Grammalecte  Diff

#!python3

"""
Operations on strings:
- calculate distance between two strings
- transform strings with transformation codes
"""


#### DISTANCE CALCULATIONS

def longestCommonSubstring (s1, s2):
    "longest common substring"
    # http://en.wikipedia.org/wiki/Longest_common_substring_problem
    # http://en.wikibooks.org/wiki/Algorithm_implementation/Strings/Longest_common_substring
    lMatrix = [ [0]*(1+len(s2)) for i in range(1+len(s1)) ]
    nLongest, nLongestX = 0, 0
    for x in range(1, 1+len(s1)):
        for y in range(1, 1+len(s2)):
            if s1[x-1] == s2[y-1]:
                lMatrix[x][y] = lMatrix[x-1][y-1] + 1
                if lMatrix[x][y] > nLongest:
                    nLongest = lMatrix[x][y]
                    nLongestX = x
            else:
                lMatrix[x][y] = 0
    return s1[nLongestX-nLongest : nLongestX]


def distanceDamerauLevenshtein (s1, s2):
    "distance of Damerau-Levenshtein between <s1> and <s2>"
    # https://fr.wikipedia.org/wiki/Distance_de_Damerau-Levenshtein
    d = {}
    nLen1 = len(s1)

        return len(s1)
    nLen1, nLen2 = len(s1), len(s2)
    i1, i2 = 0, 0   # Cursors for each string
    nLargestCS = 0  # Largest common substring
    nLocalCS = 0    # Local common substring
    nTrans = 0      # Number of transpositions ('ab' vs 'ba')
    lOffset = []    # Offset pair array, for computing the transpositions

    while i1 < nLen1 and i2 < nLen2:
        if s1[i1] == s2[i2]:
            nLocalCS += 1
            # Check if current match is a transposition
            bTrans = False
            i = 0
            while i < len(lOffset):

            nLocalCS = 0
            i1 = i2 = min(i1, i2)
    nLargestCS += nLocalCS
    return round(max(nLen1, nLen2) - nLargestCS + nTrans)


def showDistance (s1, s2):
    "display Damerau-Levenshtein distance and Sift4 distance between <s1> and <s2>"
    print("Damerau-Levenshtein: " + s1 + "/" + s2 + " = " + distanceDamerauLevenshtein(s1, s2))
    print("Sift4:" + s1 + "/" + s2 + " = " + distanceSift4(s1, s2))




#### STEMMING OPERATIONS

## No stemming

def noStemming (sFlex, sStem):
    "return <sStem>"
    return sStem

def rebuildWord (sFlex, sCode1, sCode2):
    """ Change <sFlex> with codes (each inserts a char at a defined possition).
        <I forgot what purpose it has…>
    """
    if sCode1 == "_":
        return sFlex
    n, c = sCode1.split(":")
    sFlex = sFlex[:n] + c + sFlex[n:]
    if sCode2 == "_":
        return sFlex
    n, c = sCode2.split(":")
    return sFlex[:n] + c + sFlex[n:]


## Define affixes for stemming

# Note: 48 is the ASCII code for "0"


# Suffix only
def defineSuffixCode (sFlex, sStem):

    if sFlex == sStem:
        return "0"
    jSfx = 0
    for i in range(min(len(sFlex), len(sStem))):
        if sFlex[i] != sStem[i]:
            break
        jSfx += 1
    return chr(len(sFlex)-jSfx+48) + sStem[jSfx:]


def changeWordWithSuffixCode (sWord, sSfxCode):
    "apply transformation code <sSfxCode> on <sWord> and return the result string"
    if sSfxCode == "0":
        return sWord
    return sWord[:-(ord(sSfxCode[0])-48)] + sSfxCode[1:]  if sSfxCode[0] != '0'  else sWord + sSfxCode[1:]


# Prefix and suffix

def defineAffixCode (sFlex, sStem):
    """ Returns a string defining how to get stem from flexion. Examples:
            "0" if stem = flexion
            "stem" if no common substring
            "n(pfx)/m(sfx)"
        with n and m: chars with numeric meaning, "0" = 0, "1" = 1, ... ":" = 10, etc. (See ASCII table.) Says how many letters to strip from flexion.
            pfx [optional]: string to add before the flexion
            sfx [optional]: string to add after the flexion
    """
    if sFlex == sStem:
        return "0"
    # is stem a substring of flexion?
    n = sFlex.find(sStem)
    if n >= 0:

        n = sFlex.find(sSubs)
        m = len(sFlex) - (len(sSubs)+n)
        return chr(n+48) + sPfx + "/" + chr(m+48) + sSfx
    return sStem


def changeWordWithAffixCode (sWord, sAffCode):
    "apply transformation code <sAffCode> on <sWord> and return the result string"
    if sAffCode == "0":
        return sWord
    if '/' not in sAffCode:
        return sAffCode
    sPfxCode, sSfxCode = sAffCode.split('/')
    sWord = sPfxCode[1:] + sWord[(ord(sPfxCode[0])-48):]
    return sWord[:-(ord(sSfxCode[0])-48)] + sSfxCode[1:]  if sSfxCode[0] != '0'  else sWord + sSfxCode[1:]