diff = ''.join(diff)
-    assert diff == '''\
----  
-+++  
-@@ -2,9 +2,8 @@
-%s
- 
-''' % line_uni_diff, '=' + diff + '='
-
-    sm = patience.SequenceMatcher(None, lines1, lines2)
-    buf = ''
-    for prefix, line in patience.diff_gen(lines1, lines2, sm.get_opcodes()):
-        assert prefix[1] == ' '
-        buf += prefix[0] + line
-    assert buf == line_diff + '\n \n', '=' + buf + '='

--- a/ForgeBlog/forgeblog/model/blog.py
+++ b/ForgeBlog/forgeblog/model/blog.py
@@ -1,3 +1,4 @@
+import difflib
 from datetime import datetime
 from random import randint
 
@@ -11,7 +12,7 @@ from ming import schema
 from ming.orm import FieldProperty, ForeignIdProperty, Mapper, session, state
 from allura import model as M
 from allura.lib import helpers as h
-from allura.lib import utils, patience
+from allura.lib import utils
 
 config = utils.ConfigProxy(
     common_suffix='forgemail.domain')
@@ -166,7 +167,7 @@ class BlogPost(M.VersionedArtifact):
             v2 = self
             la = [ line + '\n'  for line in v1.text.splitlines() ]
             lb = [ line + '\n'  for line in v2.text.splitlines() ]
-            diff = ''.join(patience.unified_diff(
+            diff = ''.join(difflib.unified_diff(
                     la, lb,
                     'v%d' % v1.version,
                     'v%d' % v2.version))

--- a/ForgeTracker/forgetracker/data/ticket_changed_tmpl
+++ b/ForgeTracker/forgetracker/data/ticket_changed_tmpl
@@ -1,4 +1,4 @@
-{% python from allura.lib import patience %}\
+{% python import difflib %}\
 {% python from allura.model import User %}\
 {% for key, values in changelist %}\
 {% choose %}\
@@ -9,7 +9,7 @@ Diff:
 
 ~~~~
 
-${'\n'.join(patience.unified_diff(
+${'\n'.join(difflib.unified_diff(
                             a=values[0].splitlines(),
                             b=values[1].splitlines(),
                             fromfile='old',

--- a/ForgeTracker/forgetracker/model/ticket.py
+++ b/ForgeTracker/forgetracker/model/ticket.py
@@ -1,6 +1,7 @@
 import logging
 import urllib
 import json
+import difflib
 from datetime import datetime, timedelta
 
 import bson
@@ -20,7 +21,6 @@ from ming.orm.declarative import MappedClass
 from allura.model import Artifact, VersionedArtifact, Snapshot, project_orm_session, BaseAttachment
 from allura.model import User, Feed, Thread, Notification, ProjectRole
 from allura.model import ACE, ALL_PERMISSIONS, DENY_ALL
-from allura.lib import patience
 from allura.lib import security
 from allura.lib.search import search_artifact
 from allura.lib import utils
@@ -392,7 +392,7 @@ class Ticket(VersionedArtifact):
             if old.description != self.description:
                 changes.append('Description updated:')
                 changes.append('\n'.join(
-                        patience.unified_diff(
+                        difflib.unified_diff(
                             a=old.description.split('\n'),
                             b=self.description.split('\n'),
                             fromfile='description-old',

--- a/ForgeWiki/forgewiki/model/wiki.py
+++ b/ForgeWiki/forgewiki/model/wiki.py
@@ -1,4 +1,5 @@
 import pylons
+import difflib
 pylons.c = pylons.tmpl_context
 pylons.g = pylons.app_globals
 from pylons import g #g is a namespace for globally accessable app helpers
@@ -11,7 +12,6 @@ from ming.orm.declarative import MappedClass
 from allura.model import VersionedArtifact, Snapshot, Feed, Thread, Post, User, BaseAttachment
 from allura.model import Notification, project_orm_session
 from allura.lib import helpers as h
-from allura.lib import patience
 from allura.lib import utils
 
 config = utils.ConfigProxy(
@@ -86,7 +86,7 @@ class Page(VersionedArtifact):
             v2 = self
             la = [ line + '\n'  for line in v1.text.splitlines() ]
             lb = [ line + '\n'  for line in v2.text.splitlines() ]
-            diff = ''.join(patience.unified_diff(
+            diff = ''.join(difflib.unified_diff(
                     la, lb,
                     'v%d' % v1.version,
                     'v%d' % v2.version))