| 4710c53d |
1 | #! /usr/bin/env python\r |
| 2 | \r |
| 3 | """\r |
| 4 | Module difflib -- helpers for computing deltas between objects.\r |
| 5 | \r |
| 6 | Function get_close_matches(word, possibilities, n=3, cutoff=0.6):\r |
| 7 | Use SequenceMatcher to return list of the best "good enough" matches.\r |
| 8 | \r |
| 9 | Function context_diff(a, b):\r |
| 10 | For two lists of strings, return a delta in context diff format.\r |
| 11 | \r |
| 12 | Function ndiff(a, b):\r |
| 13 | Return a delta: the difference between `a` and `b` (lists of strings).\r |
| 14 | \r |
| 15 | Function restore(delta, which):\r |
| 16 | Return one of the two sequences that generated an ndiff delta.\r |
| 17 | \r |
| 18 | Function unified_diff(a, b):\r |
| 19 | For two lists of strings, return a delta in unified diff format.\r |
| 20 | \r |
| 21 | Class SequenceMatcher:\r |
| 22 | A flexible class for comparing pairs of sequences of any type.\r |
| 23 | \r |
| 24 | Class Differ:\r |
| 25 | For producing human-readable deltas from sequences of lines of text.\r |
| 26 | \r |
| 27 | Class HtmlDiff:\r |
| 28 | For producing HTML side by side comparison with change highlights.\r |
| 29 | """\r |
| 30 | \r |
| 31 | __all__ = ['get_close_matches', 'ndiff', 'restore', 'SequenceMatcher',\r |
| 32 | 'Differ','IS_CHARACTER_JUNK', 'IS_LINE_JUNK', 'context_diff',\r |
| 33 | 'unified_diff', 'HtmlDiff', 'Match']\r |
| 34 | \r |
| 35 | import heapq\r |
| 36 | from collections import namedtuple as _namedtuple\r |
| 37 | from functools import reduce\r |
| 38 | \r |
| 39 | Match = _namedtuple('Match', 'a b size')\r |
| 40 | \r |
| 41 | def _calculate_ratio(matches, length):\r |
| 42 | if length:\r |
| 43 | return 2.0 * matches / length\r |
| 44 | return 1.0\r |
| 45 | \r |
| 46 | class SequenceMatcher:\r |
| 47 | \r |
| 48 | """\r |
| 49 | SequenceMatcher is a flexible class for comparing pairs of sequences of\r |
| 50 | any type, so long as the sequence elements are hashable. The basic\r |
| 51 | algorithm predates, and is a little fancier than, an algorithm\r |
| 52 | published in the late 1980's by Ratcliff and Obershelp under the\r |
| 53 | hyperbolic name "gestalt pattern matching". The basic idea is to find\r |
| 54 | the longest contiguous matching subsequence that contains no "junk"\r |
| 55 | elements (R-O doesn't address junk). The same idea is then applied\r |
| 56 | recursively to the pieces of the sequences to the left and to the right\r |
| 57 | of the matching subsequence. This does not yield minimal edit\r |
| 58 | sequences, but does tend to yield matches that "look right" to people.\r |
| 59 | \r |
| 60 | SequenceMatcher tries to compute a "human-friendly diff" between two\r |
| 61 | sequences. Unlike e.g. UNIX(tm) diff, the fundamental notion is the\r |
| 62 | longest *contiguous* & junk-free matching subsequence. That's what\r |
| 63 | catches peoples' eyes. The Windows(tm) windiff has another interesting\r |
| 64 | notion, pairing up elements that appear uniquely in each sequence.\r |
| 65 | That, and the method here, appear to yield more intuitive difference\r |
| 66 | reports than does diff. This method appears to be the least vulnerable\r |
| 67 | to synching up on blocks of "junk lines", though (like blank lines in\r |
| 68 | ordinary text files, or maybe "<P>" lines in HTML files). That may be\r |
| 69 | because this is the only method of the 3 that has a *concept* of\r |
| 70 | "junk" <wink>.\r |
| 71 | \r |
| 72 | Example, comparing two strings, and considering blanks to be "junk":\r |
| 73 | \r |
| 74 | >>> s = SequenceMatcher(lambda x: x == " ",\r |
| 75 | ... "private Thread currentThread;",\r |
| 76 | ... "private volatile Thread currentThread;")\r |
| 77 | >>>\r |
| 78 | \r |
| 79 | .ratio() returns a float in [0, 1], measuring the "similarity" of the\r |
| 80 | sequences. As a rule of thumb, a .ratio() value over 0.6 means the\r |
| 81 | sequences are close matches:\r |
| 82 | \r |
| 83 | >>> print round(s.ratio(), 3)\r |
| 84 | 0.866\r |
| 85 | >>>\r |
| 86 | \r |
| 87 | If you're only interested in where the sequences match,\r |
| 88 | .get_matching_blocks() is handy:\r |
| 89 | \r |
| 90 | >>> for block in s.get_matching_blocks():\r |
| 91 | ... print "a[%d] and b[%d] match for %d elements" % block\r |
| 92 | a[0] and b[0] match for 8 elements\r |
| 93 | a[8] and b[17] match for 21 elements\r |
| 94 | a[29] and b[38] match for 0 elements\r |
| 95 | \r |
| 96 | Note that the last tuple returned by .get_matching_blocks() is always a\r |
| 97 | dummy, (len(a), len(b), 0), and this is the only case in which the last\r |
| 98 | tuple element (number of elements matched) is 0.\r |
| 99 | \r |
| 100 | If you want to know how to change the first sequence into the second,\r |
| 101 | use .get_opcodes():\r |
| 102 | \r |
| 103 | >>> for opcode in s.get_opcodes():\r |
| 104 | ... print "%6s a[%d:%d] b[%d:%d]" % opcode\r |
| 105 | equal a[0:8] b[0:8]\r |
| 106 | insert a[8:8] b[8:17]\r |
| 107 | equal a[8:29] b[17:38]\r |
| 108 | \r |
| 109 | See the Differ class for a fancy human-friendly file differencer, which\r |
| 110 | uses SequenceMatcher both to compare sequences of lines, and to compare\r |
| 111 | sequences of characters within similar (near-matching) lines.\r |
| 112 | \r |
| 113 | See also function get_close_matches() in this module, which shows how\r |
| 114 | simple code building on SequenceMatcher can be used to do useful work.\r |
| 115 | \r |
| 116 | Timing: Basic R-O is cubic time worst case and quadratic time expected\r |
| 117 | case. SequenceMatcher is quadratic time for the worst case and has\r |
| 118 | expected-case behavior dependent in a complicated way on how many\r |
| 119 | elements the sequences have in common; best case time is linear.\r |
| 120 | \r |
| 121 | Methods:\r |
| 122 | \r |
| 123 | __init__(isjunk=None, a='', b='')\r |
| 124 | Construct a SequenceMatcher.\r |
| 125 | \r |
| 126 | set_seqs(a, b)\r |
| 127 | Set the two sequences to be compared.\r |
| 128 | \r |
| 129 | set_seq1(a)\r |
| 130 | Set the first sequence to be compared.\r |
| 131 | \r |
| 132 | set_seq2(b)\r |
| 133 | Set the second sequence to be compared.\r |
| 134 | \r |
| 135 | find_longest_match(alo, ahi, blo, bhi)\r |
| 136 | Find longest matching block in a[alo:ahi] and b[blo:bhi].\r |
| 137 | \r |
| 138 | get_matching_blocks()\r |
| 139 | Return list of triples describing matching subsequences.\r |
| 140 | \r |
| 141 | get_opcodes()\r |
| 142 | Return list of 5-tuples describing how to turn a into b.\r |
| 143 | \r |
| 144 | ratio()\r |
| 145 | Return a measure of the sequences' similarity (float in [0,1]).\r |
| 146 | \r |
| 147 | quick_ratio()\r |
| 148 | Return an upper bound on .ratio() relatively quickly.\r |
| 149 | \r |
| 150 | real_quick_ratio()\r |
| 151 | Return an upper bound on ratio() very quickly.\r |
| 152 | """\r |
| 153 | \r |
| 154 | def __init__(self, isjunk=None, a='', b='', autojunk=True):\r |
| 155 | """Construct a SequenceMatcher.\r |
| 156 | \r |
| 157 | Optional arg isjunk is None (the default), or a one-argument\r |
| 158 | function that takes a sequence element and returns true iff the\r |
| 159 | element is junk. None is equivalent to passing "lambda x: 0", i.e.\r |
| 160 | no elements are considered to be junk. For example, pass\r |
| 161 | lambda x: x in " \\t"\r |
| 162 | if you're comparing lines as sequences of characters, and don't\r |
| 163 | want to synch up on blanks or hard tabs.\r |
| 164 | \r |
| 165 | Optional arg a is the first of two sequences to be compared. By\r |
| 166 | default, an empty string. The elements of a must be hashable. See\r |
| 167 | also .set_seqs() and .set_seq1().\r |
| 168 | \r |
| 169 | Optional arg b is the second of two sequences to be compared. By\r |
| 170 | default, an empty string. The elements of b must be hashable. See\r |
| 171 | also .set_seqs() and .set_seq2().\r |
| 172 | \r |
| 173 | Optional arg autojunk should be set to False to disable the\r |
| 174 | "automatic junk heuristic" that treats popular elements as junk\r |
| 175 | (see module documentation for more information).\r |
| 176 | """\r |
| 177 | \r |
| 178 | # Members:\r |
| 179 | # a\r |
| 180 | # first sequence\r |
| 181 | # b\r |
| 182 | # second sequence; differences are computed as "what do\r |
| 183 | # we need to do to 'a' to change it into 'b'?"\r |
| 184 | # b2j\r |
| 185 | # for x in b, b2j[x] is a list of the indices (into b)\r |
| 186 | # at which x appears; junk elements do not appear\r |
| 187 | # fullbcount\r |
| 188 | # for x in b, fullbcount[x] == the number of times x\r |
| 189 | # appears in b; only materialized if really needed (used\r |
| 190 | # only for computing quick_ratio())\r |
| 191 | # matching_blocks\r |
| 192 | # a list of (i, j, k) triples, where a[i:i+k] == b[j:j+k];\r |
| 193 | # ascending & non-overlapping in i and in j; terminated by\r |
| 194 | # a dummy (len(a), len(b), 0) sentinel\r |
| 195 | # opcodes\r |
| 196 | # a list of (tag, i1, i2, j1, j2) tuples, where tag is\r |
| 197 | # one of\r |
| 198 | # 'replace' a[i1:i2] should be replaced by b[j1:j2]\r |
| 199 | # 'delete' a[i1:i2] should be deleted\r |
| 200 | # 'insert' b[j1:j2] should be inserted\r |
| 201 | # 'equal' a[i1:i2] == b[j1:j2]\r |
| 202 | # isjunk\r |
| 203 | # a user-supplied function taking a sequence element and\r |
| 204 | # returning true iff the element is "junk" -- this has\r |
| 205 | # subtle but helpful effects on the algorithm, which I'll\r |
| 206 | # get around to writing up someday <0.9 wink>.\r |
| 207 | # DON'T USE! Only __chain_b uses this. Use isbjunk.\r |
| 208 | # isbjunk\r |
| 209 | # for x in b, isbjunk(x) == isjunk(x) but much faster;\r |
| 210 | # it's really the __contains__ method of a hidden dict.\r |
| 211 | # DOES NOT WORK for x in a!\r |
| 212 | # isbpopular\r |
| 213 | # for x in b, isbpopular(x) is true iff b is reasonably long\r |
| 214 | # (at least 200 elements) and x accounts for more than 1 + 1% of\r |
| 215 | # its elements (when autojunk is enabled).\r |
| 216 | # DOES NOT WORK for x in a!\r |
| 217 | \r |
| 218 | self.isjunk = isjunk\r |
| 219 | self.a = self.b = None\r |
| 220 | self.autojunk = autojunk\r |
| 221 | self.set_seqs(a, b)\r |
| 222 | \r |
| 223 | def set_seqs(self, a, b):\r |
| 224 | """Set the two sequences to be compared.\r |
| 225 | \r |
| 226 | >>> s = SequenceMatcher()\r |
| 227 | >>> s.set_seqs("abcd", "bcde")\r |
| 228 | >>> s.ratio()\r |
| 229 | 0.75\r |
| 230 | """\r |
| 231 | \r |
| 232 | self.set_seq1(a)\r |
| 233 | self.set_seq2(b)\r |
| 234 | \r |
| 235 | def set_seq1(self, a):\r |
| 236 | """Set the first sequence to be compared.\r |
| 237 | \r |
| 238 | The second sequence to be compared is not changed.\r |
| 239 | \r |
| 240 | >>> s = SequenceMatcher(None, "abcd", "bcde")\r |
| 241 | >>> s.ratio()\r |
| 242 | 0.75\r |
| 243 | >>> s.set_seq1("bcde")\r |
| 244 | >>> s.ratio()\r |
| 245 | 1.0\r |
| 246 | >>>\r |
| 247 | \r |
| 248 | SequenceMatcher computes and caches detailed information about the\r |
| 249 | second sequence, so if you want to compare one sequence S against\r |
| 250 | many sequences, use .set_seq2(S) once and call .set_seq1(x)\r |
| 251 | repeatedly for each of the other sequences.\r |
| 252 | \r |
| 253 | See also set_seqs() and set_seq2().\r |
| 254 | """\r |
| 255 | \r |
| 256 | if a is self.a:\r |
| 257 | return\r |
| 258 | self.a = a\r |
| 259 | self.matching_blocks = self.opcodes = None\r |
| 260 | \r |
| 261 | def set_seq2(self, b):\r |
| 262 | """Set the second sequence to be compared.\r |
| 263 | \r |
| 264 | The first sequence to be compared is not changed.\r |
| 265 | \r |
| 266 | >>> s = SequenceMatcher(None, "abcd", "bcde")\r |
| 267 | >>> s.ratio()\r |
| 268 | 0.75\r |
| 269 | >>> s.set_seq2("abcd")\r |
| 270 | >>> s.ratio()\r |
| 271 | 1.0\r |
| 272 | >>>\r |
| 273 | \r |
| 274 | SequenceMatcher computes and caches detailed information about the\r |
| 275 | second sequence, so if you want to compare one sequence S against\r |
| 276 | many sequences, use .set_seq2(S) once and call .set_seq1(x)\r |
| 277 | repeatedly for each of the other sequences.\r |
| 278 | \r |
| 279 | See also set_seqs() and set_seq1().\r |
| 280 | """\r |
| 281 | \r |
| 282 | if b is self.b:\r |
| 283 | return\r |
| 284 | self.b = b\r |
| 285 | self.matching_blocks = self.opcodes = None\r |
| 286 | self.fullbcount = None\r |
| 287 | self.__chain_b()\r |
| 288 | \r |
| 289 | # For each element x in b, set b2j[x] to a list of the indices in\r |
| 290 | # b where x appears; the indices are in increasing order; note that\r |
| 291 | # the number of times x appears in b is len(b2j[x]) ...\r |
| 292 | # when self.isjunk is defined, junk elements don't show up in this\r |
| 293 | # map at all, which stops the central find_longest_match method\r |
| 294 | # from starting any matching block at a junk element ...\r |
| 295 | # also creates the fast isbjunk function ...\r |
| 296 | # b2j also does not contain entries for "popular" elements, meaning\r |
| 297 | # elements that account for more than 1 + 1% of the total elements, and\r |
| 298 | # when the sequence is reasonably large (>= 200 elements); this can\r |
| 299 | # be viewed as an adaptive notion of semi-junk, and yields an enormous\r |
| 300 | # speedup when, e.g., comparing program files with hundreds of\r |
| 301 | # instances of "return NULL;" ...\r |
| 302 | # note that this is only called when b changes; so for cross-product\r |
| 303 | # kinds of matches, it's best to call set_seq2 once, then set_seq1\r |
| 304 | # repeatedly\r |
| 305 | \r |
| 306 | def __chain_b(self):\r |
| 307 | # Because isjunk is a user-defined (not C) function, and we test\r |
| 308 | # for junk a LOT, it's important to minimize the number of calls.\r |
| 309 | # Before the tricks described here, __chain_b was by far the most\r |
| 310 | # time-consuming routine in the whole module! If anyone sees\r |
| 311 | # Jim Roskind, thank him again for profile.py -- I never would\r |
| 312 | # have guessed that.\r |
| 313 | # The first trick is to build b2j ignoring the possibility\r |
| 314 | # of junk. I.e., we don't call isjunk at all yet. Throwing\r |
| 315 | # out the junk later is much cheaper than building b2j "right"\r |
| 316 | # from the start.\r |
| 317 | b = self.b\r |
| 318 | self.b2j = b2j = {}\r |
| 319 | \r |
| 320 | for i, elt in enumerate(b):\r |
| 321 | indices = b2j.setdefault(elt, [])\r |
| 322 | indices.append(i)\r |
| 323 | \r |
| 324 | # Purge junk elements\r |
| 325 | junk = set()\r |
| 326 | isjunk = self.isjunk\r |
| 327 | if isjunk:\r |
| 328 | for elt in list(b2j.keys()): # using list() since b2j is modified\r |
| 329 | if isjunk(elt):\r |
| 330 | junk.add(elt)\r |
| 331 | del b2j[elt]\r |
| 332 | \r |
| 333 | # Purge popular elements that are not junk\r |
| 334 | popular = set()\r |
| 335 | n = len(b)\r |
| 336 | if self.autojunk and n >= 200:\r |
| 337 | ntest = n // 100 + 1\r |
| 338 | for elt, idxs in list(b2j.items()):\r |
| 339 | if len(idxs) > ntest:\r |
| 340 | popular.add(elt)\r |
| 341 | del b2j[elt]\r |
| 342 | \r |
| 343 | # Now for x in b, isjunk(x) == x in junk, but the latter is much faster.\r |
| 344 | # Sicne the number of *unique* junk elements is probably small, the\r |
| 345 | # memory burden of keeping this set alive is likely trivial compared to\r |
| 346 | # the size of b2j.\r |
| 347 | self.isbjunk = junk.__contains__\r |
| 348 | self.isbpopular = popular.__contains__\r |
| 349 | \r |
| 350 | def find_longest_match(self, alo, ahi, blo, bhi):\r |
| 351 | """Find longest matching block in a[alo:ahi] and b[blo:bhi].\r |
| 352 | \r |
| 353 | If isjunk is not defined:\r |
| 354 | \r |
| 355 | Return (i,j,k) such that a[i:i+k] is equal to b[j:j+k], where\r |
| 356 | alo <= i <= i+k <= ahi\r |
| 357 | blo <= j <= j+k <= bhi\r |
| 358 | and for all (i',j',k') meeting those conditions,\r |
| 359 | k >= k'\r |
| 360 | i <= i'\r |
| 361 | and if i == i', j <= j'\r |
| 362 | \r |
| 363 | In other words, of all maximal matching blocks, return one that\r |
| 364 | starts earliest in a, and of all those maximal matching blocks that\r |
| 365 | start earliest in a, return the one that starts earliest in b.\r |
| 366 | \r |
| 367 | >>> s = SequenceMatcher(None, " abcd", "abcd abcd")\r |
| 368 | >>> s.find_longest_match(0, 5, 0, 9)\r |
| 369 | Match(a=0, b=4, size=5)\r |
| 370 | \r |
| 371 | If isjunk is defined, first the longest matching block is\r |
| 372 | determined as above, but with the additional restriction that no\r |
| 373 | junk element appears in the block. Then that block is extended as\r |
| 374 | far as possible by matching (only) junk elements on both sides. So\r |
| 375 | the resulting block never matches on junk except as identical junk\r |
| 376 | happens to be adjacent to an "interesting" match.\r |
| 377 | \r |
| 378 | Here's the same example as before, but considering blanks to be\r |
| 379 | junk. That prevents " abcd" from matching the " abcd" at the tail\r |
| 380 | end of the second sequence directly. Instead only the "abcd" can\r |
| 381 | match, and matches the leftmost "abcd" in the second sequence:\r |
| 382 | \r |
| 383 | >>> s = SequenceMatcher(lambda x: x==" ", " abcd", "abcd abcd")\r |
| 384 | >>> s.find_longest_match(0, 5, 0, 9)\r |
| 385 | Match(a=1, b=0, size=4)\r |
| 386 | \r |
| 387 | If no blocks match, return (alo, blo, 0).\r |
| 388 | \r |
| 389 | >>> s = SequenceMatcher(None, "ab", "c")\r |
| 390 | >>> s.find_longest_match(0, 2, 0, 1)\r |
| 391 | Match(a=0, b=0, size=0)\r |
| 392 | """\r |
| 393 | \r |
| 394 | # CAUTION: stripping common prefix or suffix would be incorrect.\r |
| 395 | # E.g.,\r |
| 396 | # ab\r |
| 397 | # acab\r |
| 398 | # Longest matching block is "ab", but if common prefix is\r |
| 399 | # stripped, it's "a" (tied with "b"). UNIX(tm) diff does so\r |
| 400 | # strip, so ends up claiming that ab is changed to acab by\r |
| 401 | # inserting "ca" in the middle. That's minimal but unintuitive:\r |
| 402 | # "it's obvious" that someone inserted "ac" at the front.\r |
| 403 | # Windiff ends up at the same place as diff, but by pairing up\r |
| 404 | # the unique 'b's and then matching the first two 'a's.\r |
| 405 | \r |
| 406 | a, b, b2j, isbjunk = self.a, self.b, self.b2j, self.isbjunk\r |
| 407 | besti, bestj, bestsize = alo, blo, 0\r |
| 408 | # find longest junk-free match\r |
| 409 | # during an iteration of the loop, j2len[j] = length of longest\r |
| 410 | # junk-free match ending with a[i-1] and b[j]\r |
| 411 | j2len = {}\r |
| 412 | nothing = []\r |
| 413 | for i in xrange(alo, ahi):\r |
| 414 | # look at all instances of a[i] in b; note that because\r |
| 415 | # b2j has no junk keys, the loop is skipped if a[i] is junk\r |
| 416 | j2lenget = j2len.get\r |
| 417 | newj2len = {}\r |
| 418 | for j in b2j.get(a[i], nothing):\r |
| 419 | # a[i] matches b[j]\r |
| 420 | if j < blo:\r |
| 421 | continue\r |
| 422 | if j >= bhi:\r |
| 423 | break\r |
| 424 | k = newj2len[j] = j2lenget(j-1, 0) + 1\r |
| 425 | if k > bestsize:\r |
| 426 | besti, bestj, bestsize = i-k+1, j-k+1, k\r |
| 427 | j2len = newj2len\r |
| 428 | \r |
| 429 | # Extend the best by non-junk elements on each end. In particular,\r |
| 430 | # "popular" non-junk elements aren't in b2j, which greatly speeds\r |
| 431 | # the inner loop above, but also means "the best" match so far\r |
| 432 | # doesn't contain any junk *or* popular non-junk elements.\r |
| 433 | while besti > alo and bestj > blo and \\r |
| 434 | not isbjunk(b[bestj-1]) and \\r |
| 435 | a[besti-1] == b[bestj-1]:\r |
| 436 | besti, bestj, bestsize = besti-1, bestj-1, bestsize+1\r |
| 437 | while besti+bestsize < ahi and bestj+bestsize < bhi and \\r |
| 438 | not isbjunk(b[bestj+bestsize]) and \\r |
| 439 | a[besti+bestsize] == b[bestj+bestsize]:\r |
| 440 | bestsize += 1\r |
| 441 | \r |
| 442 | # Now that we have a wholly interesting match (albeit possibly\r |
| 443 | # empty!), we may as well suck up the matching junk on each\r |
| 444 | # side of it too. Can't think of a good reason not to, and it\r |
| 445 | # saves post-processing the (possibly considerable) expense of\r |
| 446 | # figuring out what to do with it. In the case of an empty\r |
| 447 | # interesting match, this is clearly the right thing to do,\r |
| 448 | # because no other kind of match is possible in the regions.\r |
| 449 | while besti > alo and bestj > blo and \\r |
| 450 | isbjunk(b[bestj-1]) and \\r |
| 451 | a[besti-1] == b[bestj-1]:\r |
| 452 | besti, bestj, bestsize = besti-1, bestj-1, bestsize+1\r |
| 453 | while besti+bestsize < ahi and bestj+bestsize < bhi and \\r |
| 454 | isbjunk(b[bestj+bestsize]) and \\r |
| 455 | a[besti+bestsize] == b[bestj+bestsize]:\r |
| 456 | bestsize = bestsize + 1\r |
| 457 | \r |
| 458 | return Match(besti, bestj, bestsize)\r |
| 459 | \r |
| 460 | def get_matching_blocks(self):\r |
| 461 | """Return list of triples describing matching subsequences.\r |
| 462 | \r |
| 463 | Each triple is of the form (i, j, n), and means that\r |
| 464 | a[i:i+n] == b[j:j+n]. The triples are monotonically increasing in\r |
| 465 | i and in j. New in Python 2.5, it's also guaranteed that if\r |
| 466 | (i, j, n) and (i', j', n') are adjacent triples in the list, and\r |
| 467 | the second is not the last triple in the list, then i+n != i' or\r |
| 468 | j+n != j'. IOW, adjacent triples never describe adjacent equal\r |
| 469 | blocks.\r |
| 470 | \r |
| 471 | The last triple is a dummy, (len(a), len(b), 0), and is the only\r |
| 472 | triple with n==0.\r |
| 473 | \r |
| 474 | >>> s = SequenceMatcher(None, "abxcd", "abcd")\r |
| 475 | >>> s.get_matching_blocks()\r |
| 476 | [Match(a=0, b=0, size=2), Match(a=3, b=2, size=2), Match(a=5, b=4, size=0)]\r |
| 477 | """\r |
| 478 | \r |
| 479 | if self.matching_blocks is not None:\r |
| 480 | return self.matching_blocks\r |
| 481 | la, lb = len(self.a), len(self.b)\r |
| 482 | \r |
| 483 | # This is most naturally expressed as a recursive algorithm, but\r |
| 484 | # at least one user bumped into extreme use cases that exceeded\r |
| 485 | # the recursion limit on their box. So, now we maintain a list\r |
| 486 | # ('queue`) of blocks we still need to look at, and append partial\r |
| 487 | # results to `matching_blocks` in a loop; the matches are sorted\r |
| 488 | # at the end.\r |
| 489 | queue = [(0, la, 0, lb)]\r |
| 490 | matching_blocks = []\r |
| 491 | while queue:\r |
| 492 | alo, ahi, blo, bhi = queue.pop()\r |
| 493 | i, j, k = x = self.find_longest_match(alo, ahi, blo, bhi)\r |
| 494 | # a[alo:i] vs b[blo:j] unknown\r |
| 495 | # a[i:i+k] same as b[j:j+k]\r |
| 496 | # a[i+k:ahi] vs b[j+k:bhi] unknown\r |
| 497 | if k: # if k is 0, there was no matching block\r |
| 498 | matching_blocks.append(x)\r |
| 499 | if alo < i and blo < j:\r |
| 500 | queue.append((alo, i, blo, j))\r |
| 501 | if i+k < ahi and j+k < bhi:\r |
| 502 | queue.append((i+k, ahi, j+k, bhi))\r |
| 503 | matching_blocks.sort()\r |
| 504 | \r |
| 505 | # It's possible that we have adjacent equal blocks in the\r |
| 506 | # matching_blocks list now. Starting with 2.5, this code was added\r |
| 507 | # to collapse them.\r |
| 508 | i1 = j1 = k1 = 0\r |
| 509 | non_adjacent = []\r |
| 510 | for i2, j2, k2 in matching_blocks:\r |
| 511 | # Is this block adjacent to i1, j1, k1?\r |
| 512 | if i1 + k1 == i2 and j1 + k1 == j2:\r |
| 513 | # Yes, so collapse them -- this just increases the length of\r |
| 514 | # the first block by the length of the second, and the first\r |
| 515 | # block so lengthened remains the block to compare against.\r |
| 516 | k1 += k2\r |
| 517 | else:\r |
| 518 | # Not adjacent. Remember the first block (k1==0 means it's\r |
| 519 | # the dummy we started with), and make the second block the\r |
| 520 | # new block to compare against.\r |
| 521 | if k1:\r |
| 522 | non_adjacent.append((i1, j1, k1))\r |
| 523 | i1, j1, k1 = i2, j2, k2\r |
| 524 | if k1:\r |
| 525 | non_adjacent.append((i1, j1, k1))\r |
| 526 | \r |
| 527 | non_adjacent.append( (la, lb, 0) )\r |
| 528 | self.matching_blocks = non_adjacent\r |
| 529 | return map(Match._make, self.matching_blocks)\r |
| 530 | \r |
| 531 | def get_opcodes(self):\r |
| 532 | """Return list of 5-tuples describing how to turn a into b.\r |
| 533 | \r |
| 534 | Each tuple is of the form (tag, i1, i2, j1, j2). The first tuple\r |
| 535 | has i1 == j1 == 0, and remaining tuples have i1 == the i2 from the\r |
| 536 | tuple preceding it, and likewise for j1 == the previous j2.\r |
| 537 | \r |
| 538 | The tags are strings, with these meanings:\r |
| 539 | \r |
| 540 | 'replace': a[i1:i2] should be replaced by b[j1:j2]\r |
| 541 | 'delete': a[i1:i2] should be deleted.\r |
| 542 | Note that j1==j2 in this case.\r |
| 543 | 'insert': b[j1:j2] should be inserted at a[i1:i1].\r |
| 544 | Note that i1==i2 in this case.\r |
| 545 | 'equal': a[i1:i2] == b[j1:j2]\r |
| 546 | \r |
| 547 | >>> a = "qabxcd"\r |
| 548 | >>> b = "abycdf"\r |
| 549 | >>> s = SequenceMatcher(None, a, b)\r |
| 550 | >>> for tag, i1, i2, j1, j2 in s.get_opcodes():\r |
| 551 | ... print ("%7s a[%d:%d] (%s) b[%d:%d] (%s)" %\r |
| 552 | ... (tag, i1, i2, a[i1:i2], j1, j2, b[j1:j2]))\r |
| 553 | delete a[0:1] (q) b[0:0] ()\r |
| 554 | equal a[1:3] (ab) b[0:2] (ab)\r |
| 555 | replace a[3:4] (x) b[2:3] (y)\r |
| 556 | equal a[4:6] (cd) b[3:5] (cd)\r |
| 557 | insert a[6:6] () b[5:6] (f)\r |
| 558 | """\r |
| 559 | \r |
| 560 | if self.opcodes is not None:\r |
| 561 | return self.opcodes\r |
| 562 | i = j = 0\r |
| 563 | self.opcodes = answer = []\r |
| 564 | for ai, bj, size in self.get_matching_blocks():\r |
| 565 | # invariant: we've pumped out correct diffs to change\r |
| 566 | # a[:i] into b[:j], and the next matching block is\r |
| 567 | # a[ai:ai+size] == b[bj:bj+size]. So we need to pump\r |
| 568 | # out a diff to change a[i:ai] into b[j:bj], pump out\r |
| 569 | # the matching block, and move (i,j) beyond the match\r |
| 570 | tag = ''\r |
| 571 | if i < ai and j < bj:\r |
| 572 | tag = 'replace'\r |
| 573 | elif i < ai:\r |
| 574 | tag = 'delete'\r |
| 575 | elif j < bj:\r |
| 576 | tag = 'insert'\r |
| 577 | if tag:\r |
| 578 | answer.append( (tag, i, ai, j, bj) )\r |
| 579 | i, j = ai+size, bj+size\r |
| 580 | # the list of matching blocks is terminated by a\r |
| 581 | # sentinel with size 0\r |
| 582 | if size:\r |
| 583 | answer.append( ('equal', ai, i, bj, j) )\r |
| 584 | return answer\r |
| 585 | \r |
| 586 | def get_grouped_opcodes(self, n=3):\r |
| 587 | """ Isolate change clusters by eliminating ranges with no changes.\r |
| 588 | \r |
| 589 | Return a generator of groups with upto n lines of context.\r |
| 590 | Each group is in the same format as returned by get_opcodes().\r |
| 591 | \r |
| 592 | >>> from pprint import pprint\r |
| 593 | >>> a = map(str, range(1,40))\r |
| 594 | >>> b = a[:]\r |
| 595 | >>> b[8:8] = ['i'] # Make an insertion\r |
| 596 | >>> b[20] += 'x' # Make a replacement\r |
| 597 | >>> b[23:28] = [] # Make a deletion\r |
| 598 | >>> b[30] += 'y' # Make another replacement\r |
| 599 | >>> pprint(list(SequenceMatcher(None,a,b).get_grouped_opcodes()))\r |
| 600 | [[('equal', 5, 8, 5, 8), ('insert', 8, 8, 8, 9), ('equal', 8, 11, 9, 12)],\r |
| 601 | [('equal', 16, 19, 17, 20),\r |
| 602 | ('replace', 19, 20, 20, 21),\r |
| 603 | ('equal', 20, 22, 21, 23),\r |
| 604 | ('delete', 22, 27, 23, 23),\r |
| 605 | ('equal', 27, 30, 23, 26)],\r |
| 606 | [('equal', 31, 34, 27, 30),\r |
| 607 | ('replace', 34, 35, 30, 31),\r |
| 608 | ('equal', 35, 38, 31, 34)]]\r |
| 609 | """\r |
| 610 | \r |
| 611 | codes = self.get_opcodes()\r |
| 612 | if not codes:\r |
| 613 | codes = [("equal", 0, 1, 0, 1)]\r |
| 614 | # Fixup leading and trailing groups if they show no changes.\r |
| 615 | if codes[0][0] == 'equal':\r |
| 616 | tag, i1, i2, j1, j2 = codes[0]\r |
| 617 | codes[0] = tag, max(i1, i2-n), i2, max(j1, j2-n), j2\r |
| 618 | if codes[-1][0] == 'equal':\r |
| 619 | tag, i1, i2, j1, j2 = codes[-1]\r |
| 620 | codes[-1] = tag, i1, min(i2, i1+n), j1, min(j2, j1+n)\r |
| 621 | \r |
| 622 | nn = n + n\r |
| 623 | group = []\r |
| 624 | for tag, i1, i2, j1, j2 in codes:\r |
| 625 | # End the current group and start a new one whenever\r |
| 626 | # there is a large range with no changes.\r |
| 627 | if tag == 'equal' and i2-i1 > nn:\r |
| 628 | group.append((tag, i1, min(i2, i1+n), j1, min(j2, j1+n)))\r |
| 629 | yield group\r |
| 630 | group = []\r |
| 631 | i1, j1 = max(i1, i2-n), max(j1, j2-n)\r |
| 632 | group.append((tag, i1, i2, j1 ,j2))\r |
| 633 | if group and not (len(group)==1 and group[0][0] == 'equal'):\r |
| 634 | yield group\r |
| 635 | \r |
| 636 | def ratio(self):\r |
| 637 | """Return a measure of the sequences' similarity (float in [0,1]).\r |
| 638 | \r |
| 639 | Where T is the total number of elements in both sequences, and\r |
| 640 | M is the number of matches, this is 2.0*M / T.\r |
| 641 | Note that this is 1 if the sequences are identical, and 0 if\r |
| 642 | they have nothing in common.\r |
| 643 | \r |
| 644 | .ratio() is expensive to compute if you haven't already computed\r |
| 645 | .get_matching_blocks() or .get_opcodes(), in which case you may\r |
| 646 | want to try .quick_ratio() or .real_quick_ratio() first to get an\r |
| 647 | upper bound.\r |
| 648 | \r |
| 649 | >>> s = SequenceMatcher(None, "abcd", "bcde")\r |
| 650 | >>> s.ratio()\r |
| 651 | 0.75\r |
| 652 | >>> s.quick_ratio()\r |
| 653 | 0.75\r |
| 654 | >>> s.real_quick_ratio()\r |
| 655 | 1.0\r |
| 656 | """\r |
| 657 | \r |
| 658 | matches = reduce(lambda sum, triple: sum + triple[-1],\r |
| 659 | self.get_matching_blocks(), 0)\r |
| 660 | return _calculate_ratio(matches, len(self.a) + len(self.b))\r |
| 661 | \r |
| 662 | def quick_ratio(self):\r |
| 663 | """Return an upper bound on ratio() relatively quickly.\r |
| 664 | \r |
| 665 | This isn't defined beyond that it is an upper bound on .ratio(), and\r |
| 666 | is faster to compute.\r |
| 667 | """\r |
| 668 | \r |
| 669 | # viewing a and b as multisets, set matches to the cardinality\r |
| 670 | # of their intersection; this counts the number of matches\r |
| 671 | # without regard to order, so is clearly an upper bound\r |
| 672 | if self.fullbcount is None:\r |
| 673 | self.fullbcount = fullbcount = {}\r |
| 674 | for elt in self.b:\r |
| 675 | fullbcount[elt] = fullbcount.get(elt, 0) + 1\r |
| 676 | fullbcount = self.fullbcount\r |
| 677 | # avail[x] is the number of times x appears in 'b' less the\r |
| 678 | # number of times we've seen it in 'a' so far ... kinda\r |
| 679 | avail = {}\r |
| 680 | availhas, matches = avail.__contains__, 0\r |
| 681 | for elt in self.a:\r |
| 682 | if availhas(elt):\r |
| 683 | numb = avail[elt]\r |
| 684 | else:\r |
| 685 | numb = fullbcount.get(elt, 0)\r |
| 686 | avail[elt] = numb - 1\r |
| 687 | if numb > 0:\r |
| 688 | matches = matches + 1\r |
| 689 | return _calculate_ratio(matches, len(self.a) + len(self.b))\r |
| 690 | \r |
| 691 | def real_quick_ratio(self):\r |
| 692 | """Return an upper bound on ratio() very quickly.\r |
| 693 | \r |
| 694 | This isn't defined beyond that it is an upper bound on .ratio(), and\r |
| 695 | is faster to compute than either .ratio() or .quick_ratio().\r |
| 696 | """\r |
| 697 | \r |
| 698 | la, lb = len(self.a), len(self.b)\r |
| 699 | # can't have more matches than the number of elements in the\r |
| 700 | # shorter sequence\r |
| 701 | return _calculate_ratio(min(la, lb), la + lb)\r |
| 702 | \r |
| 703 | def get_close_matches(word, possibilities, n=3, cutoff=0.6):\r |
| 704 | """Use SequenceMatcher to return list of the best "good enough" matches.\r |
| 705 | \r |
| 706 | word is a sequence for which close matches are desired (typically a\r |
| 707 | string).\r |
| 708 | \r |
| 709 | possibilities is a list of sequences against which to match word\r |
| 710 | (typically a list of strings).\r |
| 711 | \r |
| 712 | Optional arg n (default 3) is the maximum number of close matches to\r |
| 713 | return. n must be > 0.\r |
| 714 | \r |
| 715 | Optional arg cutoff (default 0.6) is a float in [0, 1]. Possibilities\r |
| 716 | that don't score at least that similar to word are ignored.\r |
| 717 | \r |
| 718 | The best (no more than n) matches among the possibilities are returned\r |
| 719 | in a list, sorted by similarity score, most similar first.\r |
| 720 | \r |
| 721 | >>> get_close_matches("appel", ["ape", "apple", "peach", "puppy"])\r |
| 722 | ['apple', 'ape']\r |
| 723 | >>> import keyword as _keyword\r |
| 724 | >>> get_close_matches("wheel", _keyword.kwlist)\r |
| 725 | ['while']\r |
| 726 | >>> get_close_matches("apple", _keyword.kwlist)\r |
| 727 | []\r |
| 728 | >>> get_close_matches("accept", _keyword.kwlist)\r |
| 729 | ['except']\r |
| 730 | """\r |
| 731 | \r |
| 732 | if not n > 0:\r |
| 733 | raise ValueError("n must be > 0: %r" % (n,))\r |
| 734 | if not 0.0 <= cutoff <= 1.0:\r |
| 735 | raise ValueError("cutoff must be in [0.0, 1.0]: %r" % (cutoff,))\r |
| 736 | result = []\r |
| 737 | s = SequenceMatcher()\r |
| 738 | s.set_seq2(word)\r |
| 739 | for x in possibilities:\r |
| 740 | s.set_seq1(x)\r |
| 741 | if s.real_quick_ratio() >= cutoff and \\r |
| 742 | s.quick_ratio() >= cutoff and \\r |
| 743 | s.ratio() >= cutoff:\r |
| 744 | result.append((s.ratio(), x))\r |
| 745 | \r |
| 746 | # Move the best scorers to head of list\r |
| 747 | result = heapq.nlargest(n, result)\r |
| 748 | # Strip scores for the best n matches\r |
| 749 | return [x for score, x in result]\r |
| 750 | \r |
| 751 | def _count_leading(line, ch):\r |
| 752 | """\r |
| 753 | Return number of `ch` characters at the start of `line`.\r |
| 754 | \r |
| 755 | Example:\r |
| 756 | \r |
| 757 | >>> _count_leading(' abc', ' ')\r |
| 758 | 3\r |
| 759 | """\r |
| 760 | \r |
| 761 | i, n = 0, len(line)\r |
| 762 | while i < n and line[i] == ch:\r |
| 763 | i += 1\r |
| 764 | return i\r |
| 765 | \r |
| 766 | class Differ:\r |
| 767 | r"""\r |
| 768 | Differ is a class for comparing sequences of lines of text, and\r |
| 769 | producing human-readable differences or deltas. Differ uses\r |
| 770 | SequenceMatcher both to compare sequences of lines, and to compare\r |
| 771 | sequences of characters within similar (near-matching) lines.\r |
| 772 | \r |
| 773 | Each line of a Differ delta begins with a two-letter code:\r |
| 774 | \r |
| 775 | '- ' line unique to sequence 1\r |
| 776 | '+ ' line unique to sequence 2\r |
| 777 | ' ' line common to both sequences\r |
| 778 | '? ' line not present in either input sequence\r |
| 779 | \r |
| 780 | Lines beginning with '? ' attempt to guide the eye to intraline\r |
| 781 | differences, and were not present in either input sequence. These lines\r |
| 782 | can be confusing if the sequences contain tab characters.\r |
| 783 | \r |
| 784 | Note that Differ makes no claim to produce a *minimal* diff. To the\r |
| 785 | contrary, minimal diffs are often counter-intuitive, because they synch\r |
| 786 | up anywhere possible, sometimes accidental matches 100 pages apart.\r |
| 787 | Restricting synch points to contiguous matches preserves some notion of\r |
| 788 | locality, at the occasional cost of producing a longer diff.\r |
| 789 | \r |
| 790 | Example: Comparing two texts.\r |
| 791 | \r |
| 792 | First we set up the texts, sequences of individual single-line strings\r |
| 793 | ending with newlines (such sequences can also be obtained from the\r |
| 794 | `readlines()` method of file-like objects):\r |
| 795 | \r |
| 796 | >>> text1 = ''' 1. Beautiful is better than ugly.\r |
| 797 | ... 2. Explicit is better than implicit.\r |
| 798 | ... 3. Simple is better than complex.\r |
| 799 | ... 4. Complex is better than complicated.\r |
| 800 | ... '''.splitlines(1)\r |
| 801 | >>> len(text1)\r |
| 802 | 4\r |
| 803 | >>> text1[0][-1]\r |
| 804 | '\n'\r |
| 805 | >>> text2 = ''' 1. Beautiful is better than ugly.\r |
| 806 | ... 3. Simple is better than complex.\r |
| 807 | ... 4. Complicated is better than complex.\r |
| 808 | ... 5. Flat is better than nested.\r |
| 809 | ... '''.splitlines(1)\r |
| 810 | \r |
| 811 | Next we instantiate a Differ object:\r |
| 812 | \r |
| 813 | >>> d = Differ()\r |
| 814 | \r |
| 815 | Note that when instantiating a Differ object we may pass functions to\r |
| 816 | filter out line and character 'junk'. See Differ.__init__ for details.\r |
| 817 | \r |
| 818 | Finally, we compare the two:\r |
| 819 | \r |
| 820 | >>> result = list(d.compare(text1, text2))\r |
| 821 | \r |
| 822 | 'result' is a list of strings, so let's pretty-print it:\r |
| 823 | \r |
| 824 | >>> from pprint import pprint as _pprint\r |
| 825 | >>> _pprint(result)\r |
| 826 | [' 1. Beautiful is better than ugly.\n',\r |
| 827 | '- 2. Explicit is better than implicit.\n',\r |
| 828 | '- 3. Simple is better than complex.\n',\r |
| 829 | '+ 3. Simple is better than complex.\n',\r |
| 830 | '? ++\n',\r |
| 831 | '- 4. Complex is better than complicated.\n',\r |
| 832 | '? ^ ---- ^\n',\r |
| 833 | '+ 4. Complicated is better than complex.\n',\r |
| 834 | '? ++++ ^ ^\n',\r |
| 835 | '+ 5. Flat is better than nested.\n']\r |
| 836 | \r |
| 837 | As a single multi-line string it looks like this:\r |
| 838 | \r |
| 839 | >>> print ''.join(result),\r |
| 840 | 1. Beautiful is better than ugly.\r |
| 841 | - 2. Explicit is better than implicit.\r |
| 842 | - 3. Simple is better than complex.\r |
| 843 | + 3. Simple is better than complex.\r |
| 844 | ? ++\r |
| 845 | - 4. Complex is better than complicated.\r |
| 846 | ? ^ ---- ^\r |
| 847 | + 4. Complicated is better than complex.\r |
| 848 | ? ++++ ^ ^\r |
| 849 | + 5. Flat is better than nested.\r |
| 850 | \r |
| 851 | Methods:\r |
| 852 | \r |
| 853 | __init__(linejunk=None, charjunk=None)\r |
| 854 | Construct a text differencer, with optional filters.\r |
| 855 | \r |
| 856 | compare(a, b)\r |
| 857 | Compare two sequences of lines; generate the resulting delta.\r |
| 858 | """\r |
| 859 | \r |
| 860 | def __init__(self, linejunk=None, charjunk=None):\r |
| 861 | """\r |
| 862 | Construct a text differencer, with optional filters.\r |
| 863 | \r |
| 864 | The two optional keyword parameters are for filter functions:\r |
| 865 | \r |
| 866 | - `linejunk`: A function that should accept a single string argument,\r |
| 867 | and return true iff the string is junk. The module-level function\r |
| 868 | `IS_LINE_JUNK` may be used to filter out lines without visible\r |
| 869 | characters, except for at most one splat ('#'). It is recommended\r |
| 870 | to leave linejunk None; as of Python 2.3, the underlying\r |
| 871 | SequenceMatcher class has grown an adaptive notion of "noise" lines\r |
| 872 | that's better than any static definition the author has ever been\r |
| 873 | able to craft.\r |
| 874 | \r |
| 875 | - `charjunk`: A function that should accept a string of length 1. The\r |
| 876 | module-level function `IS_CHARACTER_JUNK` may be used to filter out\r |
| 877 | whitespace characters (a blank or tab; **note**: bad idea to include\r |
| 878 | newline in this!). Use of IS_CHARACTER_JUNK is recommended.\r |
| 879 | """\r |
| 880 | \r |
| 881 | self.linejunk = linejunk\r |
| 882 | self.charjunk = charjunk\r |
| 883 | \r |
| 884 | def compare(self, a, b):\r |
| 885 | r"""\r |
| 886 | Compare two sequences of lines; generate the resulting delta.\r |
| 887 | \r |
| 888 | Each sequence must contain individual single-line strings ending with\r |
| 889 | newlines. Such sequences can be obtained from the `readlines()` method\r |
| 890 | of file-like objects. The delta generated also consists of newline-\r |
| 891 | terminated strings, ready to be printed as-is via the writeline()\r |
| 892 | method of a file-like object.\r |
| 893 | \r |
| 894 | Example:\r |
| 895 | \r |
| 896 | >>> print ''.join(Differ().compare('one\ntwo\nthree\n'.splitlines(1),\r |
| 897 | ... 'ore\ntree\nemu\n'.splitlines(1))),\r |
| 898 | - one\r |
| 899 | ? ^\r |
| 900 | + ore\r |
| 901 | ? ^\r |
| 902 | - two\r |
| 903 | - three\r |
| 904 | ? -\r |
| 905 | + tree\r |
| 906 | + emu\r |
| 907 | """\r |
| 908 | \r |
| 909 | cruncher = SequenceMatcher(self.linejunk, a, b)\r |
| 910 | for tag, alo, ahi, blo, bhi in cruncher.get_opcodes():\r |
| 911 | if tag == 'replace':\r |
| 912 | g = self._fancy_replace(a, alo, ahi, b, blo, bhi)\r |
| 913 | elif tag == 'delete':\r |
| 914 | g = self._dump('-', a, alo, ahi)\r |
| 915 | elif tag == 'insert':\r |
| 916 | g = self._dump('+', b, blo, bhi)\r |
| 917 | elif tag == 'equal':\r |
| 918 | g = self._dump(' ', a, alo, ahi)\r |
| 919 | else:\r |
| 920 | raise ValueError, 'unknown tag %r' % (tag,)\r |
| 921 | \r |
| 922 | for line in g:\r |
| 923 | yield line\r |
| 924 | \r |
| 925 | def _dump(self, tag, x, lo, hi):\r |
| 926 | """Generate comparison results for a same-tagged range."""\r |
| 927 | for i in xrange(lo, hi):\r |
| 928 | yield '%s %s' % (tag, x[i])\r |
| 929 | \r |
| 930 | def _plain_replace(self, a, alo, ahi, b, blo, bhi):\r |
| 931 | assert alo < ahi and blo < bhi\r |
| 932 | # dump the shorter block first -- reduces the burden on short-term\r |
| 933 | # memory if the blocks are of very different sizes\r |
| 934 | if bhi - blo < ahi - alo:\r |
| 935 | first = self._dump('+', b, blo, bhi)\r |
| 936 | second = self._dump('-', a, alo, ahi)\r |
| 937 | else:\r |
| 938 | first = self._dump('-', a, alo, ahi)\r |
| 939 | second = self._dump('+', b, blo, bhi)\r |
| 940 | \r |
| 941 | for g in first, second:\r |
| 942 | for line in g:\r |
| 943 | yield line\r |
| 944 | \r |
| 945 | def _fancy_replace(self, a, alo, ahi, b, blo, bhi):\r |
| 946 | r"""\r |
| 947 | When replacing one block of lines with another, search the blocks\r |
| 948 | for *similar* lines; the best-matching pair (if any) is used as a\r |
| 949 | synch point, and intraline difference marking is done on the\r |
| 950 | similar pair. Lots of work, but often worth it.\r |
| 951 | \r |
| 952 | Example:\r |
| 953 | \r |
| 954 | >>> d = Differ()\r |
| 955 | >>> results = d._fancy_replace(['abcDefghiJkl\n'], 0, 1,\r |
| 956 | ... ['abcdefGhijkl\n'], 0, 1)\r |
| 957 | >>> print ''.join(results),\r |
| 958 | - abcDefghiJkl\r |
| 959 | ? ^ ^ ^\r |
| 960 | + abcdefGhijkl\r |
| 961 | ? ^ ^ ^\r |
| 962 | """\r |
| 963 | \r |
| 964 | # don't synch up unless the lines have a similarity score of at\r |
| 965 | # least cutoff; best_ratio tracks the best score seen so far\r |
| 966 | best_ratio, cutoff = 0.74, 0.75\r |
| 967 | cruncher = SequenceMatcher(self.charjunk)\r |
| 968 | eqi, eqj = None, None # 1st indices of equal lines (if any)\r |
| 969 | \r |
| 970 | # search for the pair that matches best without being identical\r |
| 971 | # (identical lines must be junk lines, & we don't want to synch up\r |
| 972 | # on junk -- unless we have to)\r |
| 973 | for j in xrange(blo, bhi):\r |
| 974 | bj = b[j]\r |
| 975 | cruncher.set_seq2(bj)\r |
| 976 | for i in xrange(alo, ahi):\r |
| 977 | ai = a[i]\r |
| 978 | if ai == bj:\r |
| 979 | if eqi is None:\r |
| 980 | eqi, eqj = i, j\r |
| 981 | continue\r |
| 982 | cruncher.set_seq1(ai)\r |
| 983 | # computing similarity is expensive, so use the quick\r |
| 984 | # upper bounds first -- have seen this speed up messy\r |
| 985 | # compares by a factor of 3.\r |
| 986 | # note that ratio() is only expensive to compute the first\r |
| 987 | # time it's called on a sequence pair; the expensive part\r |
| 988 | # of the computation is cached by cruncher\r |
| 989 | if cruncher.real_quick_ratio() > best_ratio and \\r |
| 990 | cruncher.quick_ratio() > best_ratio and \\r |
| 991 | cruncher.ratio() > best_ratio:\r |
| 992 | best_ratio, best_i, best_j = cruncher.ratio(), i, j\r |
| 993 | if best_ratio < cutoff:\r |
| 994 | # no non-identical "pretty close" pair\r |
| 995 | if eqi is None:\r |
| 996 | # no identical pair either -- treat it as a straight replace\r |
| 997 | for line in self._plain_replace(a, alo, ahi, b, blo, bhi):\r |
| 998 | yield line\r |
| 999 | return\r |
| 1000 | # no close pair, but an identical pair -- synch up on that\r |
| 1001 | best_i, best_j, best_ratio = eqi, eqj, 1.0\r |
| 1002 | else:\r |
| 1003 | # there's a close pair, so forget the identical pair (if any)\r |
| 1004 | eqi = None\r |
| 1005 | \r |
| 1006 | # a[best_i] very similar to b[best_j]; eqi is None iff they're not\r |
| 1007 | # identical\r |
| 1008 | \r |
| 1009 | # pump out diffs from before the synch point\r |
| 1010 | for line in self._fancy_helper(a, alo, best_i, b, blo, best_j):\r |
| 1011 | yield line\r |
| 1012 | \r |
| 1013 | # do intraline marking on the synch pair\r |
| 1014 | aelt, belt = a[best_i], b[best_j]\r |
| 1015 | if eqi is None:\r |
| 1016 | # pump out a '-', '?', '+', '?' quad for the synched lines\r |
| 1017 | atags = btags = ""\r |
| 1018 | cruncher.set_seqs(aelt, belt)\r |
| 1019 | for tag, ai1, ai2, bj1, bj2 in cruncher.get_opcodes():\r |
| 1020 | la, lb = ai2 - ai1, bj2 - bj1\r |
| 1021 | if tag == 'replace':\r |
| 1022 | atags += '^' * la\r |
| 1023 | btags += '^' * lb\r |
| 1024 | elif tag == 'delete':\r |
| 1025 | atags += '-' * la\r |
| 1026 | elif tag == 'insert':\r |
| 1027 | btags += '+' * lb\r |
| 1028 | elif tag == 'equal':\r |
| 1029 | atags += ' ' * la\r |
| 1030 | btags += ' ' * lb\r |
| 1031 | else:\r |
| 1032 | raise ValueError, 'unknown tag %r' % (tag,)\r |
| 1033 | for line in self._qformat(aelt, belt, atags, btags):\r |
| 1034 | yield line\r |
| 1035 | else:\r |
| 1036 | # the synch pair is identical\r |
| 1037 | yield ' ' + aelt\r |
| 1038 | \r |
| 1039 | # pump out diffs from after the synch point\r |
| 1040 | for line in self._fancy_helper(a, best_i+1, ahi, b, best_j+1, bhi):\r |
| 1041 | yield line\r |
| 1042 | \r |
| 1043 | def _fancy_helper(self, a, alo, ahi, b, blo, bhi):\r |
| 1044 | g = []\r |
| 1045 | if alo < ahi:\r |
| 1046 | if blo < bhi:\r |
| 1047 | g = self._fancy_replace(a, alo, ahi, b, blo, bhi)\r |
| 1048 | else:\r |
| 1049 | g = self._dump('-', a, alo, ahi)\r |
| 1050 | elif blo < bhi:\r |
| 1051 | g = self._dump('+', b, blo, bhi)\r |
| 1052 | \r |
| 1053 | for line in g:\r |
| 1054 | yield line\r |
| 1055 | \r |
| 1056 | def _qformat(self, aline, bline, atags, btags):\r |
| 1057 | r"""\r |
| 1058 | Format "?" output and deal with leading tabs.\r |
| 1059 | \r |
| 1060 | Example:\r |
| 1061 | \r |
| 1062 | >>> d = Differ()\r |
| 1063 | >>> results = d._qformat('\tabcDefghiJkl\n', '\tabcdefGhijkl\n',\r |
| 1064 | ... ' ^ ^ ^ ', ' ^ ^ ^ ')\r |
| 1065 | >>> for line in results: print repr(line)\r |
| 1066 | ...\r |
| 1067 | '- \tabcDefghiJkl\n'\r |
| 1068 | '? \t ^ ^ ^\n'\r |
| 1069 | '+ \tabcdefGhijkl\n'\r |
| 1070 | '? \t ^ ^ ^\n'\r |
| 1071 | """\r |
| 1072 | \r |
| 1073 | # Can hurt, but will probably help most of the time.\r |
| 1074 | common = min(_count_leading(aline, "\t"),\r |
| 1075 | _count_leading(bline, "\t"))\r |
| 1076 | common = min(common, _count_leading(atags[:common], " "))\r |
| 1077 | common = min(common, _count_leading(btags[:common], " "))\r |
| 1078 | atags = atags[common:].rstrip()\r |
| 1079 | btags = btags[common:].rstrip()\r |
| 1080 | \r |
| 1081 | yield "- " + aline\r |
| 1082 | if atags:\r |
| 1083 | yield "? %s%s\n" % ("\t" * common, atags)\r |
| 1084 | \r |
| 1085 | yield "+ " + bline\r |
| 1086 | if btags:\r |
| 1087 | yield "? %s%s\n" % ("\t" * common, btags)\r |
| 1088 | \r |
| 1089 | # With respect to junk, an earlier version of ndiff simply refused to\r |
| 1090 | # *start* a match with a junk element. The result was cases like this:\r |
| 1091 | # before: private Thread currentThread;\r |
| 1092 | # after: private volatile Thread currentThread;\r |
| 1093 | # If you consider whitespace to be junk, the longest contiguous match\r |
| 1094 | # not starting with junk is "e Thread currentThread". So ndiff reported\r |
| 1095 | # that "e volatil" was inserted between the 't' and the 'e' in "private".\r |
| 1096 | # While an accurate view, to people that's absurd. The current version\r |
| 1097 | # looks for matching blocks that are entirely junk-free, then extends the\r |
| 1098 | # longest one of those as far as possible but only with matching junk.\r |
| 1099 | # So now "currentThread" is matched, then extended to suck up the\r |
| 1100 | # preceding blank; then "private" is matched, and extended to suck up the\r |
| 1101 | # following blank; then "Thread" is matched; and finally ndiff reports\r |
| 1102 | # that "volatile " was inserted before "Thread". The only quibble\r |
| 1103 | # remaining is that perhaps it was really the case that " volatile"\r |
| 1104 | # was inserted after "private". I can live with that <wink>.\r |
| 1105 | \r |
| 1106 | import re\r |
| 1107 | \r |
| 1108 | def IS_LINE_JUNK(line, pat=re.compile(r"\s*#?\s*$").match):\r |
| 1109 | r"""\r |
| 1110 | Return 1 for ignorable line: iff `line` is blank or contains a single '#'.\r |
| 1111 | \r |
| 1112 | Examples:\r |
| 1113 | \r |
| 1114 | >>> IS_LINE_JUNK('\n')\r |
| 1115 | True\r |
| 1116 | >>> IS_LINE_JUNK(' # \n')\r |
| 1117 | True\r |
| 1118 | >>> IS_LINE_JUNK('hello\n')\r |
| 1119 | False\r |
| 1120 | """\r |
| 1121 | \r |
| 1122 | return pat(line) is not None\r |
| 1123 | \r |
| 1124 | def IS_CHARACTER_JUNK(ch, ws=" \t"):\r |
| 1125 | r"""\r |
| 1126 | Return 1 for ignorable character: iff `ch` is a space or tab.\r |
| 1127 | \r |
| 1128 | Examples:\r |
| 1129 | \r |
| 1130 | >>> IS_CHARACTER_JUNK(' ')\r |
| 1131 | True\r |
| 1132 | >>> IS_CHARACTER_JUNK('\t')\r |
| 1133 | True\r |
| 1134 | >>> IS_CHARACTER_JUNK('\n')\r |
| 1135 | False\r |
| 1136 | >>> IS_CHARACTER_JUNK('x')\r |
| 1137 | False\r |
| 1138 | """\r |
| 1139 | \r |
| 1140 | return ch in ws\r |
| 1141 | \r |
| 1142 | \r |
| 1143 | ########################################################################\r |
| 1144 | ### Unified Diff\r |
| 1145 | ########################################################################\r |
| 1146 | \r |
| 1147 | def _format_range_unified(start, stop):\r |
| 1148 | 'Convert range to the "ed" format'\r |
| 1149 | # Per the diff spec at http://www.unix.org/single_unix_specification/\r |
| 1150 | beginning = start + 1 # lines start numbering with one\r |
| 1151 | length = stop - start\r |
| 1152 | if length == 1:\r |
| 1153 | return '{}'.format(beginning)\r |
| 1154 | if not length:\r |
| 1155 | beginning -= 1 # empty ranges begin at line just before the range\r |
| 1156 | return '{},{}'.format(beginning, length)\r |
| 1157 | \r |
| 1158 | def unified_diff(a, b, fromfile='', tofile='', fromfiledate='',\r |
| 1159 | tofiledate='', n=3, lineterm='\n'):\r |
| 1160 | r"""\r |
| 1161 | Compare two sequences of lines; generate the delta as a unified diff.\r |
| 1162 | \r |
| 1163 | Unified diffs are a compact way of showing line changes and a few\r |
| 1164 | lines of context. The number of context lines is set by 'n' which\r |
| 1165 | defaults to three.\r |
| 1166 | \r |
| 1167 | By default, the diff control lines (those with ---, +++, or @@) are\r |
| 1168 | created with a trailing newline. This is helpful so that inputs\r |
| 1169 | created from file.readlines() result in diffs that are suitable for\r |
| 1170 | file.writelines() since both the inputs and outputs have trailing\r |
| 1171 | newlines.\r |
| 1172 | \r |
| 1173 | For inputs that do not have trailing newlines, set the lineterm\r |
| 1174 | argument to "" so that the output will be uniformly newline free.\r |
| 1175 | \r |
| 1176 | The unidiff format normally has a header for filenames and modification\r |
| 1177 | times. Any or all of these may be specified using strings for\r |
| 1178 | 'fromfile', 'tofile', 'fromfiledate', and 'tofiledate'.\r |
| 1179 | The modification times are normally expressed in the ISO 8601 format.\r |
| 1180 | \r |
| 1181 | Example:\r |
| 1182 | \r |
| 1183 | >>> for line in unified_diff('one two three four'.split(),\r |
| 1184 | ... 'zero one tree four'.split(), 'Original', 'Current',\r |
| 1185 | ... '2005-01-26 23:30:50', '2010-04-02 10:20:52',\r |
| 1186 | ... lineterm=''):\r |
| 1187 | ... print line # doctest: +NORMALIZE_WHITESPACE\r |
| 1188 | --- Original 2005-01-26 23:30:50\r |
| 1189 | +++ Current 2010-04-02 10:20:52\r |
| 1190 | @@ -1,4 +1,4 @@\r |
| 1191 | +zero\r |
| 1192 | one\r |
| 1193 | -two\r |
| 1194 | -three\r |
| 1195 | +tree\r |
| 1196 | four\r |
| 1197 | """\r |
| 1198 | \r |
| 1199 | started = False\r |
| 1200 | for group in SequenceMatcher(None,a,b).get_grouped_opcodes(n):\r |
| 1201 | if not started:\r |
| 1202 | started = True\r |
| 1203 | fromdate = '\t{}'.format(fromfiledate) if fromfiledate else ''\r |
| 1204 | todate = '\t{}'.format(tofiledate) if tofiledate else ''\r |
| 1205 | yield '--- {}{}{}'.format(fromfile, fromdate, lineterm)\r |
| 1206 | yield '+++ {}{}{}'.format(tofile, todate, lineterm)\r |
| 1207 | \r |
| 1208 | first, last = group[0], group[-1]\r |
| 1209 | file1_range = _format_range_unified(first[1], last[2])\r |
| 1210 | file2_range = _format_range_unified(first[3], last[4])\r |
| 1211 | yield '@@ -{} +{} @@{}'.format(file1_range, file2_range, lineterm)\r |
| 1212 | \r |
| 1213 | for tag, i1, i2, j1, j2 in group:\r |
| 1214 | if tag == 'equal':\r |
| 1215 | for line in a[i1:i2]:\r |
| 1216 | yield ' ' + line\r |
| 1217 | continue\r |
| 1218 | if tag in ('replace', 'delete'):\r |
| 1219 | for line in a[i1:i2]:\r |
| 1220 | yield '-' + line\r |
| 1221 | if tag in ('replace', 'insert'):\r |
| 1222 | for line in b[j1:j2]:\r |
| 1223 | yield '+' + line\r |
| 1224 | \r |
| 1225 | \r |
| 1226 | ########################################################################\r |
| 1227 | ### Context Diff\r |
| 1228 | ########################################################################\r |
| 1229 | \r |
| 1230 | def _format_range_context(start, stop):\r |
| 1231 | 'Convert range to the "ed" format'\r |
| 1232 | # Per the diff spec at http://www.unix.org/single_unix_specification/\r |
| 1233 | beginning = start + 1 # lines start numbering with one\r |
| 1234 | length = stop - start\r |
| 1235 | if not length:\r |
| 1236 | beginning -= 1 # empty ranges begin at line just before the range\r |
| 1237 | if length <= 1:\r |
| 1238 | return '{}'.format(beginning)\r |
| 1239 | return '{},{}'.format(beginning, beginning + length - 1)\r |
| 1240 | \r |
| 1241 | # See http://www.unix.org/single_unix_specification/\r |
| 1242 | def context_diff(a, b, fromfile='', tofile='',\r |
| 1243 | fromfiledate='', tofiledate='', n=3, lineterm='\n'):\r |
| 1244 | r"""\r |
| 1245 | Compare two sequences of lines; generate the delta as a context diff.\r |
| 1246 | \r |
| 1247 | Context diffs are a compact way of showing line changes and a few\r |
| 1248 | lines of context. The number of context lines is set by 'n' which\r |
| 1249 | defaults to three.\r |
| 1250 | \r |
| 1251 | By default, the diff control lines (those with *** or ---) are\r |
| 1252 | created with a trailing newline. This is helpful so that inputs\r |
| 1253 | created from file.readlines() result in diffs that are suitable for\r |
| 1254 | file.writelines() since both the inputs and outputs have trailing\r |
| 1255 | newlines.\r |
| 1256 | \r |
| 1257 | For inputs that do not have trailing newlines, set the lineterm\r |
| 1258 | argument to "" so that the output will be uniformly newline free.\r |
| 1259 | \r |
| 1260 | The context diff format normally has a header for filenames and\r |
| 1261 | modification times. Any or all of these may be specified using\r |
| 1262 | strings for 'fromfile', 'tofile', 'fromfiledate', and 'tofiledate'.\r |
| 1263 | The modification times are normally expressed in the ISO 8601 format.\r |
| 1264 | If not specified, the strings default to blanks.\r |
| 1265 | \r |
| 1266 | Example:\r |
| 1267 | \r |
| 1268 | >>> print ''.join(context_diff('one\ntwo\nthree\nfour\n'.splitlines(1),\r |
| 1269 | ... 'zero\none\ntree\nfour\n'.splitlines(1), 'Original', 'Current')),\r |
| 1270 | *** Original\r |
| 1271 | --- Current\r |
| 1272 | ***************\r |
| 1273 | *** 1,4 ****\r |
| 1274 | one\r |
| 1275 | ! two\r |
| 1276 | ! three\r |
| 1277 | four\r |
| 1278 | --- 1,4 ----\r |
| 1279 | + zero\r |
| 1280 | one\r |
| 1281 | ! tree\r |
| 1282 | four\r |
| 1283 | """\r |
| 1284 | \r |
| 1285 | prefix = dict(insert='+ ', delete='- ', replace='! ', equal=' ')\r |
| 1286 | started = False\r |
| 1287 | for group in SequenceMatcher(None,a,b).get_grouped_opcodes(n):\r |
| 1288 | if not started:\r |
| 1289 | started = True\r |
| 1290 | fromdate = '\t{}'.format(fromfiledate) if fromfiledate else ''\r |
| 1291 | todate = '\t{}'.format(tofiledate) if tofiledate else ''\r |
| 1292 | yield '*** {}{}{}'.format(fromfile, fromdate, lineterm)\r |
| 1293 | yield '--- {}{}{}'.format(tofile, todate, lineterm)\r |
| 1294 | \r |
| 1295 | first, last = group[0], group[-1]\r |
| 1296 | yield '***************' + lineterm\r |
| 1297 | \r |
| 1298 | file1_range = _format_range_context(first[1], last[2])\r |
| 1299 | yield '*** {} ****{}'.format(file1_range, lineterm)\r |
| 1300 | \r |
| 1301 | if any(tag in ('replace', 'delete') for tag, _, _, _, _ in group):\r |
| 1302 | for tag, i1, i2, _, _ in group:\r |
| 1303 | if tag != 'insert':\r |
| 1304 | for line in a[i1:i2]:\r |
| 1305 | yield prefix[tag] + line\r |
| 1306 | \r |
| 1307 | file2_range = _format_range_context(first[3], last[4])\r |
| 1308 | yield '--- {} ----{}'.format(file2_range, lineterm)\r |
| 1309 | \r |
| 1310 | if any(tag in ('replace', 'insert') for tag, _, _, _, _ in group):\r |
| 1311 | for tag, _, _, j1, j2 in group:\r |
| 1312 | if tag != 'delete':\r |
| 1313 | for line in b[j1:j2]:\r |
| 1314 | yield prefix[tag] + line\r |
| 1315 | \r |
| 1316 | def ndiff(a, b, linejunk=None, charjunk=IS_CHARACTER_JUNK):\r |
| 1317 | r"""\r |
| 1318 | Compare `a` and `b` (lists of strings); return a `Differ`-style delta.\r |
| 1319 | \r |
| 1320 | Optional keyword parameters `linejunk` and `charjunk` are for filter\r |
| 1321 | functions (or None):\r |
| 1322 | \r |
| 1323 | - linejunk: A function that should accept a single string argument, and\r |
| 1324 | return true iff the string is junk. The default is None, and is\r |
| 1325 | recommended; as of Python 2.3, an adaptive notion of "noise" lines is\r |
| 1326 | used that does a good job on its own.\r |
| 1327 | \r |
| 1328 | - charjunk: A function that should accept a string of length 1. The\r |
| 1329 | default is module-level function IS_CHARACTER_JUNK, which filters out\r |
| 1330 | whitespace characters (a blank or tab; note: bad idea to include newline\r |
| 1331 | in this!).\r |
| 1332 | \r |
| 1333 | Tools/scripts/ndiff.py is a command-line front-end to this function.\r |
| 1334 | \r |
| 1335 | Example:\r |
| 1336 | \r |
| 1337 | >>> diff = ndiff('one\ntwo\nthree\n'.splitlines(1),\r |
| 1338 | ... 'ore\ntree\nemu\n'.splitlines(1))\r |
| 1339 | >>> print ''.join(diff),\r |
| 1340 | - one\r |
| 1341 | ? ^\r |
| 1342 | + ore\r |
| 1343 | ? ^\r |
| 1344 | - two\r |
| 1345 | - three\r |
| 1346 | ? -\r |
| 1347 | + tree\r |
| 1348 | + emu\r |
| 1349 | """\r |
| 1350 | return Differ(linejunk, charjunk).compare(a, b)\r |
| 1351 | \r |
| 1352 | def _mdiff(fromlines, tolines, context=None, linejunk=None,\r |
| 1353 | charjunk=IS_CHARACTER_JUNK):\r |
| 1354 | r"""Returns generator yielding marked up from/to side by side differences.\r |
| 1355 | \r |
| 1356 | Arguments:\r |
| 1357 | fromlines -- list of text lines to compared to tolines\r |
| 1358 | tolines -- list of text lines to be compared to fromlines\r |
| 1359 | context -- number of context lines to display on each side of difference,\r |
| 1360 | if None, all from/to text lines will be generated.\r |
| 1361 | linejunk -- passed on to ndiff (see ndiff documentation)\r |
| 1362 | charjunk -- passed on to ndiff (see ndiff documentation)\r |
| 1363 | \r |
| 1364 | This function returns an interator which returns a tuple:\r |
| 1365 | (from line tuple, to line tuple, boolean flag)\r |
| 1366 | \r |
| 1367 | from/to line tuple -- (line num, line text)\r |
| 1368 | line num -- integer or None (to indicate a context separation)\r |
| 1369 | line text -- original line text with following markers inserted:\r |
| 1370 | '\0+' -- marks start of added text\r |
| 1371 | '\0-' -- marks start of deleted text\r |
| 1372 | '\0^' -- marks start of changed text\r |
| 1373 | '\1' -- marks end of added/deleted/changed text\r |
| 1374 | \r |
| 1375 | boolean flag -- None indicates context separation, True indicates\r |
| 1376 | either "from" or "to" line contains a change, otherwise False.\r |
| 1377 | \r |
| 1378 | This function/iterator was originally developed to generate side by side\r |
| 1379 | file difference for making HTML pages (see HtmlDiff class for example\r |
| 1380 | usage).\r |
| 1381 | \r |
| 1382 | Note, this function utilizes the ndiff function to generate the side by\r |
| 1383 | side difference markup. Optional ndiff arguments may be passed to this\r |
| 1384 | function and they in turn will be passed to ndiff.\r |
| 1385 | """\r |
| 1386 | import re\r |
| 1387 | \r |
| 1388 | # regular expression for finding intraline change indices\r |
| 1389 | change_re = re.compile('(\++|\-+|\^+)')\r |
| 1390 | \r |
| 1391 | # create the difference iterator to generate the differences\r |
| 1392 | diff_lines_iterator = ndiff(fromlines,tolines,linejunk,charjunk)\r |
| 1393 | \r |
| 1394 | def _make_line(lines, format_key, side, num_lines=[0,0]):\r |
| 1395 | """Returns line of text with user's change markup and line formatting.\r |
| 1396 | \r |
| 1397 | lines -- list of lines from the ndiff generator to produce a line of\r |
| 1398 | text from. When producing the line of text to return, the\r |
| 1399 | lines used are removed from this list.\r |
| 1400 | format_key -- '+' return first line in list with "add" markup around\r |
| 1401 | the entire line.\r |
| 1402 | '-' return first line in list with "delete" markup around\r |
| 1403 | the entire line.\r |
| 1404 | '?' return first line in list with add/delete/change\r |
| 1405 | intraline markup (indices obtained from second line)\r |
| 1406 | None return first line in list with no markup\r |
| 1407 | side -- indice into the num_lines list (0=from,1=to)\r |
| 1408 | num_lines -- from/to current line number. This is NOT intended to be a\r |
| 1409 | passed parameter. It is present as a keyword argument to\r |
| 1410 | maintain memory of the current line numbers between calls\r |
| 1411 | of this function.\r |
| 1412 | \r |
| 1413 | Note, this function is purposefully not defined at the module scope so\r |
| 1414 | that data it needs from its parent function (within whose context it\r |
| 1415 | is defined) does not need to be of module scope.\r |
| 1416 | """\r |
| 1417 | num_lines[side] += 1\r |
| 1418 | # Handle case where no user markup is to be added, just return line of\r |
| 1419 | # text with user's line format to allow for usage of the line number.\r |
| 1420 | if format_key is None:\r |
| 1421 | return (num_lines[side],lines.pop(0)[2:])\r |
| 1422 | # Handle case of intraline changes\r |
| 1423 | if format_key == '?':\r |
| 1424 | text, markers = lines.pop(0), lines.pop(0)\r |
| 1425 | # find intraline changes (store change type and indices in tuples)\r |
| 1426 | sub_info = []\r |
| 1427 | def record_sub_info(match_object,sub_info=sub_info):\r |
| 1428 | sub_info.append([match_object.group(1)[0],match_object.span()])\r |
| 1429 | return match_object.group(1)\r |
| 1430 | change_re.sub(record_sub_info,markers)\r |
| 1431 | # process each tuple inserting our special marks that won't be\r |
| 1432 | # noticed by an xml/html escaper.\r |
| 1433 | for key,(begin,end) in sub_info[::-1]:\r |
| 1434 | text = text[0:begin]+'\0'+key+text[begin:end]+'\1'+text[end:]\r |
| 1435 | text = text[2:]\r |
| 1436 | # Handle case of add/delete entire line\r |
| 1437 | else:\r |
| 1438 | text = lines.pop(0)[2:]\r |
| 1439 | # if line of text is just a newline, insert a space so there is\r |
| 1440 | # something for the user to highlight and see.\r |
| 1441 | if not text:\r |
| 1442 | text = ' '\r |
| 1443 | # insert marks that won't be noticed by an xml/html escaper.\r |
| 1444 | text = '\0' + format_key + text + '\1'\r |
| 1445 | # Return line of text, first allow user's line formatter to do its\r |
| 1446 | # thing (such as adding the line number) then replace the special\r |
| 1447 | # marks with what the user's change markup.\r |
| 1448 | return (num_lines[side],text)\r |
| 1449 | \r |
| 1450 | def _line_iterator():\r |
| 1451 | """Yields from/to lines of text with a change indication.\r |
| 1452 | \r |
| 1453 | This function is an iterator. It itself pulls lines from a\r |
| 1454 | differencing iterator, processes them and yields them. When it can\r |
| 1455 | it yields both a "from" and a "to" line, otherwise it will yield one\r |
| 1456 | or the other. In addition to yielding the lines of from/to text, a\r |
| 1457 | boolean flag is yielded to indicate if the text line(s) have\r |
| 1458 | differences in them.\r |
| 1459 | \r |
| 1460 | Note, this function is purposefully not defined at the module scope so\r |
| 1461 | that data it needs from its parent function (within whose context it\r |
| 1462 | is defined) does not need to be of module scope.\r |
| 1463 | """\r |
| 1464 | lines = []\r |
| 1465 | num_blanks_pending, num_blanks_to_yield = 0, 0\r |
| 1466 | while True:\r |
| 1467 | # Load up next 4 lines so we can look ahead, create strings which\r |
| 1468 | # are a concatenation of the first character of each of the 4 lines\r |
| 1469 | # so we can do some very readable comparisons.\r |
| 1470 | while len(lines) < 4:\r |
| 1471 | try:\r |
| 1472 | lines.append(diff_lines_iterator.next())\r |
| 1473 | except StopIteration:\r |
| 1474 | lines.append('X')\r |
| 1475 | s = ''.join([line[0] for line in lines])\r |
| 1476 | if s.startswith('X'):\r |
| 1477 | # When no more lines, pump out any remaining blank lines so the\r |
| 1478 | # corresponding add/delete lines get a matching blank line so\r |
| 1479 | # all line pairs get yielded at the next level.\r |
| 1480 | num_blanks_to_yield = num_blanks_pending\r |
| 1481 | elif s.startswith('-?+?'):\r |
| 1482 | # simple intraline change\r |
| 1483 | yield _make_line(lines,'?',0), _make_line(lines,'?',1), True\r |
| 1484 | continue\r |
| 1485 | elif s.startswith('--++'):\r |
| 1486 | # in delete block, add block coming: we do NOT want to get\r |
| 1487 | # caught up on blank lines yet, just process the delete line\r |
| 1488 | num_blanks_pending -= 1\r |
| 1489 | yield _make_line(lines,'-',0), None, True\r |
| 1490 | continue\r |
| 1491 | elif s.startswith(('--?+', '--+', '- ')):\r |
| 1492 | # in delete block and see a intraline change or unchanged line\r |
| 1493 | # coming: yield the delete line and then blanks\r |
| 1494 | from_line,to_line = _make_line(lines,'-',0), None\r |
| 1495 | num_blanks_to_yield,num_blanks_pending = num_blanks_pending-1,0\r |
| 1496 | elif s.startswith('-+?'):\r |
| 1497 | # intraline change\r |
| 1498 | yield _make_line(lines,None,0), _make_line(lines,'?',1), True\r |
| 1499 | continue\r |
| 1500 | elif s.startswith('-?+'):\r |
| 1501 | # intraline change\r |
| 1502 | yield _make_line(lines,'?',0), _make_line(lines,None,1), True\r |
| 1503 | continue\r |
| 1504 | elif s.startswith('-'):\r |
| 1505 | # delete FROM line\r |
| 1506 | num_blanks_pending -= 1\r |
| 1507 | yield _make_line(lines,'-',0), None, True\r |
| 1508 | continue\r |
| 1509 | elif s.startswith('+--'):\r |
| 1510 | # in add block, delete block coming: we do NOT want to get\r |
| 1511 | # caught up on blank lines yet, just process the add line\r |
| 1512 | num_blanks_pending += 1\r |
| 1513 | yield None, _make_line(lines,'+',1), True\r |
| 1514 | continue\r |
| 1515 | elif s.startswith(('+ ', '+-')):\r |
| 1516 | # will be leaving an add block: yield blanks then add line\r |
| 1517 | from_line, to_line = None, _make_line(lines,'+',1)\r |
| 1518 | num_blanks_to_yield,num_blanks_pending = num_blanks_pending+1,0\r |
| 1519 | elif s.startswith('+'):\r |
| 1520 | # inside an add block, yield the add line\r |
| 1521 | num_blanks_pending += 1\r |
| 1522 | yield None, _make_line(lines,'+',1), True\r |
| 1523 | continue\r |
| 1524 | elif s.startswith(' '):\r |
| 1525 | # unchanged text, yield it to both sides\r |
| 1526 | yield _make_line(lines[:],None,0),_make_line(lines,None,1),False\r |
| 1527 | continue\r |
| 1528 | # Catch up on the blank lines so when we yield the next from/to\r |
| 1529 | # pair, they are lined up.\r |
| 1530 | while(num_blanks_to_yield < 0):\r |
| 1531 | num_blanks_to_yield += 1\r |
| 1532 | yield None,('','\n'),True\r |
| 1533 | while(num_blanks_to_yield > 0):\r |
| 1534 | num_blanks_to_yield -= 1\r |
| 1535 | yield ('','\n'),None,True\r |
| 1536 | if s.startswith('X'):\r |
| 1537 | raise StopIteration\r |
| 1538 | else:\r |
| 1539 | yield from_line,to_line,True\r |
| 1540 | \r |
| 1541 | def _line_pair_iterator():\r |
| 1542 | """Yields from/to lines of text with a change indication.\r |
| 1543 | \r |
| 1544 | This function is an iterator. It itself pulls lines from the line\r |
| 1545 | iterator. Its difference from that iterator is that this function\r |
| 1546 | always yields a pair of from/to text lines (with the change\r |
| 1547 | indication). If necessary it will collect single from/to lines\r |
| 1548 | until it has a matching pair from/to pair to yield.\r |
| 1549 | \r |
| 1550 | Note, this function is purposefully not defined at the module scope so\r |
| 1551 | that data it needs from its parent function (within whose context it\r |
| 1552 | is defined) does not need to be of module scope.\r |
| 1553 | """\r |
| 1554 | line_iterator = _line_iterator()\r |
| 1555 | fromlines,tolines=[],[]\r |
| 1556 | while True:\r |
| 1557 | # Collecting lines of text until we have a from/to pair\r |
| 1558 | while (len(fromlines)==0 or len(tolines)==0):\r |
| 1559 | from_line, to_line, found_diff =line_iterator.next()\r |
| 1560 | if from_line is not None:\r |
| 1561 | fromlines.append((from_line,found_diff))\r |
| 1562 | if to_line is not None:\r |
| 1563 | tolines.append((to_line,found_diff))\r |
| 1564 | # Once we have a pair, remove them from the collection and yield it\r |
| 1565 | from_line, fromDiff = fromlines.pop(0)\r |
| 1566 | to_line, to_diff = tolines.pop(0)\r |
| 1567 | yield (from_line,to_line,fromDiff or to_diff)\r |
| 1568 | \r |
| 1569 | # Handle case where user does not want context differencing, just yield\r |
| 1570 | # them up without doing anything else with them.\r |
| 1571 | line_pair_iterator = _line_pair_iterator()\r |
| 1572 | if context is None:\r |
| 1573 | while True:\r |
| 1574 | yield line_pair_iterator.next()\r |
| 1575 | # Handle case where user wants context differencing. We must do some\r |
| 1576 | # storage of lines until we know for sure that they are to be yielded.\r |
| 1577 | else:\r |
| 1578 | context += 1\r |
| 1579 | lines_to_write = 0\r |
| 1580 | while True:\r |
| 1581 | # Store lines up until we find a difference, note use of a\r |
| 1582 | # circular queue because we only need to keep around what\r |
| 1583 | # we need for context.\r |
| 1584 | index, contextLines = 0, [None]*(context)\r |
| 1585 | found_diff = False\r |
| 1586 | while(found_diff is False):\r |
| 1587 | from_line, to_line, found_diff = line_pair_iterator.next()\r |
| 1588 | i = index % context\r |
| 1589 | contextLines[i] = (from_line, to_line, found_diff)\r |
| 1590 | index += 1\r |
| 1591 | # Yield lines that we have collected so far, but first yield\r |
| 1592 | # the user's separator.\r |
| 1593 | if index > context:\r |
| 1594 | yield None, None, None\r |
| 1595 | lines_to_write = context\r |
| 1596 | else:\r |
| 1597 | lines_to_write = index\r |
| 1598 | index = 0\r |
| 1599 | while(lines_to_write):\r |
| 1600 | i = index % context\r |
| 1601 | index += 1\r |
| 1602 | yield contextLines[i]\r |
| 1603 | lines_to_write -= 1\r |
| 1604 | # Now yield the context lines after the change\r |
| 1605 | lines_to_write = context-1\r |
| 1606 | while(lines_to_write):\r |
| 1607 | from_line, to_line, found_diff = line_pair_iterator.next()\r |
| 1608 | # If another change within the context, extend the context\r |
| 1609 | if found_diff:\r |
| 1610 | lines_to_write = context-1\r |
| 1611 | else:\r |
| 1612 | lines_to_write -= 1\r |
| 1613 | yield from_line, to_line, found_diff\r |
| 1614 | \r |
| 1615 | \r |
| 1616 | _file_template = """\r |
| 1617 | <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"\r |
| 1618 | "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">\r |
| 1619 | \r |
| 1620 | <html>\r |
| 1621 | \r |
| 1622 | <head>\r |
| 1623 | <meta http-equiv="Content-Type"\r |
| 1624 | content="text/html; charset=ISO-8859-1" />\r |
| 1625 | <title></title>\r |
| 1626 | <style type="text/css">%(styles)s\r |
| 1627 | </style>\r |
| 1628 | </head>\r |
| 1629 | \r |
| 1630 | <body>\r |
| 1631 | %(table)s%(legend)s\r |
| 1632 | </body>\r |
| 1633 | \r |
| 1634 | </html>"""\r |
| 1635 | \r |
| 1636 | _styles = """\r |
| 1637 | table.diff {font-family:Courier; border:medium;}\r |
| 1638 | .diff_header {background-color:#e0e0e0}\r |
| 1639 | td.diff_header {text-align:right}\r |
| 1640 | .diff_next {background-color:#c0c0c0}\r |
| 1641 | .diff_add {background-color:#aaffaa}\r |
| 1642 | .diff_chg {background-color:#ffff77}\r |
| 1643 | .diff_sub {background-color:#ffaaaa}"""\r |
| 1644 | \r |
| 1645 | _table_template = """\r |
| 1646 | <table class="diff" id="difflib_chg_%(prefix)s_top"\r |
| 1647 | cellspacing="0" cellpadding="0" rules="groups" >\r |
| 1648 | <colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\r |
| 1649 | <colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\r |
| 1650 | %(header_row)s\r |
| 1651 | <tbody>\r |
| 1652 | %(data_rows)s </tbody>\r |
| 1653 | </table>"""\r |
| 1654 | \r |
| 1655 | _legend = """\r |
| 1656 | <table class="diff" summary="Legends">\r |
| 1657 | <tr> <th colspan="2"> Legends </th> </tr>\r |
| 1658 | <tr> <td> <table border="" summary="Colors">\r |
| 1659 | <tr><th> Colors </th> </tr>\r |
| 1660 | <tr><td class="diff_add"> Added </td></tr>\r |
| 1661 | <tr><td class="diff_chg">Changed</td> </tr>\r |
| 1662 | <tr><td class="diff_sub">Deleted</td> </tr>\r |
| 1663 | </table></td>\r |
| 1664 | <td> <table border="" summary="Links">\r |
| 1665 | <tr><th colspan="2"> Links </th> </tr>\r |
| 1666 | <tr><td>(f)irst change</td> </tr>\r |
| 1667 | <tr><td>(n)ext change</td> </tr>\r |
| 1668 | <tr><td>(t)op</td> </tr>\r |
| 1669 | </table></td> </tr>\r |
| 1670 | </table>"""\r |
| 1671 | \r |
| 1672 | class HtmlDiff(object):\r |
| 1673 | """For producing HTML side by side comparison with change highlights.\r |
| 1674 | \r |
| 1675 | This class can be used to create an HTML table (or a complete HTML file\r |
| 1676 | containing the table) showing a side by side, line by line comparison\r |
| 1677 | of text with inter-line and intra-line change highlights. The table can\r |
| 1678 | be generated in either full or contextual difference mode.\r |
| 1679 | \r |
| 1680 | The following methods are provided for HTML generation:\r |
| 1681 | \r |
| 1682 | make_table -- generates HTML for a single side by side table\r |
| 1683 | make_file -- generates complete HTML file with a single side by side table\r |
| 1684 | \r |
| 1685 | See tools/scripts/diff.py for an example usage of this class.\r |
| 1686 | """\r |
| 1687 | \r |
| 1688 | _file_template = _file_template\r |
| 1689 | _styles = _styles\r |
| 1690 | _table_template = _table_template\r |
| 1691 | _legend = _legend\r |
| 1692 | _default_prefix = 0\r |
| 1693 | \r |
| 1694 | def __init__(self,tabsize=8,wrapcolumn=None,linejunk=None,\r |
| 1695 | charjunk=IS_CHARACTER_JUNK):\r |
| 1696 | """HtmlDiff instance initializer\r |
| 1697 | \r |
| 1698 | Arguments:\r |
| 1699 | tabsize -- tab stop spacing, defaults to 8.\r |
| 1700 | wrapcolumn -- column number where lines are broken and wrapped,\r |
| 1701 | defaults to None where lines are not wrapped.\r |
| 1702 | linejunk,charjunk -- keyword arguments passed into ndiff() (used to by\r |
| 1703 | HtmlDiff() to generate the side by side HTML differences). See\r |
| 1704 | ndiff() documentation for argument default values and descriptions.\r |
| 1705 | """\r |
| 1706 | self._tabsize = tabsize\r |
| 1707 | self._wrapcolumn = wrapcolumn\r |
| 1708 | self._linejunk = linejunk\r |
| 1709 | self._charjunk = charjunk\r |
| 1710 | \r |
| 1711 | def make_file(self,fromlines,tolines,fromdesc='',todesc='',context=False,\r |
| 1712 | numlines=5):\r |
| 1713 | """Returns HTML file of side by side comparison with change highlights\r |
| 1714 | \r |
| 1715 | Arguments:\r |
| 1716 | fromlines -- list of "from" lines\r |
| 1717 | tolines -- list of "to" lines\r |
| 1718 | fromdesc -- "from" file column header string\r |
| 1719 | todesc -- "to" file column header string\r |
| 1720 | context -- set to True for contextual differences (defaults to False\r |
| 1721 | which shows full differences).\r |
| 1722 | numlines -- number of context lines. When context is set True,\r |
| 1723 | controls number of lines displayed before and after the change.\r |
| 1724 | When context is False, controls the number of lines to place\r |
| 1725 | the "next" link anchors before the next change (so click of\r |
| 1726 | "next" link jumps to just before the change).\r |
| 1727 | """\r |
| 1728 | \r |
| 1729 | return self._file_template % dict(\r |
| 1730 | styles = self._styles,\r |
| 1731 | legend = self._legend,\r |
| 1732 | table = self.make_table(fromlines,tolines,fromdesc,todesc,\r |
| 1733 | context=context,numlines=numlines))\r |
| 1734 | \r |
| 1735 | def _tab_newline_replace(self,fromlines,tolines):\r |
| 1736 | """Returns from/to line lists with tabs expanded and newlines removed.\r |
| 1737 | \r |
| 1738 | Instead of tab characters being replaced by the number of spaces\r |
| 1739 | needed to fill in to the next tab stop, this function will fill\r |
| 1740 | the space with tab characters. This is done so that the difference\r |
| 1741 | algorithms can identify changes in a file when tabs are replaced by\r |
| 1742 | spaces and vice versa. At the end of the HTML generation, the tab\r |
| 1743 | characters will be replaced with a nonbreakable space.\r |
| 1744 | """\r |
| 1745 | def expand_tabs(line):\r |
| 1746 | # hide real spaces\r |
| 1747 | line = line.replace(' ','\0')\r |
| 1748 | # expand tabs into spaces\r |
| 1749 | line = line.expandtabs(self._tabsize)\r |
| 1750 | # replace spaces from expanded tabs back into tab characters\r |
| 1751 | # (we'll replace them with markup after we do differencing)\r |
| 1752 | line = line.replace(' ','\t')\r |
| 1753 | return line.replace('\0',' ').rstrip('\n')\r |
| 1754 | fromlines = [expand_tabs(line) for line in fromlines]\r |
| 1755 | tolines = [expand_tabs(line) for line in tolines]\r |
| 1756 | return fromlines,tolines\r |
| 1757 | \r |
| 1758 | def _split_line(self,data_list,line_num,text):\r |
| 1759 | """Builds list of text lines by splitting text lines at wrap point\r |
| 1760 | \r |
| 1761 | This function will determine if the input text line needs to be\r |
| 1762 | wrapped (split) into separate lines. If so, the first wrap point\r |
| 1763 | will be determined and the first line appended to the output\r |
| 1764 | text line list. This function is used recursively to handle\r |
| 1765 | the second part of the split line to further split it.\r |
| 1766 | """\r |
| 1767 | # if blank line or context separator, just add it to the output list\r |
| 1768 | if not line_num:\r |
| 1769 | data_list.append((line_num,text))\r |
| 1770 | return\r |
| 1771 | \r |
| 1772 | # if line text doesn't need wrapping, just add it to the output list\r |
| 1773 | size = len(text)\r |
| 1774 | max = self._wrapcolumn\r |
| 1775 | if (size <= max) or ((size -(text.count('\0')*3)) <= max):\r |
| 1776 | data_list.append((line_num,text))\r |
| 1777 | return\r |
| 1778 | \r |
| 1779 | # scan text looking for the wrap point, keeping track if the wrap\r |
| 1780 | # point is inside markers\r |
| 1781 | i = 0\r |
| 1782 | n = 0\r |
| 1783 | mark = ''\r |
| 1784 | while n < max and i < size:\r |
| 1785 | if text[i] == '\0':\r |
| 1786 | i += 1\r |
| 1787 | mark = text[i]\r |
| 1788 | i += 1\r |
| 1789 | elif text[i] == '\1':\r |
| 1790 | i += 1\r |
| 1791 | mark = ''\r |
| 1792 | else:\r |
| 1793 | i += 1\r |
| 1794 | n += 1\r |
| 1795 | \r |
| 1796 | # wrap point is inside text, break it up into separate lines\r |
| 1797 | line1 = text[:i]\r |
| 1798 | line2 = text[i:]\r |
| 1799 | \r |
| 1800 | # if wrap point is inside markers, place end marker at end of first\r |
| 1801 | # line and start marker at beginning of second line because each\r |
| 1802 | # line will have its own table tag markup around it.\r |
| 1803 | if mark:\r |
| 1804 | line1 = line1 + '\1'\r |
| 1805 | line2 = '\0' + mark + line2\r |
| 1806 | \r |
| 1807 | # tack on first line onto the output list\r |
| 1808 | data_list.append((line_num,line1))\r |
| 1809 | \r |
| 1810 | # use this routine again to wrap the remaining text\r |
| 1811 | self._split_line(data_list,'>',line2)\r |
| 1812 | \r |
| 1813 | def _line_wrapper(self,diffs):\r |
| 1814 | """Returns iterator that splits (wraps) mdiff text lines"""\r |
| 1815 | \r |
| 1816 | # pull from/to data and flags from mdiff iterator\r |
| 1817 | for fromdata,todata,flag in diffs:\r |
| 1818 | # check for context separators and pass them through\r |
| 1819 | if flag is None:\r |
| 1820 | yield fromdata,todata,flag\r |
| 1821 | continue\r |
| 1822 | (fromline,fromtext),(toline,totext) = fromdata,todata\r |
| 1823 | # for each from/to line split it at the wrap column to form\r |
| 1824 | # list of text lines.\r |
| 1825 | fromlist,tolist = [],[]\r |
| 1826 | self._split_line(fromlist,fromline,fromtext)\r |
| 1827 | self._split_line(tolist,toline,totext)\r |
| 1828 | # yield from/to line in pairs inserting blank lines as\r |
| 1829 | # necessary when one side has more wrapped lines\r |
| 1830 | while fromlist or tolist:\r |
| 1831 | if fromlist:\r |
| 1832 | fromdata = fromlist.pop(0)\r |
| 1833 | else:\r |
| 1834 | fromdata = ('',' ')\r |
| 1835 | if tolist:\r |
| 1836 | todata = tolist.pop(0)\r |
| 1837 | else:\r |
| 1838 | todata = ('',' ')\r |
| 1839 | yield fromdata,todata,flag\r |
| 1840 | \r |
| 1841 | def _collect_lines(self,diffs):\r |
| 1842 | """Collects mdiff output into separate lists\r |
| 1843 | \r |
| 1844 | Before storing the mdiff from/to data into a list, it is converted\r |
| 1845 | into a single line of text with HTML markup.\r |
| 1846 | """\r |
| 1847 | \r |
| 1848 | fromlist,tolist,flaglist = [],[],[]\r |
| 1849 | # pull from/to data and flags from mdiff style iterator\r |
| 1850 | for fromdata,todata,flag in diffs:\r |
| 1851 | try:\r |
| 1852 | # store HTML markup of the lines into the lists\r |
| 1853 | fromlist.append(self._format_line(0,flag,*fromdata))\r |
| 1854 | tolist.append(self._format_line(1,flag,*todata))\r |
| 1855 | except TypeError:\r |
| 1856 | # exceptions occur for lines where context separators go\r |
| 1857 | fromlist.append(None)\r |
| 1858 | tolist.append(None)\r |
| 1859 | flaglist.append(flag)\r |
| 1860 | return fromlist,tolist,flaglist\r |
| 1861 | \r |
| 1862 | def _format_line(self,side,flag,linenum,text):\r |
| 1863 | """Returns HTML markup of "from" / "to" text lines\r |
| 1864 | \r |
| 1865 | side -- 0 or 1 indicating "from" or "to" text\r |
| 1866 | flag -- indicates if difference on line\r |
| 1867 | linenum -- line number (used for line number column)\r |
| 1868 | text -- line text to be marked up\r |
| 1869 | """\r |
| 1870 | try:\r |
| 1871 | linenum = '%d' % linenum\r |
| 1872 | id = ' id="%s%s"' % (self._prefix[side],linenum)\r |
| 1873 | except TypeError:\r |
| 1874 | # handle blank lines where linenum is '>' or ''\r |
| 1875 | id = ''\r |
| 1876 | # replace those things that would get confused with HTML symbols\r |
| 1877 | text=text.replace("&","&").replace(">",">").replace("<","<")\r |
| 1878 | \r |
| 1879 | # make space non-breakable so they don't get compressed or line wrapped\r |
| 1880 | text = text.replace(' ',' ').rstrip()\r |
| 1881 | \r |
| 1882 | return '<td class="diff_header"%s>%s</td><td nowrap="nowrap">%s</td>' \\r |
| 1883 | % (id,linenum,text)\r |
| 1884 | \r |
| 1885 | def _make_prefix(self):\r |
| 1886 | """Create unique anchor prefixes"""\r |
| 1887 | \r |
| 1888 | # Generate a unique anchor prefix so multiple tables\r |
| 1889 | # can exist on the same HTML page without conflicts.\r |
| 1890 | fromprefix = "from%d_" % HtmlDiff._default_prefix\r |
| 1891 | toprefix = "to%d_" % HtmlDiff._default_prefix\r |
| 1892 | HtmlDiff._default_prefix += 1\r |
| 1893 | # store prefixes so line format method has access\r |
| 1894 | self._prefix = [fromprefix,toprefix]\r |
| 1895 | \r |
| 1896 | def _convert_flags(self,fromlist,tolist,flaglist,context,numlines):\r |
| 1897 | """Makes list of "next" links"""\r |
| 1898 | \r |
| 1899 | # all anchor names will be generated using the unique "to" prefix\r |
| 1900 | toprefix = self._prefix[1]\r |
| 1901 | \r |
| 1902 | # process change flags, generating middle column of next anchors/links\r |
| 1903 | next_id = ['']*len(flaglist)\r |
| 1904 | next_href = ['']*len(flaglist)\r |
| 1905 | num_chg, in_change = 0, False\r |
| 1906 | last = 0\r |
| 1907 | for i,flag in enumerate(flaglist):\r |
| 1908 | if flag:\r |
| 1909 | if not in_change:\r |
| 1910 | in_change = True\r |
| 1911 | last = i\r |
| 1912 | # at the beginning of a change, drop an anchor a few lines\r |
| 1913 | # (the context lines) before the change for the previous\r |
| 1914 | # link\r |
| 1915 | i = max([0,i-numlines])\r |
| 1916 | next_id[i] = ' id="difflib_chg_%s_%d"' % (toprefix,num_chg)\r |
| 1917 | # at the beginning of a change, drop a link to the next\r |
| 1918 | # change\r |
| 1919 | num_chg += 1\r |
| 1920 | next_href[last] = '<a href="#difflib_chg_%s_%d">n</a>' % (\r |
| 1921 | toprefix,num_chg)\r |
| 1922 | else:\r |
| 1923 | in_change = False\r |
| 1924 | # check for cases where there is no content to avoid exceptions\r |
| 1925 | if not flaglist:\r |
| 1926 | flaglist = [False]\r |
| 1927 | next_id = ['']\r |
| 1928 | next_href = ['']\r |
| 1929 | last = 0\r |
| 1930 | if context:\r |
| 1931 | fromlist = ['<td></td><td> No Differences Found </td>']\r |
| 1932 | tolist = fromlist\r |
| 1933 | else:\r |
| 1934 | fromlist = tolist = ['<td></td><td> Empty File </td>']\r |
| 1935 | # if not a change on first line, drop a link\r |
| 1936 | if not flaglist[0]:\r |
| 1937 | next_href[0] = '<a href="#difflib_chg_%s_0">f</a>' % toprefix\r |
| 1938 | # redo the last link to link to the top\r |
| 1939 | next_href[last] = '<a href="#difflib_chg_%s_top">t</a>' % (toprefix)\r |
| 1940 | \r |
| 1941 | return fromlist,tolist,flaglist,next_href,next_id\r |
| 1942 | \r |
| 1943 | def make_table(self,fromlines,tolines,fromdesc='',todesc='',context=False,\r |
| 1944 | numlines=5):\r |
| 1945 | """Returns HTML table of side by side comparison with change highlights\r |
| 1946 | \r |
| 1947 | Arguments:\r |
| 1948 | fromlines -- list of "from" lines\r |
| 1949 | tolines -- list of "to" lines\r |
| 1950 | fromdesc -- "from" file column header string\r |
| 1951 | todesc -- "to" file column header string\r |
| 1952 | context -- set to True for contextual differences (defaults to False\r |
| 1953 | which shows full differences).\r |
| 1954 | numlines -- number of context lines. When context is set True,\r |
| 1955 | controls number of lines displayed before and after the change.\r |
| 1956 | When context is False, controls the number of lines to place\r |
| 1957 | the "next" link anchors before the next change (so click of\r |
| 1958 | "next" link jumps to just before the change).\r |
| 1959 | """\r |
| 1960 | \r |
| 1961 | # make unique anchor prefixes so that multiple tables may exist\r |
| 1962 | # on the same page without conflict.\r |
| 1963 | self._make_prefix()\r |
| 1964 | \r |
| 1965 | # change tabs to spaces before it gets more difficult after we insert\r |
| 1966 | # markkup\r |
| 1967 | fromlines,tolines = self._tab_newline_replace(fromlines,tolines)\r |
| 1968 | \r |
| 1969 | # create diffs iterator which generates side by side from/to data\r |
| 1970 | if context:\r |
| 1971 | context_lines = numlines\r |
| 1972 | else:\r |
| 1973 | context_lines = None\r |
| 1974 | diffs = _mdiff(fromlines,tolines,context_lines,linejunk=self._linejunk,\r |
| 1975 | charjunk=self._charjunk)\r |
| 1976 | \r |
| 1977 | # set up iterator to wrap lines that exceed desired width\r |
| 1978 | if self._wrapcolumn:\r |
| 1979 | diffs = self._line_wrapper(diffs)\r |
| 1980 | \r |
| 1981 | # collect up from/to lines and flags into lists (also format the lines)\r |
| 1982 | fromlist,tolist,flaglist = self._collect_lines(diffs)\r |
| 1983 | \r |
| 1984 | # process change flags, generating middle column of next anchors/links\r |
| 1985 | fromlist,tolist,flaglist,next_href,next_id = self._convert_flags(\r |
| 1986 | fromlist,tolist,flaglist,context,numlines)\r |
| 1987 | \r |
| 1988 | s = []\r |
| 1989 | fmt = ' <tr><td class="diff_next"%s>%s</td>%s' + \\r |
| 1990 | '<td class="diff_next">%s</td>%s</tr>\n'\r |
| 1991 | for i in range(len(flaglist)):\r |
| 1992 | if flaglist[i] is None:\r |
| 1993 | # mdiff yields None on separator lines skip the bogus ones\r |
| 1994 | # generated for the first line\r |
| 1995 | if i > 0:\r |
| 1996 | s.append(' </tbody> \n <tbody>\n')\r |
| 1997 | else:\r |
| 1998 | s.append( fmt % (next_id[i],next_href[i],fromlist[i],\r |
| 1999 | next_href[i],tolist[i]))\r |
| 2000 | if fromdesc or todesc:\r |
| 2001 | header_row = '<thead><tr>%s%s%s%s</tr></thead>' % (\r |
| 2002 | '<th class="diff_next"><br /></th>',\r |
| 2003 | '<th colspan="2" class="diff_header">%s</th>' % fromdesc,\r |
| 2004 | '<th class="diff_next"><br /></th>',\r |
| 2005 | '<th colspan="2" class="diff_header">%s</th>' % todesc)\r |
| 2006 | else:\r |
| 2007 | header_row = ''\r |
| 2008 | \r |
| 2009 | table = self._table_template % dict(\r |
| 2010 | data_rows=''.join(s),\r |
| 2011 | header_row=header_row,\r |
| 2012 | prefix=self._prefix[1])\r |
| 2013 | \r |
| 2014 | return table.replace('\0+','<span class="diff_add">'). \\r |
| 2015 | replace('\0-','<span class="diff_sub">'). \\r |
| 2016 | replace('\0^','<span class="diff_chg">'). \\r |
| 2017 | replace('\1','</span>'). \\r |
| 2018 | replace('\t',' ')\r |
| 2019 | \r |
| 2020 | del re\r |
| 2021 | \r |
| 2022 | def restore(delta, which):\r |
| 2023 | r"""\r |
| 2024 | Generate one of the two sequences that generated a delta.\r |
| 2025 | \r |
| 2026 | Given a `delta` produced by `Differ.compare()` or `ndiff()`, extract\r |
| 2027 | lines originating from file 1 or 2 (parameter `which`), stripping off line\r |
| 2028 | prefixes.\r |
| 2029 | \r |
| 2030 | Examples:\r |
| 2031 | \r |
| 2032 | >>> diff = ndiff('one\ntwo\nthree\n'.splitlines(1),\r |
| 2033 | ... 'ore\ntree\nemu\n'.splitlines(1))\r |
| 2034 | >>> diff = list(diff)\r |
| 2035 | >>> print ''.join(restore(diff, 1)),\r |
| 2036 | one\r |
| 2037 | two\r |
| 2038 | three\r |
| 2039 | >>> print ''.join(restore(diff, 2)),\r |
| 2040 | ore\r |
| 2041 | tree\r |
| 2042 | emu\r |
| 2043 | """\r |
| 2044 | try:\r |
| 2045 | tag = {1: "- ", 2: "+ "}[int(which)]\r |
| 2046 | except KeyError:\r |
| 2047 | raise ValueError, ('unknown delta choice (must be 1 or 2): %r'\r |
| 2048 | % which)\r |
| 2049 | prefixes = (" ", tag)\r |
| 2050 | for line in delta:\r |
| 2051 | if line[:2] in prefixes:\r |
| 2052 | yield line[2:]\r |
| 2053 | \r |
| 2054 | def _test():\r |
| 2055 | import doctest, difflib\r |
| 2056 | return doctest.testmod(difflib)\r |
| 2057 | \r |
| 2058 | if __name__ == "__main__":\r |
| 2059 | _test()\r |
projects / mirror_edk2.git / blame