0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
1 |
class Schema : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
2 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
3 |
def __init__(self, schema) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
4 |
self._schema = schema |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
5 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
6 |
def get_attrs(self, entity) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
7 |
return self._schema[entity][0] |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
8 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
9 |
def get_relations(self, entity) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
10 |
return self._schema[entity][1] |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
11 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
12 |
def get_attr_index(self, entity, attr) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
13 |
return list(self._schema[entity][0]).index(attr) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
14 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
15 |
SCHEMA = Schema({'societe': ( ('nom','ville'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
16 |
[('concerne_par','affaire'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
17 |
] ), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
18 |
'affaire': ( ('ref',), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
19 |
[('concerne','societe'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
20 |
('concerne_par', 'document') |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
21 |
] ), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
22 |
'document':( ('fichier', 'annee','mois','jour','type'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
23 |
[('concerne','affaire'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
24 |
] ), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
25 |
}) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
26 |
|
1802
|
27 |
|
0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
28 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
29 |
DATA = { 'societe': [ ('CETIAD', 'Dijon'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
30 |
('EDF_R&D', 'Clamart'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
31 |
('Logilab', 'Paris'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
32 |
], |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
33 |
'affaire': [ ('CTIA01', 'CETIAD'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
34 |
('EDFR01', 'EDF_R&D'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
35 |
('EDFR02', 'EDF_R&D'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
36 |
], |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
37 |
'document':[ ('CTIA01-040906-PRE-1-01.pdf','2004','09','06','PRE','CTIA01'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
38 |
('EDFR01-050201-CLI-1-01.pdf','2005','02','01','CLI','EDFR01'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
39 |
('EDFR01-050322-OFR-1-01.pdf','2005','03','22','OFR','EDFR01'), |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
40 |
], |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
41 |
} |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
42 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
43 |
def get_data(entity, where=[]) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
44 |
for value in DATA[entity] : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
45 |
for index, val in where : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
46 |
if value[index] != val : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
47 |
break |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
48 |
else : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
49 |
yield value |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
50 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
51 |
class PathParser : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
52 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
53 |
def __init__(self, schema, path) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
54 |
self.schema = schema |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
55 |
self.path = path |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
56 |
self._components = iter([comp for comp in self.path.split('/') if comp]) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
57 |
self._entity = None |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
58 |
self._attr = None |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
59 |
self._rel = None |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
60 |
self._restrictions = [] |
1802
|
61 |
|
0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
62 |
def parse(self) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
63 |
self._entity = self._components.next() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
64 |
try: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
65 |
self.process_entity() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
66 |
except StopIteration : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
67 |
pass |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
68 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
69 |
def process_entity(self) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
70 |
_next = self._components.next() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
71 |
if _next in self.schema.get_attrs(self._entity) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
72 |
self._attr = _next |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
73 |
_next = self._components.next() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
74 |
self._restrictions.append( (self._entity, self._attr, _next) ) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
75 |
self._attr = None |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
76 |
self._rel = None |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
77 |
self.process_entity() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
78 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
79 |
def get_list(self) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
80 |
if self._rel : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
81 |
return |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
82 |
elif self._attr : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
83 |
where = [] |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
84 |
for e,a,v in self._restrictions : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
85 |
i = self.schema.get_attr_index(e, a) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
86 |
where.append( (i,v) ) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
87 |
i = self.schema.get_attr_index(self._entity, self._attr) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
88 |
for values in get_data(self._entity,where) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
89 |
yield values[i]+'/' |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
90 |
else : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
91 |
attr_restrict = [a for e,a,v in self._restrictions] |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
92 |
for attr in self.schema.get_attrs(self._entity) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
93 |
if attr not in attr_restrict : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
94 |
yield attr+'/' |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
95 |
for data in DATA[self._entity]: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
96 |
yield data[0] |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
97 |
for nom, entity in self.schema.get_relations(self._entity) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
98 |
yield nom+'/' |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
99 |
yield entity+'/' |
1802
|
100 |
|
0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
101 |
def ls(path) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
102 |
p = PathParser(SCHEMA,path) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
103 |
p.parse() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
104 |
return list(p.get_list()) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
105 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
106 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
107 |
class SytPathParser : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
108 |
|
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
109 |
def __init__(self, schema, path) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
110 |
self.schema = schema |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
111 |
self.path = path |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
112 |
self._components = iter([comp for comp in self.path.split('/') if comp]) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
113 |
self._e_type = None |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
114 |
self._restrictions = [] |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
115 |
self._alphabet = list('ABCDEFGHIJKLMNOPQRSTUVWXYZ') |
1802
|
116 |
|
0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
117 |
def parse(self): |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
118 |
self._var = self._alphabet.pop(0) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
119 |
self._e_type = self._components.next() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
120 |
e_type = self._e_type.capitalize() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
121 |
self._restrictions.append('%s is %s' % (self._var, e_type)) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
122 |
try: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
123 |
self.process_entity() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
124 |
except StopIteration : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
125 |
pass |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
126 |
return 'Any %s WHERE %s' % (self._var, ', '.join(self._restrictions)) |
1802
|
127 |
|
0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
128 |
def process_entity(self) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
129 |
_next = self._components.next() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
130 |
if _next in self.schema.get_attrs(self._e_type) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
131 |
attr = _next |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
132 |
try: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
133 |
_next = self._components.next() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
134 |
self._restrictions.append('%s %s %s' % (self._var, attr, _next)) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
135 |
except StopIteration: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
136 |
a_var = self._alphabet.pop(0) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
137 |
self._restrictions.append('%s %s %s' % (self._var, attr, a_var) ) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
138 |
self._var = a_var |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
139 |
raise |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
140 |
elif _next in [r for r,e in self.schema.get_relations(self._e_type)]: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
141 |
rel = _next |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
142 |
r_var = self._alphabet.pop(0) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
143 |
self._restrictions.append('%s %s %s' % (self._var, rel, r_var)) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
144 |
self._var = r_var |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
145 |
try: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
146 |
_next = self._components.next() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
147 |
self._restrictions.append('%s is %s' % (r_var, _next.capitalize())) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
148 |
except StopIteration: |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
149 |
raise |
1802
|
150 |
self.process_entity() |
0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
151 |
|
1802
|
152 |
|
0
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
153 |
def to_rql(path) : |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
154 |
p = SytPathParser(SCHEMA,path) |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
155 |
return p.parse() |
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff
changeset
|
156 |
|