1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
|
# Create a Direct Acyclic Rule Graph (DARG)
|
>
>
|
>
>
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
|
"""
Grammalecte: compile rules
Create a Direct Acyclic Rule Graphs (DARGs)
"""
import re
import traceback
import json
import darg
dACTIONS = {}
dFUNCTIONS = {}
def prepareFunction (s, bTokenValue=False):
"convert simple rule syntax to a string of Python code"
s = s.replace("__also__", "bCondMemo")
s = s.replace("__else__", "not bCondMemo")
s = re.sub(r"(morph|analyse|displayInfo)[(]\\(\d+)", 'g_\\1(lToken[\\2+nTokenOffset]', s)
s = re.sub(r"(select|exclude|define)[(][\\](\d+)", 'g_\\1(lToken[\\2+nTokenOffset], dTags', s)
s = re.sub(r"(tag_before|tag_after)[(][\\](\d+)", 'g_\\1(lToken[\\2+nTokenOffset], dTags', s)
s = re.sub(r"(switchGender|has(?:Mas|Fem)Form)[(]\\(\d+)", '\\1(lToken[\\2+nTokenOffset]["sValue"]', s)
s = re.sub(r"(morph|analyse)\(>1", 'g_\\1(lToken[nLastToken+1]', s) # next token
|
| ︙ | | | ︙ | |
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
|
return s
def genTokenLines (sTokenLine, dDef):
"tokenize a string and return a list of lines of tokens"
lToken = sTokenLine.split()
lTokenLines = None
for i, sToken in enumerate(lToken):
# optional token?
bNullPossible = sToken.startswith("?") and sToken.endswith("¿")
if bNullPossible:
sToken = sToken[1:-1]
# token with definition?
if sToken.startswith("({") and sToken.endswith("})") and sToken[1:-1] in dDef:
sToken = "(" + dDef[sToken[1:-1]] + ")"
|
|
|
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
|
return s
def genTokenLines (sTokenLine, dDef):
"tokenize a string and return a list of lines of tokens"
lToken = sTokenLine.split()
lTokenLines = None
for sToken in lToken:
# optional token?
bNullPossible = sToken.startswith("?") and sToken.endswith("¿")
if bNullPossible:
sToken = sToken[1:-1]
# token with definition?
if sToken.startswith("({") and sToken.endswith("})") and sToken[1:-1] in dDef:
sToken = "(" + dDef[sToken[1:-1]] + ")"
|
| ︙ | | | ︙ | |
92
93
94
95
96
97
98
99
100
101
102
103
104
105
|
for aRule in lTokenLines:
aRule.append(sToken)
for aRule in lTokenLines:
yield aRule
def createRule (iLine, sRuleName, sTokenLine, iActionBlock, sActions, nPriority, dDef):
# print(iLine, "//", sRuleName, "//", sTokenLine, "//", sActions, "//", nPriority)
for lToken in genTokenLines(sTokenLine, dDef):
# Calculate positions
dPos = {} # key: iGroup, value: iToken
iGroup = 0
for i, sToken in enumerate(lToken):
if sToken.startswith("(") and sToken.endswith(")"):
|
>
|
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
|
for aRule in lTokenLines:
aRule.append(sToken)
for aRule in lTokenLines:
yield aRule
def createRule (iLine, sRuleName, sTokenLine, iActionBlock, sActions, nPriority, dDef):
"generator: create rule as list"
# print(iLine, "//", sRuleName, "//", sTokenLine, "//", sActions, "//", nPriority)
for lToken in genTokenLines(sTokenLine, dDef):
# Calculate positions
dPos = {} # key: iGroup, value: iToken
iGroup = 0
for i, sToken in enumerate(lToken):
if sToken.startswith("(") and sToken.endswith(")"):
|
| ︙ | | | ︙ | |
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
|
dACTIONS[sActionId] = aAction
lResult = list(lToken)
lResult.extend(["##"+str(iLine), sActionId])
yield lResult
def changeReferenceToken (sText, dPos):
for i in range(len(dPos), 0, -1):
sText = sText.replace("\\"+str(i), "\\"+str(dPos[i]))
return sText
def checkTokenNumbers (sText, sActionId, nToken):
for x in re.finditer(r"\\(\d+)", sText):
if int(x.group(1)) > nToken:
print("# Error in token index at line " + sActionId + " ("+str(nToken)+" tokens only)")
print(sText)
def checkIfThereIsCode (sText, sActionId):
if re.search("[.]\\w+[(]|sugg\\w+[(]|\\([0-9]|\\[[0-9]", sText):
print("# Warning at line " + sActionId + ": This message looks like code. Line should probably begin with =")
print(sText)
def createAction (sActionId, sAction, nPriority, nToken, dPos):
# Option
sOption = False
m = re.match("/(\\w+)/", sAction)
if m:
sOption = m.group(1)
sAction = sAction[m.end():].strip()
# valid action?
|
>
>
>
>
|
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
|
dACTIONS[sActionId] = aAction
lResult = list(lToken)
lResult.extend(["##"+str(iLine), sActionId])
yield lResult
def changeReferenceToken (sText, dPos):
"change group reference in <sText> with values in <dPos>"
for i in range(len(dPos), 0, -1):
sText = sText.replace("\\"+str(i), "\\"+str(dPos[i]))
return sText
def checkTokenNumbers (sText, sActionId, nToken):
"check if token references in <sText> greater than <nToken> (debugging)"
for x in re.finditer(r"\\(\d+)", sText):
if int(x.group(1)) > nToken:
print("# Error in token index at line " + sActionId + " ("+str(nToken)+" tokens only)")
print(sText)
def checkIfThereIsCode (sText, sActionId):
"check if there is code in <sText> (debugging)"
if re.search("[.]\\w+[(]|sugg\\w+[(]|\\([0-9]|\\[[0-9]", sText):
print("# Warning at line " + sActionId + ": This message looks like code. Line should probably begin with =")
print(sText)
def createAction (sActionId, sAction, nPriority, nToken, dPos):
"create action rule as a list"
# Option
sOption = False
m = re.match("/(\\w+)/", sAction)
if m:
sOption = m.group(1)
sAction = sAction[m.end():].strip()
# valid action?
|
| ︙ | | | ︙ | |
365
366
367
368
369
370
371
372
373
374
375
376
377
|
print("\nActions:")
for sActionName, aAction in dACTIONS.items():
print(sActionName, aAction)
print("\nFunctions:")
print(sPyCallables)
# Result
d = {
"graph_callables": sPyCallables,
"rules_graphs": dAllGraph,
"rules_actions": dACTIONS
}
return d
|
|
<
|
374
375
376
377
378
379
380
381
382
383
384
385
|
print("\nActions:")
for sActionName, aAction in dACTIONS.items():
print(sActionName, aAction)
print("\nFunctions:")
print(sPyCallables)
# Result
return {
"graph_callables": sPyCallables,
"rules_graphs": dAllGraph,
"rules_actions": dACTIONS
}
|